Hardware Hacker 2
Hardware Hacker 2
Hardware Hacker 2
Don Lancasters
Hardware Hacker
Selected reprints volume III Radio-Electronics series (January 1990 - December 1991)
Copyright c. 1989 by Don Lancaster and Synergetics Box 809, Thatcher, AZ 85552 (602) 428-4073 Book-on-demand self-published using the Apple IIe computer, a duplex LaserJet and the PostScript language. All graphics were done in their entirety by ProDOS AppleWriter 2.1.
ii
Introduction Welcome to the third volume in our ongoing series of my Hardware Hacker
reprints from Radio Electronics magazine. Who have recently renamed themselves Electronics Now. This volume includes the January 1990 through December 1991 columns. As before, I do use the term "hacker" in its original context, namely anyone who is genuinely interested in pushing all the limits of software and hardware and is willing to do so in a fully cooperative, low cost and sharing network environment. Ive tried to expand the topics beyond the strict boundaries of traditional electronics. Going into the science and math subjects many of you have asked for over our voice helpline. And even looking at controversial pseudoscience and "free energy" topics. My goal in flirting with the "not even wrong" pseudoscience and perpetual motion stuff is to pile it up on center stage and shine a bright light on it. To get you to conclude that "Yup, that sure is a big pile all right." Very little of this can withstand critical inquiry. Some of the more popular subjects in this volume include fractals, chaos science, alternate energy, electronic levels, magnetic refrigeration, vortex coolers, video compression, wavelets, caller id, data curve fitting, and GPS navigation. Plus dozens more. Preprints and reprints of most of my columns are also available electronically via GEnie1 PSRT , presented in Hacker Data Format . For GEnie access, have your modem dial (800) 638-8369. Just as soon as you verify your software connect, enter an HHH . After the prompt, enter JOINGENIE . When you are asked for your keycode, enter DMD524. I continue to offer my no-charge voice helpline at (602) 428-4073 , and offer free referrals from our Synergetics Consultants Network. Ive also got a free and newly expanded Hardware Hacker Inside Secrets and catalog waiting for you when you call or write. You can also grab this as SYNCAT1.PS off GEnie PSRT , or else request one through my Internet address of SYNERGETICS@GENIE.GEIS.COM. The Electronics Now subscription number is (516) 293-3000.
iii
About the Author As he said in his classic Incredible Secret Money Machine II, Don Lancaster
writes books. And quests tinajas. Microcomputer pioneer and guru Don Lancaster is the author of 28 books, 2 videos, and countless articles. These include his TTL Cookbook and CMOS Cookbook internationally published million seller classics. Don is considered by some to be the father of the personal computer, for his early ground-breaking work with hacker digital electronics and low cost video terminal displays. He is considered by others to be the patron saint of the Walter Mitties of the World. And by yet others as the Uh, I guess we better skip that one. Don is the Hardware Hacker in Electronics Now (516) 293-3000, runs his Resource Bin in Nuts & Volts (714) 371-8497, and is the Blatant Opportunist in Midnight Engineering (719) 254-4558 . Don is also the sysop for GEnie PSRT , a leading Hardware hacking, PostScript, laser publishing, and money machine on-line resource. His Ask the Guru columns continue to be published electronically on PSRT. For fast GEnie modem access, dial (800) 638-8369. Then HHH followed by JOINGENIE and DMD524. Don is also the head honcho of Synergetics, a new-age design and consulting firm that specializes in computer applications, laser printing, PostScript development work, electronic prototypes, Book-on-demand publishing, technical writing, and innovative software design. His avocations include firefighting, cave exploration, bicycling, and, of course, tinaja questing. Don maintains a no-charge voice helpline at (602) 428-4073. He does welcome your calls and letters. The best calling times are 8-5 weekdays, Mountain Standard Time. Referrals to Dons Synergetics Consultants Network are available through this same number. 3
iv
Table of Contents
PostScript fractal fern Unusual cooling ideas Chaos science resources Thermoelectric modules Superintelligent machines Dew sensing circuits Direct toner pc update Midnight Engineering Hacker cold fusion kits Alarm & security resources LVDT position detectors Treasure finding circuitry Switchable analog inverter Synchronous demodulators Desktop accessories contest Organic vapor detectors Low melting point alloys Ultraviolet flame detector Alternatie energy resources Five band equalizer circuit Hall effect resources Shape memory alloys Cycolor printing secrets Low noise amplification Unusual mechanical stuff Electronic levels Low cost memory New form of matter Cases and enclosures LaserWriter tech info
Electronic tune chips Unusual book resources Ten speed blender control Power fundamentals A hacker interchange standard Huffman codes video line drivers Run length encoding New compaction chips Amaterur radio astronomy Pertual motion Audio log voltmeter Independent research The Magnetocaloric effect Refrigeration breakthrough Using SAW devices Designing active filter Digital signal processing Switched capacitor filters Unique digital thermometer Using vortex coolers Polyphase generators Walking ring counters Magnetic refrigeration Electronic trade journals Discrete cosine transforms Digital sinewave generator More on video compression Santa Claus machine update Audiovisual switcher circuit
more . . .
Audio level meters A cold fusion update Alternates to patenting Understanding decibels Wholesale surplus sources Low cost visible lasers Electric motor resources Induction motor controls New wavelet math theory dual digital potentiometer Machine shop resources Infrared people detection Elegant simplicity contest Liquid crystal battery tester Exploiting paradigm shifts Reloading toner cartridges New electronic phone book More Santa Claus machines Secret laser printer manuals VHF & microwave resources Electronic tuning diodes Two unusual newsletters Parametric amplification Association book resources Preventing modem dropouts Power electronic resources Video crosshatch generator More wavelet breakthroughs Secrets of video compression Understanding transformation
Electric power research info Focused X-ray breakthrough C-60 and the new Buckyballs Getting telephone information Caller number delivery secrets Driving inductive loads More on phone caller ID Bakerizing and laminating Alternators as stepper motors Programmable logic resources Curve fitting fuzzy data Tesla Coils and resources Meeting FCC regulations Flashlight battery supplies Thoughts on E field machines SMPTE video time codes Electronic dog tag contest Case & enclosure resources Solar energy breakthrough? Photovoltaic panel bargains Fractals & chaos update Standards resource info More on FM stereo chips Home energy monitoring Current transformer ideas Effective shielding ideas Computer monitor basics VGA adaptor for Mac LC Recurrent flyback sweeps GPS navigation resources
vi
Don Lancasters
Hardware Hacker
January, 1990
et us start off with some new update info for our previous columns. The 38.4 Kilohertz crystal needed for the stereo broadcaster is available for around $3 from Statek, who stock quite a wide variety of small and cheap low frequency crystals. Because 38.4 kHz is now a standard modem value, it might end up a better choice than the usual 38.0 kHz stereo multiplex frequency, and should work out just about as good. Probably your best way to assemble this project, though, is to go to one of the kits offered in the classified section. UPDATE: Improved circuits can be found in Hardware Hacker #52. One interesting use for this chip is to measure the torque on a rotating shaft. A $1.50 part can easily replace a $1100 telemetry system, at a lower power and smaller size. Two additional sources for heavy water (deuterium) are that Canadian Atomic Energy Commission and their Isotech reps here in the US. Price is in the forty cents per gram range. But do note that heavy water weights twice as much as regular water, or roughly two grams per cubic centimeter. A key cold fusion paper appears on pages 301-308 of Volume 261 of the Journal of Electroanalytical Chemistry. The EPRI Journal (the Electric Power Research Institute) for August 89 also has some interesting info. A few of you callers on our hacker helpline have suggested faking those high priced linear actuators that we looked at two columns back. To do this, select an ordinary stepper and couple it to a piece of threaded rod by using shrink tubing. Then add two nuts and some sort of a platform to make up a traveling slider. Various tricks could eliminate any backlash, and you should be able to replace a $57 list price actuator with a surplus $2 stepper this way. As per usual, all of our names and numbers have been gathered together into the ending appendix. We seem to have a mixed bag for this month
PostScript fractal fern Unusual cooling ideas Chaos science resources Thermoelectric modules Superintelligent machines
$20-30 range.
Thermoelectric Coolers
There sure appears to be a lot of hacker interest in the thermoelectric cooling devices, so let us have at it. Well start off with the bottom line Besides their very high costs, these devices are pretty much useless, since they are so incredibly and horribly inefficient. They are only workable for very specialized and arcane applications where nothing else will do. Thermo coolers will work only when you go to ridiculous extremes on insulation and heatsink efficiency, and then only are capable of pumping tiny amounts of heat. And, to further shatter some hacker dreams, no way will they make ice in the real world. At any rate, a thermoelectric module is a solid state cooling device. All you do is input a dc current, and heat moves through the TE device, cooling one side and heating the other. Figure one shows you a typical device. This ones the Cambion model 801-2007. A similar component is also now available through Melcor. This module measures about two inches square and can move up to 22 watts of heat when powered by a 12 volt car battery at 7 amps. Cost new is in the
Solid state cooling modules use the Peltier effect. Figure two shows you details. What you do is start off with bismuth telluride or some other semiconductor. Then you cut it up into blocks while heavily p doping some of them and n doping others. When a current is applied in the proper direction, the electron carriers in the n material and the hole carriers in the p material will "drag" thermal energy along with their motions. In both cases, the towed heat travels downward, creating a hot side and a cold side. The amount of heat you can move depends highly upon the temperature differential you have to maintain. As figure three shows us, this particular device can provide a 60 degree C difference between the hot and cold side provided you pump zero heat through it. On the other hand, if you do keep both sides at the same temperature, you can pump up to 22 watts of heat through the device. In most uses, of course, you will both want to cool something off and transfer heat, so you must use the middle of the curve. For instance, you could pump eight watts of heat with a 40 degree centigrade drop. To put this
Fig. 1 THERMOELECTRIC COOLING modules are fairly cheap and easy to apply, but they are so inefficient that they are pretty much useless for any higher-power heat pumping.
25.1
Hardware Hacker
0.001 inch over its entire area. It must then get carefully covered with a thin layer of thermal grease. Obviously, those portions of a module not contacting the heatsink can not deliver any useful cooling. The real killer to most TM module applications lies in the thermal drop between the heatsink and the ambient temperature. For higher power uses, a fan or water cooling is essential. For instance, a very large plain old heatsink will have at best a 1 degree C per watt temperature rise above the ambient. Say you want to pump the 8 watts at a 40 degree centigrade drop. The heatsink will have to sink these 8 watts, plus an additional 32 watts or so of the input power for a total of 40 watts. At 1 degree C per watt heatsink rise, the heatsink will warm up to 40 degrees C above your ambient room temperature. And youve done nothing but break even! The cooling module has a 40 degree drop across it. Your heatsink has a 40 degree rise across it. And the net cooling is zero. You have nothing at all to show for all your effort. One more time. In this previous example, your final cooling is zero, because the hot side of the heatsink rises just as much above ambient as the amount of temperature drop you are trying to achieve! And a a one degree per watt heatsink is a darn good one. And large, too. So, an extremely low thermal drop from heatsink to a room ambient is absolutely essential. Also essential is an incredible quantity of insulation around whatever you are cooling, for any net heat flux into your area being cooled will only make matters much worse than they already are. The modules can get reversed by reversing the current. Because of the inefficiency, the heating mode is five times more effective than the cooling mode. Watch this detail. One side effect of this inefficiency is that you must run the modules off of a nearly pure direct current. The least amount of power supply ripple will dramatically chop your cooling ability, since the ripple troughs will heat five times better than the ripple peaks will cool your load. A car battery is often your best choice to power your TE modules. You might temporarily demonstrate a
cold N P
(+) hot
(-)
Fig. 2 A THERMOELECTRIC MODULE consists of blocks of bismuth telluride that have been alternately n- and p-doped. The blocks are placed electrically in series and thermally in parallel. When a DC current is applied, electrons in the n material and holes in the p material will tow heat along with themselves. Heat is thus pumped from cold to hot.
in perspective, this module could, in theory, make a very small ice cube in around one hour. Assuming youve got ideal conditions. The trouble is that conditions are never ideal for TE modules. There are a number of nasty gotchas which will gang up on you when you try to use them in the real world. Thermoelectric modules are rather expensive, mostly because they have not been improved one iota in the last
two decades. Worse yet, all of these modules are horribly inefficient. In a typical use, youd have to input five watts of input power to provide just one watt of cooling. Mechanical air conditioners are almost fifty times more efficient than this! Worse yet, you have to apply this extra heat to exactly where you do not want it. The heatsink you connect the hot side of the module to must be absolutely flat, and be machined to within
20
10
0 0 20 40 60
Fig. 3 THE AMOUNT OF HEAT PUMPED by a thermoelectric module depends on the temperature drop you require. Note that the heat transferred at the maximum temperature differential is zero. Note also that the heat differential is with respect to the hot end of your heat sink, and not to ambient.
25.2
January, 1990
module simply by holding a module between your thumb and finger and then connecting up the module to a single aklaline "D" cell. Do not use the "D" cell for more than a very few seconds at a time. What good are TE modules? You definitely cant buy a 4 x 8 foot panel of these and use them to air condition your house. But they are lots of fun to play with. These are a sure-fire topic for a winning science fair project or student paper. Besides that, though, thermoelectrics are limited to those specialized and low load uses where nothing else can do the job. Important uses include coolers for infrared detectors, dew point humidity sensors, chillers for microscope stages, and for thermal management control in satellites. I have often wondered why these do not get used in satellite receiver front ends. I guess the answer is that the gain of the input amplifier drops faster with temperature than the noise figure improves. Both Cambion and Melcor do have lots of interesting tech literature on thermoelectrics, so you might want to contact them both for more info. across the room for a power source. At least the extra heat isnt generated at precisely the wrong place, as is the case with TE modules. It turns out that rapidly moving air accepts heat from slowly moving air. Thermodynamics and all. So, if you build a suitable vortex nozzle device having one inlet for shop air and two outlets, you can arrange things so that cold air comes out one end and hot air out the other. With zero moving parts. Believe it or not, you can get as cold as the -40 degrees Farenheit dry ice temperatures, or else you can trade off flow rates to get maximum heat pumping at lower differentials. One obvious use for these is as needle coolers for industrial sewing machines. Any heat-caused needle breakage can be almost completely eliminated by using a vortex nozzle. The leading manufacturer of vortex cooling nozzles seems to be Vortec. Besides offering bunches of free data and free cassette tapes on vortex cooling, they also have a number of other unsual nozzles that can do such things as air amplification. Another source is Exair. As a third approach to effective cooling, dry ice machines are amazingly low cost and simple. Edmund Scientific is one source. All you do is connect the machine to an also cheap carbon dioxide tank from your local gas supplier and turn it on. Slightly over half the gas expands and boils off, while the rest of it will cool and solidify into one solid block. Presto. Instant cold. Naturally, you do this only in a highly vented area while you are wearing safety glasses. You also will
Other Coolers
Theres three other approaches to unusual cooling techniques that are hacker hackable. You just might want to pick up on these instead. First, Japanese air conditioners are ridiculously better than the American ones, and have been for the past few years. It seems that all the American manufacturers refuse to pick up on this new technology, blaming the building codes, and claiming that a customer demand does not exist. These new air conditioners have EER efficiency ratings approaching a value of 15, meaning that they use 33 percent less electrical energy than ours do. Thats $60 per month or so. They are also much quieter and far more comfortable. Four of the secrets include a new beastie called a scroll compressor, a motor that provides a near infinitely variable speed on the compressor, a similar variable speed blower on the air handler, and a total programmed controller. I guess you would call this second device a solid state cooler, although you will need a large air compressor
/problistcreate {mark /counter 0 def probabilities {128 mul round cvi {transforms counter get} repeat /counter counter 1 add def} forall counttomark 128 sub neg dup 0 gt { [1 0 0 1 0 0] repeat} {pop} ifelse] /problist exch def} bind def /doit {problistcreate 1 1 20 {problist rand -24 bitshift get transform 2 copy moveto 0.001 10 rlineto} repeat newpath numdots {problist rand -24 bitshift get transform 2 copy moveto 0.001 0 rlineto stroke} repeat} bind def % /// demo - remove before use. /// /numdots 6000 def % increase for denser image; decrease to print faster /transforms [ [0.5 0 0 0.5 0 0] [0.5 0 0 0.5 1 0] [0.5 0 0 0.5 0.5 0.5] ] def /probabilities [ .33 .33 .34 ] def 1 setlinecap 0 setlinewidth 200 300 translate 100 dup scale doit showpage %quit
Fig. 4 A SERPENSKI TRIANGLE STRANGE ATTRACTOR done by using PostScript. While you can enter this into most any word processor on any computer, a printer that accepts PostScript is needed for your final output.
25.3
Hardware Hacker
want to prevent the instant frostbite caused by carelessly handling the finished product, or otherwise letting it come into skin or eye contact. memory bits or so. Thats equal to a dozen CD ROM platters or a small shoebox full of the 1 Meg x 8 SIMM strips. Its also equal to a mere eight of the new 64 Meg x 8 SIMMs just beginning to emerge from their preliminary lab research stage. Projecting these ongoing memory trends, in around ten years or so, machines will definitely be smarter than people. So, for this months contest, just write me a short essay on exactly what happens when machines will inevitably become the most intelligent life form on earth. Well have all the usual Incredible Secret Money Machine book prizes for the dozen or so best entries, with an all-expense-paid (FOB Thatcher, AZ) tinaja quest for two going to the top entry of all. As usual, all entries have to be on paper, and should get sent to me here at Synergetics, rather than to Radio-Electronics. Now, theres a memory effect that nobody talks about very much, and its implications for machine intelligence are genuinely scary. I like to call this the threshold effect. Your threshold effect goes like this: If you double the size of any memory, the performance capabilities of that memory usually will increase by much more than a factor of two. This happens because the arrangement and the use of your memory can now be better organized and can more effectively perform new higher level functions. For instance, some of your smallest microcontrollers only have room for a single hand-coded program. Add a little memory, and you can step up to a system monitor that gives you more power and control. More memory and you step up to some interpreted language. More still, and you pick up a sophisticated operating system. The next level gives you fancy toolboxes, and integrated programs. The level beyond that can give you powerful new levels of organizing the programs themselves. HyperCard and its mutants are an obvious example here. Beyond that lies all the expert systems, massively parallel architectures, and artificial intelligence. Because of these threshold effects, going from a 4 Megabit memory to a 4 gigabit one will gain you far more than a paltry 1000:1 in performance abilities. I feel that something like a million to one is more like it. It would seem to me that there are probably three or four brand new and previously unthunk of thresholds that lie between where we are today and where we will be when we get up to routinely being human brain sized a decade from now. What are these thresholds and when will they get crossed? Knowing the correct answers to this one could make you filthy rich in a hurry.
Another Contest
In case you have not noticed, the bottom has once again fallen out of the memory chip market, eliminating that insane two-year pricing blip that was a direct result of some incredibly stupid and monumentally misdirected government meddling in free trade. Your jellybean 1 Megabyte dynamic RAM chips can now be gotten in quantity for as little as $8 each. That translates to around 800 microcents per bit. And hackers can once again rejoice. Yea team. Now, depending on who is doing the counting, the human brain will contain the equivalent of four billion
/problistcreate {mark /counter 0 def probabilities {128 mul round cvi {transforms counter get} repeat /counter counter 1 add def} forall counttomark 128 sub neg dup 0 gt { [1 0 0 1 0 0] repeat} {pop} ifelse] /problist exch def} bind def /doit {problistcreate 1 1 20 {problist rand -24 bitshift get transform 2 copy moveto 0.001 10 rlineto} repeat newpath numdots {problist rand -24 bitshift get transform 2 copy moveto 0.001 0 rlineto stroke} repeat} bind def % /// demo - remove before use. /// /numdots 6000 def % increase for denser image; decrease to print faster /transforms [ [0 0 0 .16 0 0] [.2 .23 -.26 .22 0 1.6] [-.15 .26 .28 .24 0 .44] [.85 -.04 .04 .85 0 1.6] ] def /probabilities [ .01 .07 .07 .85 ] def 1 setlinecap 0 setlinewidth 200 300 translate 30 dup scale doit showpage %quit
Fig. 5 A FRACTAL FERN DONE IN POSTSCRIPT. Note that only 28 data values are needed to define this image. But whats really astounding is that you dont really "draw" this fern instead, you simply let it out!
25.4
January, 1990
probably can be called chaos science. Since chaos science is futuristic, utterly and totally bizarre, and more than eminently hackable, it should be of foremost interest to most of you Radio-Electronics hackers. Our Chaos Science Resources sidebar will give you a quick summary of several of the main places to go for further info on this exciting new field. Youll probably want to start off with James Gleicks book Chaos Making a New Science. This one is a thorough and easy-to-read but often shallow introduction. The horses whatever document, of course, is Mandlebrots Fractal Geometry of Nature. While more than pretty to look at, this text is pretty nigh unreadable and is laughingly and pompously egocentric. The book I am currently the most impressed with is the new Science of Fractal Images, especially in chapter five. A good summary did appear in the January 1988 Byte under the title A Better Way to Compress Images. One of the things that struck me about this great book and the Byte summary article was that the sample BASIC and C programs were forever making those complicated translaterotate-scale transformations. All the kind of stuff that PostScript does free, invisibly, and automatically all of the time. Contrary to all the dire negative comments found early in the red book, PostScript should end as up nearly ideal for any fractal geometry uses, along with nearly anything else involving chaos science.
FRACTAL AND CHAOS SCIENCE RESOURCES Beauty of Fractals, by Peitgen and Richter Earlier text that is both a coffee table book and a source of useful algorithms. From the same publisher that did the Science of Fractal Images. (Springer-Verlag, 1986)
Byte Magazine, One Phoenix Mill Lane, Peterborough, NH 03458 (603) 924-9281
Chaos Making of a new Science, by James Gleick Interesting and esaily read survey of the entire field of fractals and chaos science. Frustratingly shallow at times. (Viking, 1987)
Viking Penguin, 40 West 23rd Street, New York, NY 10010 (212) 337-5200
Fractal Geometry of Nature, by B.B. Mandlebrot The original "horses whatever" book that started it all. Mind blowing artwork in an otherwise unreadable and excessively egocentric text. (W.H. Freeman Press, 1982)
W.H Freeman, 4419 West 1980 South, Salt Lake City, UT 84104 (801) 973-4660
Science of Fractal Images, by M.F. Barnsley et. al. Outstanding collection of useful hands-on fractal information. A good starting place to extend your own research. Includes algorithms and a good bibliography (Springer-Verlag, 1988)
Scientific American, 415 Madison Avenue, New York, NY 10017 (212) 754-0550
SIGGRAPH Proceedings Far and away the best annual computer graphics show. Includes the latest and best of fractal and chaos science developments, on all three levels theoretical, tutorial, and artsy-craftsy. (Siggraph-ACM Press, annually)
Siggraph/ACM Press, 11 West 42nd Street, New York, NY 10036 (212) 869-7440
25.5
Hardware Hacker
Even though there still is an awful lot of slop left in this code, I was quite pleased with the final results and their speed under PostScript. I did elect to create a single custom probability table once ahead of time, rather than making slower individual probability selections on the fly. By providing for 128 entries in this table, rather than the expected 100 , does give us some further speedup and simplification. But a lot more cries to be done. When you are experimenting with these on your own, be sure to change the numdots variable to something in the high thousands for your original fast printing samples that will image in a few seconds to a minute or two. Later on, you could increase numdots as high as you need to for top quality final images that might take hours to print in their larger sizes. Ultimately, you can convert the fern into a super fast printing image. Full details in #336 FASTFERN.PS on my GEnie PSRT. One other gotcha: The six element matrices used by the original fractal people and PostScript differ slightly. To get from one to the other, simply interchange your second and third numbers in the array. Never mind why. There are some amazingly bizarre things about the fern image of figure three. First and foremost, larger sizes of this fern original are hauntingly beautiful. In fact, Ive never seen any PostScript code image anywhere that comes even remotely close. I do hope our Radio-Electronics printing process left at least a hint of this for you. Second, the smaller pieces of the fern are definitely not a smaller, or selfsimilar copy of your whole picture. Unlike ordinary pictures, as you magnify, additional and uniquely different detail appears. But that is what fractal stuff is all about. Third, although an apparently random process is used that picks four tasks by using random probabilities in random orders, you always end up with the same final image! This will happen regardless of which random numbers get used in which order. This is a stunning example of a strange atttractor, otherwise known as the holy grail of fractaldom. As you might expect, others are seeking out those possible strange attractors thought to underly the stock market prices and weather patterns. Despite this, a sick random number generator will definitely not perform. You do have to use a good one. Short repetition sequences appear to be a no-no. Many more details on this in my Apple Assembly Cookbook. Fourth, believe it or not, the entire fern is coded as only 28 numbers in a tiny data array! Theres no difference whatsoever between your Serpenski triangle and the fern, except for these 28 numbers. Nor is there any other difference between these and untold billions of other wildly differing images. So, we seem to have an absolutely incredible picture compaction scheme going on here. For a full page PostScript fern, your data compaction could approach 300,000:1 ! Unfortunately, though, your image regeneration times do get out of hand with compaction levels that are this high. You can forget about real time. In theory, by going to just a few more numbers, we can create any image at all. Which gets heavy fast. Fifth, and finally, this image really does look just like an authentic Black Spleenwort Fern. Now, Black Speenwort Ferns have been around for a rather long time, and they sure are pretty, but nobody ever accused them of being very bright. Could the exact same process be used by the real fern to teach itself how to grow? To me, the odds are overwhelming that this does seem to be the case. We thus appear to be tampering with some heavy duty stuff here. What really astounds me over this fern, though, is that you do not really draw it. All you do is let it out, and it leaps out of its cage with a vengence. I guess that is where the "strange" in strange attractors comes from in the first place. Oh yes. If you do find the strange attractor that underlies stock market prices and send it to me, Ill be most happy to send you five of my money machine books. I might even consider a free tinaja quest as well. metric gears, especially the Module 1.0 size. Any suggestions? Yes, Stock Drive Products does carry them, but at five times what I want to pay for them. A free book for your trouble.
Help Wanted
How about you helping me for a change? Im having trouble finding some sources of cheap stock plastic
25.6
Don Lancasters
Hardware Hacker
February, 1990
Dew sensing circuits Direct toner pc update Midnight Engineering Hacker cold fusion kits Alarm & security resources
dozen earlier papers, gotten through the
INSPEC service within Dialog.
heres lots of exciting news this month, including cold fusion kits, an update on the direct toner printed circuit process, a major new hacker magazine, and some details on unusual dew sensor parts. So, lets just jump right on in
hundred key cold fusion papers should now be available by way of the Dialog In- formation Service that you will find through your local library. Figure one should get you started. Ive listed two
The costs of generating your own complete and up-to-date cold fusion abstract listing should be around $35,
Cold fusion in metals J. Kondo, Journal of the Physical Society of Japan, June 89, Vol 58 #56, pp 1869-1870. Deuterium nuclear fusion in metals at room temperature P. Tomas, Fizika (Yugoslavie), vol 21 #2, 1989, pp 209-214. Exact upper bound on barrier penetration probabilities for cold fusion A. Lettitt, Physics Review Letters, (USA), Vol 63 #2, 10 July 89, pp 191-194. Cold fusion: how close can deuterium atoms come inside palladium? Z. Sun. Physics Review Letters (USA), Vol 63 #1, 3 July 89, pp 59-61 Search for fusion reactions between deuterium atoms inplanted into titanium R. Behrisch, Nuclear Fusion (Australia), Vol 29 #7, July 89, pp 1187-1190. Can solid-state effects enhance the cold-fusion rate? A. Leggett, Nature, (UK), Vol 340 #6228, 6 July 89, pp 45-46. Upper limits on neutron and gamma ray emission from cold fusion M Gai, Nature (UK), Vol 340 #6228 6 July 89, pp 29-34. Calculated fusion rates in isotopic hydrogen molecules S. Koonin, Nature (UK), Vol 339 #6227, 29 June 89, pp 690-691. Search for neutrons during heavy water electrolysis on palladium electrodes S. Blagus, Physics of the atomic neculi (Germany), Vol 333 #3, 1989, pp 321-322. Screening corrections in cold deuterium fusion rates K. Langanke, Physics of the atomic neculi (Germany), Vol 333 #3, 1989, pp 317-318. Cold fusion in the solid state? J. Goedkoop, Eneergiespectrum (Netherlands), Vol 13 #6, June 89, pp 156-162. Nuclear fusion P. Bullo, Elettrificazione (Italy), #5, May 89, pp 57-61. Theoretical considerations on the cold nuclear fusion in condensed matter F. Parmigiani, Nuovo Cimento D (Italy), Vol 11D ser. 1 #6, June 89, pp 913-919. Chemical forces associated with deuterium confinement in palladium J. Mintmire, Physics Letters A (Netherlands), Vol 138 #1-2, 12 June 89, pp 51-54. Cold fusion in condensed matter: is a theoretical description possible? W. Schommers, Modern Physics Letters B (Singapore), Vol 3 #8, 20 May 89, pp 597-604. Doubts grow as many attempts at cold fusion fail B. Levi, Physics Today (USA), Vol 42 #6, June 89, pp 17-19. Muon catalyzed fusion C. Petitjean, Fusion Engineering Design (Netherlands), Vol 11 #1-2, June 89, pp 255-264. Computation of the cold fusion rate P. Dong, New Physics (South Korea), Vol 29 #2, April 89, pp 233-234. The observation of 2.2 MeV gamma-ray in electrochemical cell Y. Park, New Physics (South Korea), Vol 29 #2, April 89, pp 231-232. First steps toward an understanding of "cold" nuclear fusion T. Bressani, Nuovo Cimento A (Italy), Vol 101A ser. 2 #5, May 89, pp 845-849. Emission of neutrons as a consequence of titanium-deuterium interaction A. De Ninno, Nuovo Cimento A (Italy), Vol 101A ser. 2 #5, May 89, pp 841-844. Fermi gas like hypothesis for Freischmann-Pons experiment O. Rossler, Z. Nat. Phys. Chem. Kosmophys (W Germany), Vol 44A #4, April 89, pp 329. Observation of cold nuclear fusion in condensed matter S. Jones, Nature (UK), Vol 338 #6218, 27 April 89, pp 737-740. Electrochemically induced nuclear fusion of deuterium M. Fleischmann, Jour. Electroan. Chem. (Switzerland), Vol 261 #2A, 10 April 89, pp 301. Cold nuclear fusion A. Rusu, Energetica (Romania), Vol 3 #6, June 88, pp 258-261. On the practical use of the cold fusion (problems and prospects) Y. Petrov, Muon Catalyzed Fusion (Switzerland), Vol 3 #1-4, 1988, pp 525-535.
Fig. 1 SOME KEY COLD FUSION PAPERS. For a complete and up to date list, use the Diaglog Information Service at your local library. Start with their INSPEC resource data base. Fleischmanns paper started it all.
26.1
Hardware Hacker
ten or so. Thus, your estimated odds are only one in one hundred that you will be able to prove your cell actually works. So, what we have here appears to be a crap shoot. On the other hand, here you do have a sure fire winner for a science fair entry, show-and-tell, or school report. And the "touchy-feely" and "See what Ive got!" aspects of all these kits are completely off scale. Golly gee, Mr. Science. Guy also offers fancier cold fusion kits and similar products for serious researchers. Contact him directly for more information. Needless to say, be sure and let us know the outcome of all ongoing test experiments. This is a wide open field with mind-boggling hacker potential. UPDATE: As of fall 1992, cold fusion has largely failed to prove itself. No solid atomic evidence has ever been demonstrated, and the phenonomen has pretty much been downgraded to a lab curiosity. But dedicated researchers do continue their work.
Fig. 2 LOW COST HACKER FUSION KITS consist of a small test tube filled with heavy water, a short palladium rod cathode, and a nickel wire cathode. While the odds of positive results are rather low, this is a great show-and-tell item.
and should take twenty minutes. As a reminder, the best sources for current info on cold fusion are in the News and Comments section found in Science magazine, and the Technology section that is usually found on page B-4 of your Wall Street Journal. On to those kits. Guy Wicker is a name-brand hardware hacker and a well-known energy researcher. By a special arrangement, Guy has offered to put together a few new cold fusion mini-kits for all you Radio-Electronics hackers. The kits cost only $25, plus $2 shipping and handling. As figure two shows you, this kit consists of a small test tube full of a 0.1 molar deuterium and lithium hydroxide solution, a short piece of 50 mil palladium rod, and a small loop of
()
(+)
Fig. 3 A TYPICAL COLD FUSION test setup. Apply a dc current aound 10 mils. The anode and cathode should not contact. You measure for excess temperature, neutrons, and tritium.
nickel wire. Nickel is used instead of platinum for the anode. A classic cold fusion cell is built up something like figure three. Use a coiled nickel wire anode, and a nickel suspended palladium rod cathode. You apply a test DC current in the 4 to 20 milliampere range, with positive to the nickel anode and negative to your palladium cathode. Be sure to carefully monitor your temperatures at several locations. If possible, also monitor for radiation. A "charging" time of several hours or a few days seems normal for this size palladium rod. Look for both a very low level and a separate "burst mode" excessive heat production effects. Some sort of a continuous temperature recorder, possibly based on a personal computer, is probably a very good idea. Youll want to work inside of a styrofoam block which sits inside a picnic cooler or other insulated box. Guy reports that very small cells like this one are far more likely to produce a low level tritium radiation than excess heat production. Getting both to happen at once when someone is watching is usually frustratingly difficult. It is also the Holy Grail of cold fusion. Some warnings here. Your cell is extremely small and the excess heat production, if any, is likely to end up correspondingly tiny. Opportunities of cell contamination are very great. Your odds of your cold fusion cell working at all are only something like one in ten, and those odds of your being able to observe that operation when and as it occurs is also estimated as around one in
Midnight Engineering
Theres a brand new magazine out known as Midnight Engineering. It is specifically aimed at you hardware hackers and software developers who are trying to market their high tech products on a small-scale or startup basis. Its shoot-from-the-hip style is very much cast in the same mold as my Incredible Secret Money Machine. Emphasis is on real-world solutions and new hacker opportunities. Free sample copies are available to you. This one is a "must have."
26.2
February, 1990
work at 1X , 2X , 4X , frontwards or backwards, normal or reversed, to as many as eight layers at once; has no upper size or any complexity limit. It is also totally host independent, which allows you to use any old word processor on any old personal computer; simply lets you pass layouts over any BBS system in the world as a plain old textfile; and is compatible with virtually all of the existing CAD/CAM programs, either through a direct PostScript driver or by a PostScript HPGL emulator. Similarly, using the heated roller method works far better than an iron. The obvious reasons for this are the far more uniform pressure and temperature. You also can gain bunches in dimensional stability, since the mylar transfer sheet does not get any chance to distort all at once. As weve seen I do recommend a fake Kroy Kolor machine to do the transfer. Well look at some cheaper alternatives for this in a future issue. Several readers have now suggested substituting Kapton film instead of the polyester overhead sheets since it has better high temp stability. Two of the sources for Kapton include DuPont and Rogers Corp. It also seems a very good idea to do your final cleaning wash of your bare copper board with distilled water. A brief pre-etch before transferring your toner image does appear to be quite important as well. The etch will both guarantee an extra clean board and gives you just the right "tooth" for the toner to grab onto. Some other suggestions: preheat the copper board as much as you can before doing the transfer. The idea is to keep the copper from acting as a giant heat sink. Something around 150 degrees Farenheit should work. Note that gloves are an absolute must. And rub an ice cube on the back of the transfer sheet before seperating it from the pc board. This seems to improve the transfer process bunches. Be sure and send in your own tips and techniques for geting perfect 1:1 toner transfers. When all of the dust settles, well try to separate the black magic from the reality and standardize on a useful direct toner process. UPDATE: Vastly improved direct toner products appeared in fall of 1992. Especially from DynaArt Designs and Technics Inc..
Jensen Tools and TechniTool. Sadly, both of these yuppieized outfits are rather pricey. But all of their products are both first rate and top quality. There are now hundreds of security consultants that will be glad to help you for a sane and reasonable fee. Two examples here are Wyatt Palmer (no relation) of Valley Security, and Jeff Lancaster (my widdle brudder) of his Cain Security Systems. Please let me know if you know of any other resources you think should be added to this list.
Dew Sensors
Weve looked at humidity sensors in several earlier columns. One unique and low cost humidity detector is called a dew sensor. These cost only a dollar each and are intended to protect moisture-sensitive VCR drums. Three typical part numbers would be the Murata HOS 101-100 , the Taiyo Yuden TD-P-100 , or that Panasonic EYH-S0210. Unlike hygrometer sensors, these devices should have a fairly low and uniform dc resistance whenever no moisture condensation takes place. A typical dry resistance value is 10K. At a 100 percent humidity, the air retains all of the moisture it possibly can. Beyond this, the moisture in the air will condense out, forming rain or a dew coating. A special moisture sensing paint
1. The PostScript language is essential to the direct toner process. Its totally overwhelming advantages include precise scaling to 1:1, total format flexibility, host and printer independence, easy BBS downloads, and the ability to use nothing but a word processor. 2. Heated roller transfer methods work much better than an iron. Use a fake Kroy Kolor machine or its equivalent. The reasons include better stability and tighter temperature control. 3. Use a "dry" (silicon oil free) fuser wiper pad for several copies before your transfer sheet is run. 4. For extreme dimensional stability, try using a Kapton film, rather than polyester or mylar. 5. Do your final washing of your cleaned copper board with distilled water. 6. A brief etch before your final washing step is essential to give the proper "tooth" for toner transfer. 7. Preheat the copper to 150 degrees or so before the transfer to minimize any heat sinking effects. 8. Rub an ice cube over the back of the transfer sheet before seperating it from the pc board.
Fig. 4 HERES AN UPDATE on the direct toner printed circuit prototypes we first looked at in the December 89 issue.
26.3
Hardware Hacker
dozen or so entries, with an all expense paid (FOB Thatcher, AZ) tinaja quest for two for the best of all. As usual, all entries must be written and should be sent directly to me here at Synergetics, rather than on over to Radio-Electronics editorial. Show me what you can come up with on this.
ALARM & SECURITY RESOURCES Ademco 165 Eileen Way Syosset, NY 11791 (800) 645-7492 Alarm Installer & Dealer 2512 Artesia Boulevard Redondo Beach, CA 90278 (213) 376-8788 ARIUS 959 Concord Street Framingham, MA 01701 (800) 241-3278 Cain Security Systems 2593 Wexbayne Rd, S301 Sewickley, PA 15143 (412) 934-1000 Jensen Tools 7815 South 46th Street Phoenix, AZ 85044 (602) 968-6231 King Alarm 145 English Street Hackensack, NJ 07601 (800) 526-0162 Security Dealer 210 Crossways Park Drive Woodbury, NY 11797 (516) 496-8000 Security Distributing 1350 E Touhy Avenue Des Plaines, IL 60017 (312) 390-2117 Valley Security PO Drawer J Thatcher, AZ 85552 (602) 428-2142
dramatically raises its resistance well beyond 1 Megohm whenever it gets wet. Thus, your sensor is typically at 10K when dry and 1 Meg when wet. A good decision point is to split up the difference on a log basis, and do your tripping at a 100K level. Figure five shows you one possible sensing circuit. Because you have to keep the total system power under two milliwatts, the maximum dc voltage you are allowed to apply is only 0.8 volts. The sensor acts as a current robber. At low humidities, the sensors resistance is around 10K and steals so much current, that the NPN transistor remains off. When wet, the sensor impedance goes above a megohm, and conventional current flows into the base of the NPN transistor, turning it on and dropping its output voltage. The base-emitter junction of the transistor guarantees that the sensor voltage should never exceed 0.6 volts. This is an example of a current steering mode of operation.
That inverting CMOS Schmidt trigger gives you an on-off snap action. Your red light emitting diode lights when wet and the green when dry. You do have to be fairly careful with these. Fingerprints, solvents, or even ions from your city water can damage the sensor, as can organic chemicals from certain mounting adhesives. Solder flux and flux removers are also a no-no. Once again, your highest permitted applied voltage is only 0.8. Any selfheating is not allowed. There are a number of other dew sensors. Some of these work only under ac drive, and others may have different impedance levels. Be sure to have the data sheet for your actual device on hand if you try to use any of these alternates. Tellyawhat. For this months contest, just tell me about a new or unique use for dew sensors. There will be all of those usual Incredible Secret Money Machine book prizes going to the best
750K
100K 74HC14
Fig. 5 A DEW SENSOR TEST CIRCUIT. The red LED lights when the humidity exceeds 100 percent; the green one turns on in the absence of dew.
26.4
February, 1990
26.5
January, 1990
25.7
Don Lancasters
Hardware Hacker
March, 1990
LVDT position detectors Treasure finding circuitry Switchable analog inverter Synchronous demodulators Desktop accessories contest
doubly balanced modulator, a phasesensitive detector, I-Q demodulator, synchronous rectifier, a phase-locker, or, going back to some really ancient history, as a homodyne detector. Figure one can also be used as a modern and handy sync demod circuit. What is especially nice is that this is a single-ended circuit which needs no transformers of any kind. In general, any sync demod is an electronic multiplier that extracts the sum and the difference between your input frequency signal and your reference switching frequency. For instance, figure two shows you what happens if we keep the input and switching frequencies identical, but shift their phases. Lets first assume a zero phase difference between signal and reference. The amplifiers gain will be (+1) for the positive signal peaks, and (-1) for the negative signal peaks, giving us a plain old full wave rectifier. The "dc term" here will be a direct current level equal to the strength or amplitude of the input signal. The "lumps" will be a double-frequency and higher "ac term" that is usually filtered out. Thus, the zero phased synchronous demodulation should extract only the amplitude of an input signal. Suppose we next shift the phase by 180 degrees. This time, your gain is (-1) for the positive lumps and (+1) for
his month, I figured wed look at a rather mis-named integrated circuit that has an unbelievable future hacker potential. But first, as usual, lets pick up some background
quarter of a CMOS 4066 is often an ideal choice for this sort of thing. As we will shortly see, there are zillions of uses for this circuit. One obvious place is as a video inverter, used to create negative video images for special needs. To create a video inverter, you take this circuit and add a sync separator, such as the National LM1881 , and set things up to invert only your video portions but not the sync portions. Naturally, a videoquality op-amp would be needed for this use. Additional bias current can be switched in as needed to get the correct signal levels. But, well save that for some other time, because what I really want to get into here are the secrets behind
Synchronous Demodulation
The switch in figure one could be turned off and on at very slow rates compared to your input signal frequency. Or at equal rates. Or even at much higher rates. Now, if you turn the switch on and off at precisely the same speed as your input signal, you create a beastie known as a synchronous demodulator. Syncronous demodulators form a super-important electronic concept used just about everywhere. A sync demods name can change all over the place, being otherwise known as an autocorrellator, a lock-in amplifier, a
10K 10K
Input
5K
5K
Output op-amp
Fig. 1 A PROGRAMMABLE GAIN AMPLIFIER. When the switch is closed, the gain is -1. When the switch is opened, the gain is +1. One important circuit use is for full wave synchronous demodulation.
27.1
Hardware Hacker
complex signal at the same time. Lets look at some additional uses for synchronous demodulation. If you hard limit the input signal to your sync demod so it is a clean square wave, then you will provide phase detection. Go through the math, and you will find that your output voltage is a triangle which linearly equals the phase angle between the input and the reference. You can then use your sync demod to measure the phase angle between two input signals. Suppose we would synchronously demodulate one signal and at the same time apply an interfering signal of a different frequency. Your "wanted" signal will always be in phase and will always produce a dc output equal to its amplitude. The "unwanted" signal will create a sinewave as it "slips cycles" with respect to the wanted one. Any sinewave averages to zero over a long enough time. We apparently have a way of detecting one signal while we reject an interfering one. Certain types of sync demods are called lock-in amplifiers. These can form an extremely narrow bandpass filter that automatically centers on the frequency of interest, and can yield tremendous improvements in signal to noise ratios. As an example, it is not unknown for your lock-in amplifier to cleanly extract a signal that is buried in noise that can be as much as 120 decibels stronger than the signal you want. Important uses here include extracting data from deep space probes, and performing laboratory and medical instrumentation that measures very weak signals in the presence of much larger interfering noise and ac hum. A not-quite synchronous demodulation will output a sinewave equal to the difference between the reference and input frequencies. This is a simple example of a downconverting, or mixing of one signal against another to get a frequency difference. By using a pair of sync demods, you can not only extract the frequency difference, but also determine which of the frequencies are higher. The doppler signal extraction in a side-looking radar is one major use. Separately, synchronously demodulating a shaft encoder can give you both speed and direction information. Flangers for electronic music synth-
INPUT
REFERENCE
When the reference and the signal are in phase (0 degrees), a strong dc term is produced proportional to the strength of the input signal. The second harmonic and higher ac frequencies are usually strongly filtered out, leaving only a dc term.
OUTPUT
INPUT
REFERENCE
When the reference and the signal are in quadrature (90 degrees), a zero dc term is produced, totally rejecting the input signal. Values very near a quadrature or 90 degree phase are often used for phase measurements or for FM demodulation. When the reference and the signal are out of phase by 180 degrees, a negative dc term is produced that is proportional to the input signal. Note that the circuit of figure one is 180 degree phased, unless you purposely invert the reference input signal.
OUTPUT
INPUT
REFERENCE
OUTPUT
INPUT
REFERENCE
When the reference and the signal are at -90 or +270 degrees, a zero dc term is produced, once again totally rejecting the input signal. Because of quadrature rejection, twice as much information can be put onto a synchronously demodulated carrier.
OUTPUT
Fig. 2 SYNCHRONOUS DEMODULATOR WAVEFORMS. At 0 degrees reference phasing, the "dc" portion of the output is proportional to the input amplitude, and interfering signals are strongly rejected. At 90 degrees, the "dc" portion is zero, and the input signal itself gets strongly rejected.
the negative lumps of the signal. We get a negative output, and we could conclude that any 180 degree phased synchronous demodulation extracts only the negative of the amplitude of the input signal. Now, lets get interesting. Suppose we shift the phase to 90 degrees. What happens? Well, nothing at all. During the time the switch is closed, we have half of a positive cycle and half of a negative cycle, so our net (or average) dc output is zero. Very handily, any synchronous demodulation at a 90 degree phasing produces a zero output. And, since nothing upside down is still nothing, a similar cancellation happens at -90 or +270 degrees. But wait a minute. If we get zero output at a 90 degree phase, cant we
double the information placed on a carrier, simply by having an in-phase term and a 90 degree, or quadrature term? We sure can, and it gets done all the time. For instance, the color information on an NTSC (Never The Same Color) television is placed onto a magic subcarrier of 3.57545 Megahertz. The hue of the color is the phase angle, while the saturation of the color sets your amplitude. At the receiver, a phaselock loop does an in-phase "I" and a quadrature "Q" demodulation. After further processing, all the separate amplitudes are extracted for the red, blue, and green guns in the display tube. Thus, we have used a pair of synchronous demodulators to extract both the amplitude and phase of a
27.2
March, 1990
isizers use a similar principle. For a really strange use of a sync demod, high power rectifiers always have a forward drop associated with them of a volt or more that can cause all sorts of inefficiencies, especially with low voltage solar panel power conversions. But power transistors can have significantly lower forward drops if they are properly overdriven as a switch. To improve the efficiency of a solar power conversion system, you create a syncrhonous rectifier that uses a sync demod and a pair of power transistors to replace your inefficient power diodes normally used. Finally, one early way of detecting an am radio station was known as a homodyne detection. Here you would synchronously demodulate your incoming signal against a reference of the same frequency as the transmitter carrier and then directly extract your audio in a single step. Unfortunately, these howled a lot as they were tuned and tended towards instability, so they were flushed in the late 1920s in favor of superhetrodyne circuits that used intermediate frequency amplification, often at 455 Kilohertz. Today, the old homodyne deserves a fresh new look, especially when combined with digital synthesis, and the ability to put highly stable audio gain and good filtering into very small packages. So, sync demods are easy to build and can be amazingly versatile. But, before we explore a great sync demod chip and more real-world uses, lets briefly look at some little-used hacker components known as
secondaries
secondaries
primary
primary
(A) SCHEMATIC
(C) RESPONSE
Fig. 3 LINEAR VARIABLE DIFFERENTIAL TRANSFORMERS, short for LVDTs, can be used for extremely precise position-to-voltage transducers. The output voltage follows the core position. Microweighing is one use.
LVDTs
An LVDT, or a Linear Voltage Differential Transformer is shown in figure three. This is a transformer which has a moveable core, a single winding on the primary, and a pair of secondary windings. The secondary windings are connected in opposition, so that the output voltage will be the difference between the two. When the moveable core is in the center, equal and opposite voltages are induced in the secondaries, and the output voltage is zero. As your core moves up, a 0 degree phased sinewave appears in your output. As it moves down, a 180 degree phased sinewave appears instead. With careful design, you can get a
linear sinewave output voltage whose amplitude changes with position, and whose phase is (0) degrees for positions above center and (180) degrees for positions below center. In short, a LVDT is a very precise and ultra-sensitive position to voltage transducer. Some LVDT devices can easily sense any motions or position changes as small as a thousandth of an inch or less. They can also be made large enough to measure distances of several feet or more. Unfortunately for hackers, LVDTs are rather pricey, since they are both low-volume and precision components. One useful surplus source is AST Servo, while others advertise in the Sensors and Measurement and Control trade journals. Several LVDT uses? Weigh scales, especially for microweighing. Torque sensing. Accelerometers. Distance measurement. Inclinometers. Pressure transducers. For seismometry. Load cells. Micropositioners. Or anywhere else you want to convert a very small motion or distance change into a useful electrical signal. For precision results, your LVDT must get driven from a pure audio sinewave of a fixed and known amplitude. Distortion could lead to bad harmonics which will in turn create output errors and other difficulties. To further up the LVDT precision, you can use a LVDT in its servo mode. Here, youd use feedback to move,
balance, shove, or otherwise continually coerce the LVDT back to its null position. This is known as null seeking, and, because of the feedback, many non-linearities can be greatly reduced if not cancelled outright. While LVDT s are usually superprecise, theres no reason you cant throw one together on your own to create a low cost position sensor. Be sure to let me know what you do come up with on this. By another of those astounding coincidences that infest this column, it looks like we need a sync demod to extract the position info from a LVDT. But, if we have a simple and cheap circuit that does this, why limit it to LVDT uses, when so much more can be done so much better with it? Which brings us around to
27.3
Hardware Hacker
head of an induction balance metal locator is really a LVDT in disguise. In the absence of a buried object, the voltages induced into the output sensing coils are equal and opposite. A buried ore or a metal object will distort the transmitted field, and unbalance the output voltages. Now for the neat part. Any "metal" objects return an in-phase component to the output signal, while "mineral" deposits, such as a well rusted can, return us a quadrature signal. Nicely separating the goodies from all the grunge and garbage. Thus, the NE5520N circuit can be used for in-phase discrimination of metal objects, or for quadrature discrimination of mineral objects. To do the quadrature synchronous demodulation, just shift the phase of your reference by 90 degrees. Add a second NE5520N , and you can also add such advanced features as automatic ground tracking and the "native" soil background cancellation effects featured on higher priced locators. Similarly, over in those receivertransmitter styles of metal locators often chosen for pipe finding, fiber optics can be used to optically couple from the transmitter to the receiver, minimizing any field distortions an actual cable might create. A sync demod at the receiver could then be used for improved sensitivity and for metal/mineral discrimination. The NE5520N costs around $7 in smaller quantities. Supply current is around seven milliamperes, easily provided by a nine volt battery. While you can run the chip at +5 volts, its stability will not be as good. For additional NE5520N circuit details and bunches of applications info, see the Industrial Linear Data Manual II offered by Philips. Check out the newer NE5521N as well.
+9 Vdc
0.1 F
Vcc
Vref
5.1K 0.1F
1K
2
out
+ -
0.47F
3
sync demod
LVDT
Fig. 4 THE PHILLIPS/SIGNETICS NE5520N has an internal precision sinewave generator, a sync demod, and an extra op-amp that gets used here as an output filter. This is how to apply it as an LVDT signal conditioner. Since this chip can do so much more, it does seem a shame to waste it on LVDT uses.
uncommitted op-amp you can use for output filtering, meter driving, or inphase to quadrature conversions. The circuit shows you how to power and sense the output of a LVDT. While you actually could use one of these chips with a LVDT , the beast should work well for an extremely wide variety of hacker stuff. Where else could we use an audio source and
a sync demod? One place could include modulated infrared alarms and communicators that can ignore both sunlight and room lighting. A second might be in the fluxgate magnetometer used in solid state digital compasses. How about treasure finding? I can think of at least a dozen uses here. Figure five shows us how the search
`
BURIED PIPE
secondaries
primary
Printing Resources
As you know, each month I try to feature a Resources Sidebar which shows you where to go for the real insider stuff on unusual subjects of hardware hacking interest. Sources that are difficult or impossible to quickly pin down on your own. Be sure to tell me what you want to see in future sidebars. This months sidebar gives you a rundown of the major sources of
Fig. 5 THE SENSING HEAD of an induction-balance treasure finder or a metal locator can be thought of as an LVDT in disguise. The target acts as a moveable core. Synchronous demodulation can separate metal from mineral detection.
27.4
March, 1990
information on printing and printshops. What does this have to do with hardware hacking? Just this Theres a total desktop publishing revolution going on out there, and traditional printing equipment and machinery is so utterly and outrageously priced that it simply wont hack it. What we need instead is for all you hackers to come up with low cost and low-end, do-it-yourself hardware kit solutions for new desktop printing and book-on-demand publishing needs. To do this, hacking skills and a hacker mentality are essential. And the opportunities are pretty nigh unlimited. Whats needed? Well, for openers, heres a tiny part of my wish list
(1) A low cost Kroy Kolor machine that doubles as a laminator and as a printed circuit direct-toner transfer processor. (2) A true perfect binder that gives fully professional cold-glue results for under $99. (3) A sanely priced clamping paper cutter which can accurately trim 200 sheets at a time. (4) A die punch that in one whomp cuts out twelve business cards from a single sheet of cover stock. (5) A low cost padding press. (6) An economical and programmable folder. (7) A simple-to-use pad printer for the "real" printing of pens, golf balls, mugs, keyholders, and such. (8) A cheap corner rounder. (9) A workable paper drill. (10) Automated conversions from a laser printed image onto T-Shirts, for rubber stamps, and for CAD/CAM machining of a plastic mockup or even actual hardware. (11) A super scungy vacuum packer and/or shrink wrapper. (12) A method to convert an orbital jitter sander into a paper jogger. (13) A new combination scoring, perforating, and die-cutting machine, possibly with a hot foil die-stamping capability as well. (14) A very low cost power stapler strong enough to handle Jiffy Bags. (15) Solutions to economical custom glass etching. (16) Vinyl hot-knife sign cutters and PostScript driven wood and/or aluminum routers.
PRINTING RESOURCES American Printer 29 N Wacker Drive Chicago, IL 60606 (312) 726-2802 Font & Function PO Box 7900 Mountain View, CA 94039 (800) 833-6687 Graphic Arts Monthly 249 W 17th Street New York City, NY 10011 (212) 463-6834 Graphic Arts Prod. News 29 North Wacker Drive Chicago, IL 60606 (312) 726-2802 High Volume Printing Box 368 Northbrook IL, 60065 (312) 564-5940 In-Plant Printer Box 368 Northbrook, IL 60065 (312) 564-5940 In Plant Reproductions 401 North Broad Street Philadelphia, PA 19108 (215) 238-5300 Instant Printer Box 368 Northbrook, IL 60065 (312) 564-5940 Print Equipment News Box 5540 Glendale, CA 91221 (818) 954-9495 The Printers Shopper PO Drawer 1056 Chula Vista, CA 92012 (800) 854-2911 Printing Impressions 401 North Broad Street Philadelphia, PA 19108 (215) 238-5300 Printing Journal Box 91447 Pasadena, CA 91109 (818) 793-7901 Printing News 245 W 17th Street New York City, NY 10011 (212) 463-6727 Publishing Technology 401 North Broad Street Philadelphia, PA 19108 (215) 238-5300 Quick Printing 1680 SW Bayshore Blvd Port St Lucie, FL 34984 (407) 879-6666 Screen Printing 407 Gilbert Avenue Cincinatti, OH 45202 (513) 421-2050 Type World 15 Oakridge Circle Wilmington, MA 01887 (617) 658-6876 U&lc 2 Dag Hammarskjold Plaza New York, NY 10017 (212) 371-0699
cost to the end user on any and all of these should end up in the $25 to $99 range, or under one-tenth of the sales commissions of the "real" print machinery that all this stuff is inevitably going to shoot out of the saddle. Your hacker abilities are obviously needed to hold all the costs down. And lots of electronics and power control does seem to be involved. This is one hot topic right now, with a potential market in the millions of units.
Two Contests
Let us have two contests for this month. As usual, there will be an Incredible Secret Money Machine for the dozen or so best entries, with an all-expense paid (FOB Thatcher, AZ) tinaja quest for two awarded to the very best of all. Either (A) show to me a new and unique hacker circuit involving the NE5520N with some unusual synchronous demodulation use, or else (B) show me a low-cost solution to some desktop publishing peripheral hardware of one sort or another.
wheelbarrow full of fat, new, and free data books. Check out their Light Emitting Diode; RF Power Transistor; High-Voltage Transistor; RF Power Module; Thyristor; Small Signal Transistor; Surface Mounted Semiconductor; and the PowerMOS data books for openers. Theres bunches more where these came from. Linear Technology has issued a great new 1990 Linear Data Book that is crammed full of great hacker integrated circuits. Free software this month includes the latest Specs in Secs Discrete Data from Motorola, and the PCBII printed circuit layout demo from OrCAD. And unusual free samples for this month include the MC33034 Brushless Motor Controller, again from Motorola, the Advanced Bus Interface Product Samples from Texas Instruments, and the various low cost jellybean integrated circuits from Calogic. Our reminder here that Midnight Engineering is a great new hacker magazine aimed at all small-scale hardware and software productions. Free samples are available. Rounding out the selections, the Chomerics folks have a new Cho-Flex 440-X line of force sensitive inks. Lets hear from you3
27.5
Hardware Hacker
27.6
Don Lancasters
Hardware Hacker
April, 1990
Organic Vapor Detectors Low Melting Point Alloys Ultraviolet Flame Detector Alternate Energy Resources Five Band Equalizer Circuit
down the drain if (A) you are on a city sewer system, (B) you dilute it with great heaping bunches of very cold water, and (C) you are doing this with less than two quarts of etchant each week. Do note that your ammonium persulfate is intended for dissolving copper and your sink drain trap is usually made from copper or a copper alloy. Hot and concentrated etchant here is obviously a no-no.
e appear to have a really mixed bag of outstanding goodies for this month, so lets just jump right in. As usual, please do note that all of our referenced sources appear in either the alternate energy resource sidebar or in the names and numbers appendix
Hazardous Materials
Ive gotten several calls asking how to dispose of any spent printed circuit etchant. Since Im part of the Haz-Mat team at our local fire department, I thought wed look at the big picture first. Back to square one. We can define a hazardous material as any substance that is out to do you in. It might burn, explode, be chemically reactive, be biologically active, poisonous, radioactive, or might interact violently with water or air. As a hardware hacker, you probably expose yourself to hundreds of hazardous materials. Knowing ahead of time what these nasties are and what they can do to you is common sense. Most industrial chemicals and byproducts have to provide a Materials Safety Data Sheet. Obviously, youll want to read and collect these. If you have any employees, you are required by law to make any relevant sheets available to them on request. Unfortunately, most MSDS listings do a total whitewash job and dwell on stupid trivia. For instance, the sheet on Kapton films goes out of its way to tell you that these films are slippery and you can trip on them if you leave them on the floor. But they make no obvious mention that Kapton turns from an insulator into a conductor above a certain temperature, totally trashing any computer or electronic system it is involved with. Always watch for understatement, weasel words, and any misdirection when you are reading these sheets. One of the best Haz-Mat references is the "yellow" book from DOT, and otherwise known as the Emergency Response Guidebook. This lists most industrial chemicals and tells competent professionals how to deal with
them under emergency conditions. Single copies are obtainable at no charge on request. A second resource is the 24- hour emergency telephone number from Chemtrec at (800) 424-9300. This is a free chemical industry service that can instantly place you in contact with experts on any hazardous material. But note this is an emergency response number, sort of a nationwide 911. Do not use it for idle chat. An interesting free trade journal that deals with all hazardous materials is Pollution Equipment News. This one also exposes you to some real pneumatic and electronic gems that you might not otherwise see. Naturally, any genuine Haz-Mat incident should be handled by the professionals, usually contacted with your nearby 911 number. One of the hardest things for the pros to learn are that their preconditioned "Save my Baby!" super-aggressive initial fire attack strategy is most often dead wrong in a Haz-Mat incident. Doing nothing at all is usually far better than becoming part of the problem. The quantity of the material and the way you store it will also make a big difference. Common sense is an obvious factor here. In the case of ammonium persulfate printed circuit etchant, storage of the dry etchant in small and sealed plastic containers is recommended. Tupperware works quite well for this. Some ammonium persulfate etchants also include a highly poisonous mercury activator. So, you never work in your kitchen, always be in a well ventilated area, and always wear safety glasses and plastic gloves. Note that ammonium persulfate is an oxidizer. If you do use a wooden spoon to stir it, and come back a week or two later, chances are youll have nothing but a burned-off spoon handle left. Do avoid all contacts between ammonium persulfate and any organic materials. Now, dont tell EPA I told you this, but you can simply flush spent ammonium persulfate etchant solution
LMA-117 Melts at 117 Degrees Farenheit. Contracts slightly on cooling. About $7.50 per ounce. LMA-158 Melts at 158 Degrees Farenheit. Expands slightly on cooling, About $1.35 per ounce. LMA-255 Melts at 255 Degrees Farenheit. Expands slightly on cooling. About $1.00 per ounce. LMA-281 Melts at 281 Degrees Farenheit. Expands slightly on cooling. About $2.20 per ounce.
Fig. 1 LOW TEMPERATURE alloys you can get through Small Parts. Two of these can be melted and cast with boiling water. What hacker uses can you think of?
28.1
Hardware Hacker
Please be sure to send your written entries directly to me and not over to Radio-Electronics editorial.
WARNING: Extreme shock and "hot chassis" danger in this circuit as shown! 1 F 1N4005 10 Meg 4.7 K cathode 115 VAC POWER LINE 1N4005 1 F All capacitors are 500 volts.
A Flame Detector
At one time way back when, far and away the numero uno hacker component was the NE-2 neon lamp. It was incredible what you could do with these. For panel lamps, night lights, polarity detectors, spike protectors, winking oscillators, tiny disco stars, tone generators for electronic organs, vacuum tube filament checkers, hot chassis ac checkers, triggers, voltage regulators, and bunches more. Sadly, the good old ten cent neon lamp isnt even on most hacker charts any more, done in by its need for high voltage, and its limited brightness and color range. Sigh. But the Hamamatsu folks have come up with an interesting variation on the traditional neon lamp. It is the UVTRON R2868 flame sensor. Figure two shows you a typical circuit. Any open flame has lots of ultraviolet energy associated with it. This circuit is a simple relaxation oscillator. In the absence of strong ultraviolet energy, the oscillator only fires every few seconds or so. When sensing the oscillator fires as much as 5000 times a second. Thus, you get an audio tone out in the presence of the ultraviolet energy from a flame, and only a few random clicks otherwise. This is sensitive enough to easily detect a match at 10 feet. While the recommended supply voltage should be within ten percent of 325 volts, you can usually work with the simple line-operated voltage doubler shown. Note that this is a "hot chassis" type of circuit with a severe shock hazard potential. Youll probably want to use an optocoupler on the output in your final circuit, and an input isolation transformer would also be a good idea. The output pulses are normally around eight volts high and about five microseconds wide. Two use tips. Be certain the anode resistor is physically rather close to your flame sensor. And avoid having two sensors near each other, since they produce ultraviolet in use and could badly interact or even latch. Hamamatsu also makes small and complete flame sensor circuit cards ready for use that run off the usual 24
R2868 UVTRON
OUTPUT PULSES
220 pf
10K .001
Fig. 2 AN ULTRA-VIOLET FLAME SENSOR. The circuit is basically a neon lamp relaxation oscillator. Its frequency is a few counts per minute in the absence of a flame, and a mid to high audio frequency with a flame present. $28 to $78 per pound. Unfortunately, the 117 degree one is by far the most
expensive, since it is around one-fifth indium, a precious metal. . Those two alloys with the lowest melting point can easily be cast from a "crucible" placed inside boiling water on your stove. Plaster or silicon rubber molds should work just fine. While you could, in theory, mix the alloys together to get different melting points, it pays to know what you are
doing. Otherwise, instead of a free flowing eutetic liquid, you might end up with so much dross. I can think of all sorts of unusual and off-the-wall uses for these, but why dont you tell me instead? As our contest this month, just dream up a unique application for a low melting point alloy. Therell be all the usual Incredible Secret Money Machine book prizes, along with an all-expense paid (FOB Thatcher, AZ) tinaja quest for two for the best entry of all.
OUTPUT 4.7 K
KA2223
B1 NF1 B2 NF2 B3 NF3 B4 NF4
1 2 3 4 5 6 7 8
.039 0.68
.012 0.22
3900 .068
1200 .022
390 6800
INPUT + 4.7 K 3.3 F (108 Hz) (343 Hz) (1080 Hz) 100K x 5 (3430 Hz) (10.8 kHz)
Fig. 3 A 5-BAND GRAPHIC EQUALIZER that uses a single Samsung integrated circuit. A pair of chips can be used for ten bands.
28.2
April, 1990
volt ac thermostat transformer. While intended for flame detection in gas heaters and in air conditioners, there should be plenty of other interesting hacker uses. What can you come up with here? substances. Figure four summarizes many of their key models. Quantity costs are in the $5 range. In general, each sensor consists of a sintered block of tin oxide, with variations that will optimize it for any particular sensed vapor. An internal heater is used to raise the surface temperature to the 200 to 400 degrees Centigrade range. Through surface adsorption, the resistance of the tin oxide will drop in the presence of an organic vapor. The amount the resistance drops depends on the vapor and its concentration. For instance, figure five shows you how the TGS822 gets used as a breath alcohol detector. You apply a 5 volt ac or dc heater voltage at a current of 120 milliamperes or so. You apply a main supply voltage of 24 volts ac or dc, and sense an output voltage across a 3.9 K load resistor. In the presence of 300 parts per million alcohol, the sensor resistance will drop by 10:1 . At 3000 parts per million alcohol, the sensor resistance will drop by 50:1. Thus, both the presence and the strength of the organic vapor can be measured. The response time is typically a few seconds. Full response curves appear on the data sheets and ap-notes available directly from Figaro. Do note that these are relatively broad spectrum devices. They may well respond more strongly to certain other vapors than the one you are actually looking for. Thus, they will work best in those situations where one and only one vapor or contaminant is expected. Ads for similar sensors often appear in Pollution Equipment News. Let me know if you want to see any more details on sensors of this type. evaporator and the condenser were both designed to operate at the same temperature and pressure. Oh well. I guess an awful lot of dreams died here. Somewhere along the way. On the other hand, the solar pump factory down the street is bursting at the seams with new activity. It seems a dose of reality has at long last caught up with alternate energy. The old order fadeth. So, where does this leave us? What are the key hardware hacker alternate energy resources for the nineties? This months resource sidebar gives us a clue or two. The very core of todays alternate energy appears to be a funky little magazine known as Home Power. Besides its no-nonsense shirtsleeves tech articles, this gem is full of ads for all
CMS-302 Oxygen combustion sensor. Externally heated to 700-900 C. 12 volt supply. TGS-100 For fumes and odors in air. 1 volt, 440 milliwatt heater. 100 volt supply. TGS-109 For natural gas and gasoline. 1 volt, 440 milliwatt heater. 100 volt supply. TGS-203 For carbon monoxide. 0.8 volt, 700 milliwatt heater. 12 volt supply. TGS-813 For propane and methane. 5 volt, 830 milliwatt heater. 24 volt supply. TGS-822 For DWI breath alcohol testers. 5 volt, 830 milliwatt heater. 24 volt supply. TGS-880 For automatic microwave cooking. 5 volt, 830 milliwatt heater. 24 volt supply.
Fig. 4 VAPOR DETECTORS from Figaro. Each has been optimized for specific sensing tasks.
28.3
Hardware Hacker
the leaders in the field. Cost is around $6 per year. Still at the same old stall is the good old Whole Earth Review, the ongoing continuation of the original Whole Earth Catalog and its progeny. Of the 497 magazines I personally subscribe to, this one is number two, right up there behind MAD. Nothing else even comes remotely close. Today the WER folks are heavily into CD ROM distribution and their major alternate lifestyle BBS system widely known as The Well. While both Mother Earth News and Popular Science have lots of useful alternate energy stuff in them, there is a key difference. PS will admit they have contacted a terminally incurable case of Yuppus dementus, while MEN does not. UPDATE: MEN folded. Ive found that HVAC News, a free air conditioning trade journal, also has a surprising amount of the alternate energy stuff in it. Although most of the other solar and the wind magazines have folded, Solar Energy remains as one useful, albeit a highly technical resource. Since subscription costs on this are totally absurd, youll want to go to a larger tech library to view a copy. The feds have bunches of alternate energy info available, spread over a dozen agencies. The Department of Energy often sponsors seminars and technical paper presentations. So does the National Bureau of Standards. Try
ALTERNATE ENERGY RESOURCES Arco Solar PO Box 6032 Carmarillo, CA 93010 (805) 482-6800 Assc Energy Engineers 4025 Pleasantdale S420 Atlanta, GA 30340 (404) 447-5083 Bergey Windpower 2001 Priestley Avenue Norman, OK 73069 (405) 364-4212 Energy Depot 61 Paul Drive San Rafael, CA 94903 (415) 499-1333 Home Power PO Box 130 Hornbrook, CA 96044 (916) 475-3179 HVAC Product News 400 N. Michigan Avenue Chicago, IL 60611 (312) 222-2000 Mother Earth News PO Box 70 Hendersonville, NC 28793 (704) 693-0211 Photocomm Inc 930 Idaho Maryland Road Grass Valley, CA 95945 (800) 544-6466 Real Goods 966 Mazzoni Sreet, #4B Ukiah, Ca 95482 (800) 762-7325 Sandia National Labs PO Box 5800 Albuquerque, NM 87185 (505) 844-5678 Snow Belt Solar 286 Wilson Sreet Amherst, WI 54406 (715) 824-3982 Solar Energy Maxwell House, Fairview Elmsford, NY 10523 (914) 592-7700 Solar Energy Res. Inst. 1617 Cole Blvd. Golden, CO 80401 (303) 231-1000 SolarJack 325 East Main Safford, AZ 85546 (602) 428-1092 Solarex 1335 Piccard Drive Rockville, MD 20850 (301) 948-0202 Solec International 12533 Chadron Avenue Hawthorne, CA 90250 (213) 970-0065 Solvonics Solar 1100 West Maple Road Troy, MI 48084 (313) 362-4170 Southwest Windpower PO Box 22178 Flagstaff, AZ 86002 (602) 526-0997 Whole Earth Review 27 Gate Five Road Sausalito, CA 94965 (415) 332-1716 Yellow Jacket Solar Box 253 Yellow Jacket, CO 81335 (303) 562-4884 Zomeworks PO Box 25805 Albuquerque, NM 87125 (505) 242-5354
any large technical library that has a government documents section. One fed document Ive found most useful is the Stand-Alone Photovoltaic Systems The Handbook of Recommended Design Practices. This one is available through Sandia National Laboratories.
FIGARO TGS-822
+24 VDC
OUTPUT VOLTAGE
3.9 K +5 VDC
Fig. 5 A DWI BREATH ALCOHOL TESTER that uses the Figaro TGS-822. The presence of 500 parts per million of grain alcohol lowers the cell resistance by a factor of 10, raising the ouput voltage.
There are lots of distributors of alternate energy products. One of the largest is Real Goods, who also offer a 320 page alternate energy sourcebook for $6.50 . Their competitors include Photocomm Inc, the Energy Depot, Yellowjacket Solar and Snow Belt Solar. Get their catalogs. Two establishment associations that have useful resource availability do include the Solar Energy Research Institute and the Association of Energy Engineers. Write them for their literature and service lists. The leading manufacturers of solar panels include Arco Solar, Solarex, Solec, and Solvonics. Two windpower sources include Bergey Windpower and Southwest Windpower. Finally, two of the alternate energy "good guys" definitely include Steve Baer of Zomeworks, who is heavy into solar trackers and energy management; plus Jim Allen of SolarJack, who specializes in higher efficiency solar pumps and controllers. Please let me know what else you think needs added to this list. And be sure to tell me which new resource sidebars you want to see included in
28.4
April, 1990
future issues.3
28.5
Don Lancasters
Hardware Hacker
May, 1990
Hall Effect Resources Shape Memory Alloys Cycolor Printing Secrets Low Noise Amplification Unusual Mechanical Stuff
Some of the capsules get selectively hardened when exposed to light. A pressure roller then squashes the soft capsules, releasing their dye onto a suitable final paper, plastic, or slide material. Like stomping on baloons. A brief heating completes the imaging process. Neat stuff. A free and interesting brochure on Cycolor is obtainable through Mead Imaging, while the leading companies offering Cycolor machines include Brother, Noritsu, Pic-Mount, and the folks at Seiko-Mead. Probably the best way to start off hacking this stuff is to get free demos of the various machines and then rip off small quantities of the supplies for your own testing and use. I can think of dozens of specialized hacker uses for this new technology. What can you come up with?
f you ever find out whatever it is we are doing here, please be sure and let me know. At any rate, this months goodies include a brand new color imaging process, low noise amplifiers, Hall Effect devices, and some unusual new non-electronic items that cry out for hacker use. As usual, please observe that all of our referenced sources do appear in either the upcoming Hall Effect resource sidebar or in the Names and Numbers appendix. Do check these resources before using our free help line. And please make your product and your literature requests to anyone listed in either sidebar specific, rather than general. Onward and upward
The cyliths shells are sensitive to different color light. On exposure, the cyliths will harden. For instance, red light will harden the cyan cyliths but leave the magenta and yellow ones soft. White light will harden all of the cyliths, while black (no light at all) will leave all the cyliths soft. To create a color print or slide, an image is projected onto the developer sheet containing all the cyliths. After exposure, your developer sheet will consist of a mixture of hard and soft cyliths. You now have a latent image on your sheet that is somewhat similar to a color negative. The developer sheet is then placed in pressure contact with a suitable paper, slide, or transparency material. The pressure breaks all of those soft cyliths, releasing their internal dyes. The hard cyliths remain intact and do not release their dye. For instance, if youve got hard cyan cyliths and soft magenta and yellow cyliths, then the magenta and yellow dyes will get released and will mix to form a red image. For a blue image, those yellow cyliths harden, leaving the magenta and cyan cyliths soft. Squashing the soft cyliths mixes a magenta and cyan to produce a deep blue. And so on. One more time. Bunches of cylith capsules are on the polyester carrier.
CYAN DYE
MAGENTA DYE
YELLOW DYE
POLYESTER CARRIER
Fig. 1 THE CYCOLOR PROCESS uses tiny cylith microspheres. Each of these cyliths is initially soft, but hardens under exposure to certain light wavelengths. Crushing the soft cyliths releases a final image dye.
29.1
Hardware Hacker
from an ordinary ic to a low noise amplifier in such instances will not in any significant manner reduce your apparent output noise. It could even make things much worse. Third, the wider the bandwidth, the more noise you are going to have. The wideband circuits are all inherently noisier than narrow band ones. All other things being equal, noise will increase as the square root of your bandwidth. Most often, if you quadruple your bandwidth, you double the noise involved. Fourth, integrated circuit noise does tend to increase very dramatically at extremely low frequencies. If a precision dc response is not needed, you should be better off capacitively coupling an ac or audio amplifier. Fifth and finally, circuit noise is a very strong function of frequency and source impedance. There is no such thing as a universal "low noise" amplifier. In fact, a "low noise" amp run in the wrong frequency range or at the wrong input impedance will actually generate more noise than might a plain old op-amp. Thus, if you want low noise, it is super important that you have all the data sheets and applications literature on hand for the device you want to use, plus a thorough understanding of the options available in other chips from other sources. There are two types of noise that the first stage of an integrated amplifier can generate. These are voltage noise and current noise. It turns out that current noise is directly proportional to the square root of the first stage collector current, while voltage noise is inversely proportional to the square root of the collector current. Iffen the right one dont git ya, the left one will. Voltage noise tends to be constant with frequency and more important at low impedance levels, while current noise tends to increase sharply below a noise corner (or "1/ f") frequency in the low audio range. Current noise is more important at higher input impedance levels. Voltage noise is usually measured in nanovolts per square root of bandwidth, with any value under 1.0 being fairly decent. Current noise gets measured in units of picoamperes per square root of
LT1028
LT1037
LT1001
LT1012
LT1055
LT1037 10 Hz 50 500 5K
LT1001
LT1012
50K
500K
5 Meg
50 Meg
Fig. 2 ANY LOW NOISE AMPLIFIER must get specified over a certain input impedance level and a certain frequency range. Here are Linear Technologys selections for the "best" low noise op-amps. Note that a misapplied "low-noise" amplifier can actually be noisier than a regular one.
While all of these sources have lots of free and excellent technical info available, I am most impressed with the Linear Technology offerings at the present time. But, before we really get into this, let us shatter several hacker myths involving low noise amplification. Firstoff, in no way will a low noise amplifier by itself reduce the noise
from an already noisy source. All a low noise amplifier can do is limit the amount of extra noise added by its own amplification process. Second, in the overwhelming majority of hacker circuits, most noise problems are caused by an improper circuit layout, sloppy decoupling, by instabilities, poor shielding, and by undesired ground currents. Switching
1115
VOS IN- IN+ V1 2 3 4
1 F -18V
316K 1K
30K
316K 1K
Microphone Input
Fig. 3 AN ULTRA-LOW NOISE BALANCED PREAMP circuit that uses the Linear Technology LT1115. Noise and distortion is under 0.028 percent.
29.2
May, 1990
bandwidth, again with anything under 1.0 being desirable. Figure two does show us several possible selections for Linear Technology amplifiers that will give you the "best" noise performance for a given impedance level or frequency range. Of the previous generation devices, that Precision Monolithics OP-37 is considered the "parent" of these newer low noise devices in the same way that the original 741 started the whole op-amp universe. Lets look at a typical device. The Linear Technology LT1115 is an ultra low noise and low distortion audio op amp. It works best when ac coupled at higher gains and over an optimum input source impedance of 30 to 300 Ohms. Figure three does show you a simple microphone preamplifier that receives balanced inputs (for hum cancellation) and provides a singleended output with a gain of 300. The total of combined noise and harmonic distortion is under 0.028 percent over a 20 Hertz to 20 Kilohertz bandwidth. Should you truly and genuinely need the lowest voltage noise possibly get and already have total and absolute control over all the usual ground, hum, stability, and stray noise sources that usually swamp hacker circuit first stage noise figures, there is one sneaky trick you can pull. Simply arrange several low noise amplifiers in parallel and sum all of their outputs into a second-stage low noise amplifier. The gain goes up with the number of amplifiers, while the noise only goes up as the square root of the number of amplifiers in use. Five parallel amplifiers seems a practical limit, which, in theory can reduce your first stage noise by a factor of slightly over two. In the real world, this only works for input impedances less than ten ohms. Worse yet, circuit strays are likely to eat you alive if you attempt this. More details on the LT1028 data sheet. A new material known as Trovicel looks like a really good board product for use on everything hacker from cases through prototype models and displays. Trovicel is a rigid foamcenter plastic sheet that comes in lots of attractive colors and in several thicknesses. It is lightweight, tough, paintable, outdoor rated, printable, pretty, and self-extinguishing. You work it with pretty near any hand or power tool, although hacksawing and power shearing are not recommended. You can bond it with ordinary PVC cement, and even grind or weld it. Free samples are available. Should you need any curved plywood in tubes, quarter-round, halfround, or with a radiused end, try out some of the Deco Shapes product from Laminates Incorporated. There seems to be a lot of interest lately in machinable waxes. Besides being obviously ideal for CAD/ CAM training and program debugging, the material is well suited for one-ofa-kind mockups, proof-of-concept models, and for "does-it-fit?" product verification. With a machinable wax, your tools stay sharp, the chips are safe, and you can easily recycle your final material simply by remelting it. The product is available from two to eight inches thick; in cylinders up to nine inches in diameter; and in bulk. You could easily recast their wax into any shape you want to. It melts at 310 degrees Farenheit. One good source for machinable waxes is Freeman Supply. Prices start at $4 per pound. The American Safety Razor folks are obviously one sharp outfit. Their catalog lists zillions of different low cost blades, including ripple ones for cutting potato chips and big mutha ones nearly eighteen inches long! The variety here is amazing. One obvious project around here is a $5 cutter for laser printed business cards. Im wondering if you couldnt use their blades for a clamping paper cutter as well. A proper support, of course, would be essential. In what just might be the most obscure free trade journal in the entire world, Power and Bulk Solids covers such things as bin level detectors and similar goodies. Amazingly, several hacker helpline callers per week are asking for this sort of stuff, especially for agricultural electronics. A new product called Scotch 9703 Conductive Adhesive Transfer Tape from 3-M has a very unusual property. It can conduct electricity through its thickness but not along its length.
IN
A G E
PU
L VO
TA
GE
LL HA OR S EN
MA
GN
ET
Fig. 4 THE HALL EFFECT in a semiconductor will produce a transverse output voltage that is proportional to the strength and direction of a magnetic field. Analog or digital Hall devices are available.
29.3
Hardware Hacker
From face to face, the resistance is under 0.1 Ohms. Yet, along its length, it insulates to the point of allowing 500 volts across a ten mil gap! Embedded silver-plated particles apparently are the secret. Adhesion values are in the 50 ounces per inch range for most substrates. Free test samples are avilable. Finally, short shape metal alloy wire samples are available free from Beta Phase. These magic wires can be bent into any shape at all and will stay that way until heated to hot coffee temperatures. Then theyll rapidly spring back to their original size and shape. You can repeat their shapeand-restore process zillions of times. Applications include everything from high density zero-insertion-force connectors to dentistry. Tellyawhat. For this months contest, just dream up an off-the-wall use for any of these unique new materials. Theryll be the usual dozen or so Incredible Secret Money Machine book prizes, along with an all-expense-paid (FOB Thatcher, AZ) tinaja quest for two going to the very best of all. As usual, send all your written entries directly to me at Synergetics, rather than over to Radio-Electronics editorial offices.
N UG
30
13
3 2 1 MA GN
ET
Fig. 5 THE SPRAGUE UGN3013 is a low cost Hall Effect sensor having a snap action digital output. The open collector output remains high in the absence of a magnetic field and goes low when a field is present.
across the film, and there will be no Hall output voltage. Now, should a magnetic field be applied into the film, the current will shift towards one edge or the other, and a Hall voltage will be generated whose strength is proportional to the magnetic field and whose polarity depends on the polarity of the applied magnetic field. This Hall action is related to the magnetic deflection in a television picture tube and follows the "right hand rule" where the resultant force of a magnetic field causes an electron beam current to move sideways. The Hall voltage is usually rather low, and is typically in the tens of millivolts. In most Hall applications, fairly strong magnets must be placed quite close for reliable operation. In general, there are two types of Hall Effect sensors, linear and digital. Linear sensors are used whenever you wish to measure the actual strength of a magnetic field. A Gaussmeter is one instrument that can use Hall Effect devices to measure medium-strength magnetic fields. F.W. Bell is a major source for the precision linear Hall devices and instruments. Digital sensors are used whenever the presence or absence of a magentic field is intended to produce a "yes-no" output. Figure five shows you one typical digital Hall Effect sensor, the Sprage UGN-3013 . Which is a three terminal device about the size and shape of a small signal transistor. The price is under a dollar. A supply voltage is applied. While this can range from 4.5 to 25 volts, the usual +5 is typical. There is an open collector output that is off in absence of a magnetic field. Note that you have to provide an external pullup resistor to get any output. Typical supply current is three milliamperes, plus the "on" current through the load resistor. This chip is guaranteed to turn on with a magnetic field of 500 Gauss, or roughly 500 times the strength of the Earths magnetic field. Its also guaranteed to turn off with a magnetic field of less than 30 Gauss. To prevent chattering or noise, a snap-action hysteresis of 75 Gauss or so is usually provided via an input Schmitt trigger. A cheap Alnico magnet a third of an inch in diameter and a quarter inch thick will activate your chip from a
29.4
May, 1990
distance of 0.15 inches. Sensitivity drops off dramatically with distance. On the other hand, flux concentrating pole pieces can be added for extended range. As can stronger magnets. Other devices are available with different sensitivity. Some are bipolar in which you purposely reverse the applied field, rather than dropping it to zero. This can give you a more reliable sensing as alternate poles of a magnet go fliping on by. Two excellent and sometimes free books on the Hall Effect include the Hall Effect Sensors data book SN-500 from Sprague, and the Hall Effect Transducers applications book from Micro Switch. The latter shows a $10 optional list price. There are bunches of hacker uses for Hall Effect devices, since they can sense though a modest non-contacting distance in a dirty environment. Obvious digital uses now include high reliability keyboards, automotive ignition contacts, shaft angle encoders, position detectors, for model railroad controls, speed sensing, brushless dc motor commutation, and tilt switches. Linear applications include current sensing, circuit breakers, and direct magnetic field measurements. Hall devices also show up in all the usual surplus catalogs, and in many of the ads that youll find right here in Radio-Electronics.
HALL EFFECT DEVICES AND RESOURCES F. W. Bell 6120 Hanging Moss Road Orlando, FL 32807 (407) 678-6900 M & C Magazine 2994 West Liberty Avenue Pittsburgh, PA 15216 (412) 343-9666 Micro Switch 11 West Spring Street Freeport, IL 61032 (815) 235-6600 Sensors Magazine 1244 Reamwood Avenue Sunnyvale, CA 94089 (408) 744-1500 Siemens 19000 Homestead Road Cupertino, CA 95014 (408) 257-7910 Sprague 70 Pembroke Road Concord, NH 03301 (603) 224-1961 Texas Instruments PO Box 809066 Dallas, TX 75380 (800) 232-3200 Wabash Magnetics 1375 Swan Street Huntington, IN 46750 (219) 356-8300
CATV Systems, and the other RF Components for Satellite Receivers. And two free new surplus flyers MWK Industries for lasers and optics, plus Silicon Valley Surplus for great robotics prices. What appears to be a stupendous breakthrough is described in A light source smaller than the optical wavelength, beginning on page 59 of the January 5, 1990 issue of Science. This one is downright scary in its potential. Free samples this month include a MC14507 single supply RS-232 driver from Motorola, a Johnny come lately imitation of those Maxim chips that pioneered this field. Also, a bunch of freebie chips from Texas Instruments. These include their TL499A battery backup DC-DC converter, a TL1451A pulse width modulator chip, and their SN75C188 line driver circuit. TI also has free samples of all their new OTP (one time programmable) EPROM s. These work like a "real" EPROM, except they are in a cheaper plastic case and can only get programmed one time. You use them for repeat copies of known good code. From NEC, the new Digital Signal Processor and Speech Processor data book. And from Silicon Systems, a Microperipheral Products Data Book that covers chips used in floppy disks, hard disks, and cassette tapes. Free software this month includes a Generic Array Logic Demo through SGS-Thomson and a CUPL Demo Disk From Logical Devices. Included here is a My First PAL Design booklet.
Tektronix has a new and free Scope Evaluation Kit available that includes a test circuit board which purposely generates overshoots and hard to see glitches. They also have several free oscilloscope videotapes available. Surprisingly cheap ($3.95) analog hygrometer humidity instruments are available from Klockit. Turning to other mechanical stuff, Stock Drive Products has a new data book and master catalog available on such things as gears, belts, couplings, robotics breadboards and such. To improve the appearance of your prototypes, the DonJer folks have low cost "instant suede" applied via liquid glop and a cabbage duster. 28 colors are available, as is a $52 professional starter kit. A nice effect. Yet another reminder here that the Midnight Engineering is a great new hacker magazine aimed at all you small-scale hardware and software productions. Free samples are available by special arrangement. Turning to my own products, for the fundamentals of digital integrated circuits, check into both my classic million-selling TTL and CMOS Cookbooks. I do have some autographed copies on hand here for you. I also have a new and free mailer for you which includes dozens of insider hardware hacking secret sources. Do write or call for a copy. As always, this is your column and you can get technical help and offthe-wall networking per that Need Help? box in the appendix. 3
29.5
Don Lancasters
Hardware Hacker
June, 1990
omeone at long last has done something right for a change. Hacker friendly too. A giant Yea Team! and a sixpack of attaboys to all of the folks at Dallas Semiconductor, a manufacturer of innovative new integrated circuits. Dallas has set up a free 24 hour direct order line with no hassles and no minimums. If you want one or two of their chips at 3:42 in the morning, just whip out your VISA card and give them a call. For next day delivery. Whats really sad is that virtually all of those other "good guy" innovative chip manufacturers invariably use by far the most Neanderthal and the most hacker vicious of the old-line distribution channels. So here and now, I issue an outright challenge to all the rest of the "good guys" - Maxim, Sprague, Linear Technology, Samsung, Mitsubishi (who has far and away the most mesmerizingly awful reps and distributors in the industry, bar none), Signetics, Reticon, SGS, Statek, Teltone, Rohm, Intersil, Sony, Siemens, Philips, et. al. For your own good, please, please, find out why Dallas is laughing all the way to the bank. And please set up your own direct order lines that do not go out of their way to kick sand in the collective faces of your most promising future customers. Your present distributors are your worst enemy. As usual, please observe that all of our referenced sources do appear in either the upcoming Cases and Enclosures resource sidebar or in the Names and Numbers box. Please do check out these sidebars before using our free help line. And please do make all your product and your literature requests to anyone listed in either sidebar specific, rather than general. We return you now to our column already in progress
Electronic Levels Low Cost Memory New Form of Matter Cases and Enclosures LaserWriter Tech Info
down to the individual component parts level. Thus, you can get a fuser assembly from them, but not a fuser roller. One source Ive found useful for any and all LaserWriter parts on any level is Don Thompson, who also stocks detailed repair literature, low cost rebuilds, toner supplies, and the handy and useful repair tools. One additional source for a LaserWriter schematic is Bomarc Services. As we have seen before, you can easily refill your own SX or CX toner cartridges in two minutes for $7.50 . The job is much simpler than properly packaging your cartridge for shipping to a commercial refilling service. Besides Don Thompson, another major source of refilling supplies that I personally use is Arlin Shepard of Lazer Products. Arlin also now offers "infinite life" recoated SX drums. One Im now using is on its 14th refill. From what I have been able to determine, much of the rest of the toner refilling industry is an outright zoo. We have people selling you shoe polish and calling it drum recoating, others gold plating things that do not
their LaserWriter printers. Outside of the limited "white book" LaserWriter Reference on all of those connections and the commands (in stock here at Synergetics), virtually nothing at all is available from them on their printers. Such stupidity ends up monumentally short-sighted. Fortunately, Hewlett-Packard is as hacker-friendly as Apple is hackervicious. The Canon CX and SX engines used by both companies in their printers are nearly identical. Over 95 percent of the mechanical parts are fully interchangeable. And HP , Like Dallas Semiconductor above, also has a free VISA order line with overnight express delivery. HP has a pair of outstanding service and repair manuals that are directly applicable to the Apple printers. Their manual #33440-90920 is for the older CX engine as used on the LaserWriter and LaserWriter Plus. Their manual #33440-90904 is for the newer SX engine that is used on the LaserWriter NT and LaserWriter NTX. Unfortunately, HP sells only the replacement modules, and rarely goes
DS 1
222 2
2 3
5 +1.
VDC
DA
TA
Fig. 1 THE ECONORAM IS A MICROPOWER memory of 256 bits. These can cost less than a quarter each in quantity. Amazingly, all addressing, data transfers, and all read/write control are done through one very busy pin!
30.1
Hardware Hacker
hold the output low for between 1 and 15 microseconds. The EconoRAM should respond by remaining low for another 15 microseconds if a zero is stored, and by immediately going high if a one is stored. Thus, to read, bring the output low for 15 microseconds. Wait 5 microseconds. Then sense the one or zero. The first 8 bits in your sequence are called the command word. The first five are all ones to write, or all zeros to read. This is followed by a pair of zero bits which allow a later memory expansion. The eighth header bit is always a one. The remaining 256 bits are your data, arranged in a sequentially addressed order. Which is much like a shift register would work. To change a single bit, you first read all of the bits, store them in host memory, and then rewrite all of the bits, changing what you want on the fly. Obviously, this RAM is best suited for long term storage applications only ocasionally read from or written to. A 5K pullup resistor is recom- mended when writing or reading. The output pin is TTL compatible. Now for the tricky part. To first initialize your EconoRAM, you send it 264 write zero states. This automatically resets the internal address counter. After that you send a new sequence to do whatever reading or writing you care to. Remember that you must go completely through a sequence each and every time or the EconoRAM will get confused. Four EconoRAMs are easily used together for a full 1K that should be enough for a complete name, address, and account number. All you do is change the address bits in your command string through 00, 01, 10, or 11. As many as 32 bits could also get permanently factory written in for a positive and genuinely non-volatile message header. Let me know what new hacker uses you might come up with using this innovative new approach to small memory needs.
header D 0 D 1 D 2 D 3
header D 0 D 1 D 2 D 3
Fig. 2 THE ECONORAM GETS INTERROGATED by use of 264 sequential timing pulses. Once started, the sequence must get completed in its entirety. See the DS2222 data sheet for additional details.
need gold plated, and yet others who are making useless "gapping" adjustments that try and compensate for the shoddy quality of their toner. For a good glimpse into the entire toner refilling industry, check out the highly interesting and informative Recharger trade journal. Full details on the toner cartridge reloading appear in previous columns and in my Hardware Hacker II bookon-demand published reprints.
The cost is under a quarter in large enough quantities. Supply voltage is between 1.2 and 4 volts dc. Standby operating current is a mere 100 nanoamperes. Less than one milliampere is needed for a read or write operation. A typical small coin cell is good for 100 million transactions over several years of operation. What boggles the mind is that there are only three pins on the DS-2222 . Power, ground, and a single pin that, believe it or not, performs all of the addressing, data transfer, and control of reading and writing. You just about have to use a host microprocessor or personal computer to read or write to this chip, since the process is so involved. Heres how it usually works: To read or write any one bit, all bits must get read or written. Figure two shows details. Your host contoller provides a high output and then issues a string of 264 pulses. The timing can be any interval longer than 60 microseconds per pulse. Each pulse event begins when the output drops from supply to ground. When writing, a logic one must last between 1 and 15 microseconds. A logic zero must last at least 60 microseconds. And so, the length of your write pulse determines whether you enter a one or a zero. When reading, you have your host
30.2
June, 1990
Obviously, if you just need a plain old box, your local Radio Shack has several bargain priced ones in stock, as does Mouser Electronics, as well as nearly all of our other fine RadioElectronics surplus advertisers. The trick is to find something beyond a plain-jane box that looks sharp, makes a statement you can relate to, and isnt outrageously expensive. Lets briefly run down the major case and enclosure manufacturers: Bud , of course, who is pricey, very old line and uninspired; Vero having a foreign look about them; Hammond with a good selection of ABS mini- cases; Polycase who feature the wall mounted and plug-in stuff; PacTec with an incredible variety of sloping small consoles; Keystone for built-in battery compartments and standoffs; Serco for some fancier, highly styled boxes; and Vector who are both low cost and look it. Want to wrap your own instead? Figure three shows you three of my own favorite home-brew packages from way back when. While most of the Zero Manufacturing cases are ludicrously expensive, they also do offer a line of plain old deep-drawn aluminum boxes with rounded corners in zillions of sizes that are reasonably priced. These annodize beautifully, and many models have lids which exactly provide an outside or an inside fit. And the photographic dialplate materials from either Metalphoto or Fotofoil can integrate beautifully with these cases. For dozens of examples of these, check into the back issues of Popular Electronics and Radio-Electronics in the 1965-1975 time frame. Vinyl clad materials are used by all the big folks. Getting them in small quantities can be a real hassle. Instead, you simply go on down to your local Yellow Front or an equivalent yuppy pseudo-surplus store, buy some fake Naugahyde by the yard, and glue it onto plain old aluminum or steel. The "clamshell" design of figure 2-B is easy to do. You can simply sand the bottom half to get a satin finish, and glue vinyl onto the upper half to get a professional final result. The wooden rail ploy of figure 2-C is both easy to hack and looks great. Just get yourself some exotic wood from EDLCO or Constantine (Coco-
The ZERO BOX uses a deep drawn aluminum shell and looks particularly good with a Metalphoto dialplate. Shallow versions can also be used face up. These anodize beautifully.
The VINYL CLAMSHELL has a brushed aluminum chassis and a cover coated with glued-on decorator vinyl. Tilting end brackets can also be added, as can ventillation holes.
The WOODEN END RAIL ploy uses exotic wood caps that have suitable grooves routed in them. The chassis itself can be painted aluminum or a polished brass or chrome.
Fig. 3 HOMEBREW CASES FROM the "golden age" of hardware hacking. All of these are easily built, yet can produce outstanding professional results.
bolo or Wenge are fine choices), mill some slots in it with a hobby motor tool, slide in a bent metal frame, and you are home free. Ive found a local blacksmith or sheet metal shop to be real handy at times, so youll want to find a good one of these on your own. Better yet, see if you cant find a horse trailer or hitch works. They have the machinery to properly cut and bend the heavier stuff without costing an arm and a leg. And theres usually enough scrap on the floor under the shear. Four interesting package support outfits are the Fomeboards people which do stock all sorts of beautiful prototyping sheet materials; Coburn who is heavy into unusual finishes
such as prismatics, glow-in-the-darks, diffraction gratings, foils, etc; instant dust-on flock materials from DonJer, and the Ultra-Suede from Red Spot, a textured urethane finish having a soft fuzzy suede or smooth leather touch. All the enclosure and packaging people advertise in most of the free electronic trade journals. Electronic Component News seems about the best for cases and such. Two other trade journals with useful fit and finish ideas in them do include both Electronic Packaging and Appliance. If you have any favorite hombrew packaging ideas of your own, please send them in to me so we can share them with the others. There are all sorts of possibilities here.
30.3
Hardware Hacker
dielectric constant, a medium viscosity, is non-wetting, non-corrosive, and has a very high vapor pressure. Although propylene is one possible choice, there are lots of silicone oils that should serve as well. After filling, your case is sealed. When level, each of those butterfly plates will be immersed just as deep in the dielectric liquid. As the sensor tilts, one plate goes deeper and the other will become shallower, thus changing the capacitance. The higher the dielectric constant of the liquid compared to air, the more profound the capacitance change. To sense, you let each wing of the butterfly set your pulse width of a monostable built out of a pair of 555 timers or one single 556 . You then measure the time difference between the two to determine the angle. Obviously, a low-end microcontroller could greatly simplify your digital display. Far and away my favorite here would be a Mitsubishi M50734 or one of its single-chip alternatives. To get fancier, place a pair of the sensors back-to-back and sharing a common ground. This gives you two big advantages. First, your sensed capacitance change is now doubled, which should give you more accuracy. Better yet, your cross axis sensitivity should drop dramatically. This will happen because an unwanted forward or reverse tilt increases the depth on one side and decreases it on the other. The shape of the plates determines the linearity of the capacitance versus the slope angle. Sometimes, you may like to purposely change the plate shape to get a non-linear response. One use might be to automatically calculate compound miter cut depth on a table saw. It seems to me you could easily make up a capacitance sensor using nothing but a pair of printed circuit boards, a spacer and a large O- ring. The bottom board would form the ground plane. The spacer would have a hole in it somewhat larger than the Oring and would act as a compres- sion stop, setting a fixed width. And the top one would have the butterfly pair on it. Er, on second thought, why dont you tell me? For this months contest, either (A) show to me an eminently hackable design for the capacitance inclinometer, or (B) dream us a new
Fig. 4 A SIMPLE INCLINOMETER can be created by using nothing but a plumb bob and a protractor. You sight along the base of the protractor base and then read the angle.
Fig. 5 AN ELECTROLYTIC LEVEL is one older electronic inclinometer. As it tilts, the bridge resistance between the sensing pins will change. Bromine is one possible liquid.
Electronic Inclinometers
The folks at Wedge Innovations are now retailing intelligent electronic levels which display your choice of degrees, pitch, slope percent, level & plumb, autocalibration, and even a simulated bubble. The suggested list price is under $80. Related electronic protractors are being offered by Lucas Sensing Systems. I just thought we might take a quick look at some of the principles of electronic level sensing. You can easily build your own level sensor for under $4. In general, there are two popular ways of telling which way is up. One quite expensive method is the vertical gyro. This is simply a gyroscope that is spun up while level and stays that way when the world around it moves. Two surplus sources of vertical gyros include Fair Radio Sales and the folks at C&H Sales. A much simpler method is the inclinometer. As figure four shows us, an inclinometer can be as simple as a plumb bob and a protractor. Your gravity sensing plumb bob usually points straight down. As the protractor is rotated, its slope angle can be read. There have been several older attempts at getting an electrical output off an inclinometer. Obviously, you can simply wipe a potentiometer, but stiction, wear, and hysteresis can end up as problems. Other early schemes
used mercury, a hazardous element thats both poisonous and rather low in impedance. Figure five shows you another early attempt at an electronic inclinometer. This one is known as an electrolytic sensor. You place three probes in a conductive liquid in a sausage-shaped enclosure. As the sensor tilts from level, the deeper probes resistance drops, while the shallow one will increase. A simple op-amp bridge circuit can convert this differential resistance into an output voltage. One obvious choice for a liquid is bromine, one of those few elements which remain liquid at normal temperatures. Two older sources of the electrolytic level detectors are Hamlin and Spectron. These devices are both fragile and expensive. Optical shaft encoders could be used, but an absolute encoder would have to have at least twelve bits of resolution for decent accuracy. Most of the bets these days, though, are on the capacitance inclinometer shown you in figure six. This one is simple, cheap, low power, and clean. Picture an insulated enclosure the size and shape of a Magicians fake silver dollar. The rear of the enclosure is a grounded metal plate. The front consists of a pair of butterfly shaped capacitor plates. The case gets filled exactly halfway with a magic liquid that is an inert insulator, has a high
30.4
June, 1990
use for an electronic angle measuring device. There will be all the usual Incredible Secret Money Machine book prizes, with an all expense paid (FOB Thatcher, AZ) tinaja quest for two going to the very best entry of all. As always, send all your written entries directly to me at Synergetics, and not to Radio-Electronics editorial.
CASE AND PACKAGING RESOURCES Appliance 1110 Jorie Blvds, CS 9019 Oak Brook, IL 60522 (708) 990-3484 Coburn 1650 Corporate Road West Lakewood, NJ 08701 (201) 367-5511 Constantine 2050 Eastchester Road Bronx, NY 10461 (212) 792-1600 DonJer Products Co Ilene Court Building 8 Belle Mead, NJ 08502 (800) 336-6537 EDLCO PO Box 5373 Asheville, NC 28813 (704) 255-8765 Electronic Component News 1 Chilton Way Radnor, PA 19089 (215) 964-4345 Electronic Packaging 1350 East Touhy Avenue Des Plaines, IL 60018 (708) 635-8800 Fomeboards 2211 North Elston Chicago, IL 60614 (312) 278-9200 Fotofoil/Miller Dial 4400 North Temple City Blvd El Monte, CA 91734 (818) 444-4555 Hammond 1690 Walden Avenue Buffalo, NY 14225 (716) 894-5710 Keystone Electronics Corp 31-07 20th Road Astoria, NY 11105 (718) 956-8900 Metalphoto 18531 South Miles Road Cleveland, OH 44128 (216) 475-0555 Mouser Electronics 11433 Woodside Avenue Santee, CA 92071 (800) 346-6873 PacTec Enterprise & Executive Avenues Philadelphia, PA 19153 (215) 365-8400 Polycase 4726 Superior Avenue Cleveland, Ohio 44103 (216) 391-0444 Red Spot PO Box 418 Evansville, IN 47703 (812) 428-9100 Serco 612 Commercial Avenue Covina, CA 91723 (818) 331-0517 Vector Electronic Co 12460 Gladstone Avenue Sylmar, CA 91342 (818) 365-9661 Vero 1000 Sherman Avenue Hamden, CT 06514 (203) 288-8001 Zero Halliburton PO Box 3339 Pacoima, CA 91333 (818) 897-7777
A free Designing With Plastics; The Fundamentals booklet is obtainable from Hoechst Celanese. Turning to my own products, I am now self publishing nearly a dozen titles using my new book-on-demand PostScript technology. Four of these you might find interesting are the Hardware Hacker II reprints, (which contain all of my Radio-Electronics columns, updated, edited, and cross indexed), my Ask The Guru volumes I and II, and my brand new LaserWriter Secrets book-disk combo.
Ive also started up a major new PostScript and desktop publishing BBS on Genie. Our goal is to have a thousand free downloads very soon. See you there. Finally, Ive got a new and free mailer for you which includes dozens of insider hardware hacking secret sources. Do write or call for a copy. As always, this is your column and you can get technical help and offthe-wall networking per that Need Help? box. The best calling times are weekdays 8-5. 3
30.5
Hardware Hacker
30.6
Don Lancasters
Hardware Hacker
July, 1990
s weve seen in past columns, within a decade we just may reasonably expect electronic hardware having computation abilities that are as complex as those of the human brain. Driven primarily by ultra-cheap and ultra-dense dynamic RAM technology. Now, admittedly, my timing on this might be off by as much as twenty minutes in either direction. Things seem to be going even faster now than they were a few issues ago. Several Japanese outfits have now actually firmly scheduled their future production of 256 Meg x 17 SIMM chips. And it appears another level of threshold effect just might have been crossed with the concept of OOPS or object oriented programming which is speading like wildfire. So how can you tune yourself into and profit from this near-term inevitability of human brain functionality? As usual, those free industry trade journals are the answer. In this case, the E.E. Times is a very good choice. They are very big on dynamic RAM developments, on neuron computing, and on "fuzzy logic" in general. Cold fusion. This is being written on cold fusions first birthday. And according to the Wall Street Journal for April 3, 1990, on page B4, "a small but growing number of scientists believe they are seeing entirely new phenomenon," and "There is now a growing consensus that you cant explain the excess energy by errors." It appears, in fact, that several distinct and valid cold fusion mechanisms might be involved. Sixteen labs have recently reported fresh and positive new results. A very interesting report titled as Electricity by Serendipity did appear in the March 2, 1990 issue of Science. This describes a new type of fuel cell which permits the premixing of the hydrogen and the oxygen gases, something previously thought to be utterly impossible. These cells also involve hydrogen being absorbed or adsorbed into palladium, in cells remarkably similar to the cold fusion setups.
Electronic tune chips Unusual book resources Ten speed blender Control Power control fundamentals A hacker interchange standard
listings, charts, or most any other tech writing or tech illustrations? Putting any of these in the magazine does take up valuable space, and the third-generation photocopies or litho negatives you sometimes end up with are often of low accuracy and quality. And it sure gets hard to make a correction three months later on 360,000 already printed pages, some of which are buried in a lesser-known suburb of Moose Jaw, Montana. On the other hand, the BBS downloading will only work if the sender and sendee do have fully compatible hardware and software. What to do? I would like to suggest an obvious solution. It is called an EPS file, and is short for Encapsulated PostScript. An EPS file can easily handle any printed circuit layout, schematic, isometric or perspective drawing, any dialplate, or text or graphics of any sort to arbitrarily high resolution. The EPS files are extremely device independent. Since an EPS file is an ordinary ASCII textfile, it could easily be used with any word processor or any editor with any brand of personal computer of any age and CPU. At the output end, an EPS file can be used directly on a laser printer, by a typesetter, or by a photoplotter. With
The obvious unanswered question is "how much of cold fusions excess heat is really this previously unknown type of fuel cell?" And vice versa. Our usual reminder here that most of the resources mentioned have been gathered together into the Names and Numbers appendix or the Unusual Book Resources sidebar. You can contact all the listees directly for more info. UPDATE: As of fall 1992, the cold fusion has largely failed to prove its case. No conclusive evidence of any atomic processes have been shown to date. Cold fusion appears to have been downgraded to a laboratory curiosity having few supporters. But labwork does continue. Lets start with a lively topic
Engineering, Computer Shopper, Nuts and Volts, Byte, Audio Amateur, and Circuit Cellar Ink have all been separately grappling with a common problem: How can clean and accurate printed circuit layouts get easily and cheaply put in the end users hands? Or dialplates, drilling templates, detailed test and debug info, software
3300
T2
GATE
T1
Fig. 1 A TRIAC CAN BE USED as a line operated ac power switch which can directly control lamps, heaters, or motors. A brief and small current pulse into the gate turns the triac on; it remains on until the main current reverses.
31.1
Hardware Hacker
layout package? Surely you could not provide a free copy of these to every reader in the country, could you? Especially if you are on a Mac and they are on a PC. There is a simple little trick you can pull that is called pseudo-compiling. With pseudo-compiling, you might convert any source of PostScript code into a Just the facts, Maam form that uses nothing except the simplest of put-and-place PostScript commands. Anyone can run a pseudo-compiled EPS file, and absolutely nothing will remain of the original code or the way it was generated. Better yet, that pseudo-compiled PostScript often will run ridiculously faster than the original. And pseudocompiling is easily and quickly done by using Adobe Systems brand new and cheap Distillery program, by one of my gonzo justification routines, or your own custom code. What about screen images? In any broad based user interchange standard, you certainly would not want to use any screen images if they hurt the device independence in any way. Fortunately, systems using display PostScript or the display PostScript emulation are now becoming readily available, as is software that returns printer bitmaps for screen display. So are programs and applications which internally capture and display EPS. How does an EPS file differ from an ordinary PostScript textfile listing? Actually by very little. There are a few required remarks at the start of your file plus a few optional ones. And a very few PostScript commands (such as initgraphics) are not allowed if they would somehow corrupt the program or system that the EPS file is being imported into. Several other obscure commands (such as settransfer) must be carefully saved and restored if they are altered. EPS files are also limited to a single page each, but you can use as many of them as you like. To get started on all of this, get yourself copies of that Encapsulated PostScript Files Specification, v 2.0 , and the Document Structuring Conventions Specification, v 2.1. Both of these are available free upon request through Cynthia Johnson at Adobe Systems, or else through one of the PostScript BBS systems. You should not create or use any EPS files without
+OUT NC -OUT
2N6154 TRIAC
+IN -IN NC
1 2 3
Fig. 2 A MICROCOMPUTER TO TRIAC interface uses a phototriac optoisolator to let safety isolated logic signals directly control high power loads. Depending upon the input waveforms and the load, this same circuit can be used for either an on-off or a proportional phase control. A low input powers the lamp or other load.
simple and crude emulators such as Freedom of the Press, GoScript, and UltraScript, you can also print your results on ordinary dot matrix or ink jet printers, albeit to a lower resolution. The latest of fax machines are set up to directly handle EPS files, some to device independent resolutions that can exceed 2650 DPI . There is no high-end limit to print quality. Better yet, practically all of those
either now use PostScript internally or at the very least are capable of generating EPS output files. The EPS files can also be extremely compact. For instance, a fancy circuit schematic should need no more than a 10K textfile maximum. So, EPS stores cheap and downloads fast. But what if you are using an expensive or proprietary CAD/CAM or
100 watt light bulb 22K 250K diac 0.1 F 2N6154 TRIAC 110 VAC
Fig. 3 PHASE CONTROLLED DIMMER delays triac turn-on to a chosen point in each ac half cycle. Use this circuit ONLY for incandescent lamps, heaters, soldering irons, or "universal" motors that have brushes.
31.2
July, 1990
having these two on hand. The fundamentals on PostScript are covered by Adobes blue PostScript Tutorial and red PostScript Reference Manual, while ready-to-use printed circuit layout, schematic, isometric, and perspective routines are found in my PostScript Show and Tell. Let me know your thoughts on a standardized hacker interchange format based upon standard EPS files. While triacs are fun to experiment with and easy to use, do note you are dealing directly with the ac power line here. You have an extreme, and possibly lethal, shock hazard. Always do keep one hand behind your back. Never leave power applied after a test or experiment. And always use an isolation transformer should you want to view any scope waveforms. Figure two shows you one simple computer to ac power interface. You take a special optocoupler known as a phototriac isolator and use it as a safety interface. When your computer port goes low, a LED in the isolator turns on, which fires up an internal baby phototriac, that in turn whomps the main triac. This safety isolates your computer from the power line, yet still allows a small logic signal to directly control a kilowatt or more of ac line power. Even with the safety isolation, you still have a severe shock hazard on the right side of figure two. Observe that this interface works backward from what youd expect. A low input powers your load, and vice versa. This makes far better use of the micros current sinking ports. Under certain conditions, a triac can be used to control the brightness or speed of your power load. One way to do this is by using phase control, such as is shown in figure three. If you do precisely delay the gate pulse to some point in the ac half cycle, only partial ac half cycles will get through to the load. Since full power is only applied for part of the time, you end up with an average lower power sent to the load. But note that phase control will not work with any old power load. Phase control is fine for heaters, soldering irons and other resistive loads, and is really great for an incandescent lamp dimming. While phase control can be used to control the speed of a universal ac-dc motor which uses brushes, a
0.68 F
270 1 Watt
1N4004 220F
13 V
470 pf
100
16
15
14
13
12
11
10
LS7314
PL3 PL4 PL5 PL6 PL7 PL8 PL9 PL10
1 2 3 4 5 6 7 8
Fig. 4 A 10-SPEED TOUCH CONTROL BLENDER circuit that uses the low cost LS7314 chip by LSI Systems. The eleventh touch pad is for power off.
31.3
Hardware Hacker
slightly fancier feedback circuit that senses the motor back emf will often work considerably better. Note that a simple phase control absolutely, positively must never be used with a fluorescent lamp or an ac induction motor. While it is possible to use triacs to dim a fluorescent or to control the speed of an induction motor, these need much fancier and carefully crafted circuits. Here is how your simple dimmer control works: The diac shown consists of two four layer diodes back to back. It behaves sort of like a neon lamp. It is an open circuit until its terminal voltage gets up to fifty volts or so. Then it will suddenly turn on, dumping the capacitor charge into the triac gate. The diac turns back off when your capacitor is discharged. Varying the brightness control decides how fast the capacitor charges, and thus when in the cycle the triac fires. The later in the cycle, the less the percentage of the time that power is applied to the load, and the dimmer the lamp. Since this all happens 120 times per second, the repeated off-on actions are well beyond your critical persistence of vision limit, and no flicker will be noticeable. The long thermal inertia of the lamp filament also helps eliminate flicker. Diac trigger diodes are usually sold by the same sources who supply triacs. These days, though, it is far better to digitally generate all the power line phase delays by either using a microprocessor or a custom power control chip. One example might be triac. A reference input on pin 12 synchronizes an internal phase lock loop to the power line zero crossings. Double resistors are used on the touch plates as a safety feature. The active touch plate inputs can also double as LED output drivers, thus lighting one-of-ten selected LED lamps to show your current power or speed level. I have left the RFI noise filter off the output to keep our circuit simple here. You can also customize all of your firing angles to improve linearity for your selected load. Consult the data sheet for full details on these extras.
9 volt battery
LS3404
SP1 SP2 RCN RCS
1 2 3 4
(tempo) 0.1
(duration)
Fig. 5 A HIGH QUALITY MELODY circuit. The slow decay waveforms create chime-like notes. Pitch, tempo, and duration are adjustable.
31.4
July, 1990
personally found of interest. Lets do a quick rundown The foremost source for hacker titles has to be Lindsay Publications, who do stock an amazing variety of mechanical, machine shop, antique radio, craft, and weird science titles. For automotive electronics, nobody can beat SAE. And for electronic music, the MIX Bookshelf is first rate. For astronomy and space books, try Frontier Space Books. For robotics, use Educational Products. For all the alternate energy and self-reliance, try Real Goods. For establishment energy and conservation stuff, AEE Energy Books. For signmaking, its SignCraft. For metals and properties of materials, ASM International. It appears that two of the finest technical bookstores in the country are across the street from each other. So, you will want to check out both the Stanford Bookstore and the Computer Literacy Bookshop. For a direct mail source of most technical and industrial titles, you also may want to check out Omega Engineering. By far the best place Ive found for in-depth reviews of all unusual and genuinely useful books is the quarterly Whole Earth Review. I cannot say enough good things about these wonderful folks. I try to stock my own titles and only the very best PostScript titles of others in my own Synergetics. It seems Ive always been attracted to what I call "Taint Likely McGee" publications on Tesla-was-an-alien free energy, perpetual motion, Pogue carburetors, Newmann engines, and similar off-the-wall goodies. Although Lindsay stocks and honestly reviews a few of these titles, a second and more typical source would be H&A Industries. As a final favorite of mine, Singing Wind does specialize in the Western Americana, Southwest Literature, lost mine lore, and an otherwise amazingly eclectic collection of arcania. Since Singing Wind has an unlisted address, getting there is half the fun. If youre not one of their kind of people, youll never find the place. Well, maybe one hint. Go north on Ocotillo Road out of Benson, Arizona till it feels about right. Then turn right just beyond the seventh cow. Keep the gates the way you find them.
UNUSUAL BOOK RESOURCES Antique Radio Books 498 Cross Street Carlisle, MA 01741 (508) 371-0512 AEE Energy Books PO Box 1026 Lilburn, GA 30226 (404) 925-9558 ASM International 9639 Kinsman Road Materials Park, OH 44073 (800) 368-9800 Computer Literacy 2590 N First Street San Jose, CA 95131 (408) 435-1118 Dialog Info Service 3460 Hillview Avenue Palo Alto, CA 94304 (415) 858-2700 Educational Products PO Box 606 Mineola, NY 11501 (516) 689-8409 Frontier Space Books 2400 Foshay Tower Minneapolis, MN 55402 (800) 245-8627 SAE 400 Commonwealth Dr Warrendale, PA 15096 (412) 776-4841
SignCraft H&A Industries 1938 Hill Avenue Rt 2 Box 35E Bowling Green, MO 63334 Fort Myers, FL 33901 (813) 939-4644 Lindsay Publications PO Box 12 Bradley, IL 60915 (815) 468-3668 MIX Bookshelf 6400 Hollis St Ste 12 Emeryville, CA 94608 (800) 233-9604 Omega Engineering One Omega Drive Stamford, CT 06907 (800) 826-6342 Real Goods 966 Mazzoni Street Ukiah, CA 95482 (800) 762-7325 Singing Wind Bookshop Ocotillo Road Box 2197 Benson, AZ 85602 (602) 586-2425 Stanford Bookstore 135 University Avenue Palo Alto, CA 94305 (800) 533-2670 Synergetics Box 809 Thatcher, AZ 85552 (602) 428-4073 Whole Earth Review 27 Gate Five Road Sausalito, CA 94965 (415) 332-1716
Tellyawhat. Im sure Ive missed plenty of good sources for unusual books. For this months contest, just tell me about some other genuinely useful place to find specialized or oddball titles. Do include a mailer if possible. We will have a dozen Incredible Secret Money Machine book prizes, along with an all-expense-paid (FOB Thatcher, AZ) tinaja quest for two going to the very best. As usual, send all of your written entries directly to me at Synergetics, rather than to Radio-Electronics.
have a free supplement Direct Connection: Interfacing to the Telephone Network available just for you. Ask for Application Note #102. Reasonably priced vacuum forming machinery is available through Ron Charles & Associates. While intended for hobby modelmaking, there should be lots of hacker uses involving prototypes and limited production runs. Two interesting new prototyping and model materials are the Komacel and Celtec foamed vinyls. Some free samples are available. Turning to my own products, Ive put together a special new Lancaster Classics package of six of my finest hardware books plus that complete Hardware Hacker II reprint set which covers everything which has appeared in these RE columns. Finally, Ive got a new and free mailer for you which includes dozens of insider hardware hacking secret sources. Write or call for info. As always, this is your column and you can get technical help and offthe-wall networking per that Need Help? box. The best calling times are weekdays 8-5 in Mountain Standard Time. Lets hear from you. 3
31.5
Hardware Hacker
31.6
Don Lancasters
Hardware Hacker
August, 1990
heres a pair of stupendous new high end compaction breakthroughs for this month, so I thought it might be a good time to go over some of the fundamentals of
Huffman Codes Video Line Drivers Run Length Encoding New Compaction Chips Amateur Radio Astronomy
Several popular areas of compression interest include ASCII text strings, fax printouts, delta modulation, disk data formats and full color HDTV video images.
and typical pictures are extremely redundant, these limits often end up surprisingly generous. But do note that the less redundant the information, the more intolerable any errors will become. Further, any simple errors tend to become far more destructive. For instance, any uncompressed error which misspells any single word could instead totally trash your entire message in your compacted file. This error intolerance, along with the long processing times, are the two serious tradeoffs that get involved in extreme compaction. Compression schemes which are content dependent are often (but not always) more effective than schemes that can accept anything as an input. In general, the more you know about what you want to compress, the more you can squash it. For instance, the compacted data string alincgburgadd should not take you four score and seven years to expand into its full text version, especially if you have got a library handy. Or, as a second example, you could use a city, state, and a zip code consisting of the single digit 0 for your own town in a local mailing list.
Symbol "A" Symbol "B" Symbol "C" Symbol "D" Symbol "E" Symbol "F" Symbol "G" Symbol "H" Symbol "I" Symbol "J" that is 20% that is 18% that is 10% that is 10% that is 10% that is that is that is that is that is 6% 6% 4% 4% 4%
Text Compression
Eight bit bytes of 256 states are a wasteful way of showing ASCII text letters, which at the most needs 127 characters, and often require only 26. Early Infocom text adventures used a scheme which crammed three ASCII characters into a pair of 8- bit bytes, a compression to 67 percent. Which also made any snooping for hints real tough. The Adams adventures instead used some character pairs. The most commonly used occurrences of two characters got placed in unused code slots, again giving you something like a 3:2 compression on the average. In spelling checkers, the words are most often arranged alphabetically. Since each new word will share many initial characters with a previous one, you could provide a leading number instead. For instance, to get from fixation to fixative, the second word could be coded as 6ve. You can also apply character pairs as well as a trailing byte that could represent the
likely codes as 10. likely codes as 000. likely codes as 011. likely codes as 110. likely codes as 111. likely codes as 0101. likely codes as 00100. likely codes as 00101. likely codes as 01000. likely codes as 01001. likely codes as 00110. likely codes as 001110. likely codes as 001111.
Fig. 1 A HUFFMAN COMPRESSION USES SHORT CODES for often occurring letters or symbols, and longer codes for those less often used. On the average, your overall message will end up shorter. Here is a simple example code which involves thirteen symbols.
32.1
Hardware Hacker
256 most popular combined ways of ending words, s, ing, ly, ed, and so on. When all is said and done, it takes only slightly over 9 bits to represent an average English word in a spelling checker dictionary. Advanced compaction schemes often are first described in the Communications of the ACM from the Association for Computer Machinery. Going to a non- ASCII interpretive code can also help you bunches. For instance, to describe a single curved portion of a PostScript character, six signed integers, seven spaces, and the word curveto get involved. For a total of around 42 characters. In Adobes Type I Fonts, a special character string code gets used, of which 32 entries are commands, and the rest signed numerics of one or two self-spacing bytes. Typically, only six bytes will be needed for the same curveto example, giving you a compaction of 6:1 or so. More details in Adobes Type I Font black book. I do have a few of these in stock. The ASCII letter "e" usually needs eight bytes for transmission. So does an ASCII "z". But the "e" gets used far more often. So, why not set up a code where an "e" needs four bits, while a "z" needs fourteen? On the average, all of your text messages will end up somewhat shorter. Such a scheme is called Huffman coding. The original horses mouth document is A Method for the Construction of Minimum-Redundancy codes, as found in the September 1952 Proceedings of the IRE. An example of a Huffman code appears in figure one. Huffman codes are based upon making the length of a character code inversely proportional to the popularity of that characters use. The good news here is that a Huffman coding is fairly fast and simple, and often can be applied in combination with other compaction techniques. The bad news is that a variable bit length word can be awkward to handle in an 8-bit or 16-bit micro, and that the total compaction you get is often rather modest and typically in the 20 percent range. Worse, errors under a Huffman coding do tend to trash the entire message, rather than a single word or two. Variations and improvements on
sssssss ssssssss sssssssss sssssssss ssssssssss sssssssssss ssssssssssss ssssssssssss sssss sssssss sssss sssssss sssss sssssss sssss sssssss sssss sssssss sssss sssssss sssss sssssss sssss sssssss sssss sssssss ssssssssssssssssssssssssss sssssssssssssssssssssssss sssssssssssssssssssssssss sssssssssssssssssssssss sssssss sssssss sssssss ssssssssssssss ssssssssssssss ssssssssssssss ssssssssssssss
00 00 01 0F 7C 7F 00
07 3F F7 87 07 FF 7F
F0 F0 C0 C0 C0 FF FE
00 00 00 00 00 00 00
00 00 03 1F FF 00 00
0F 7F E7 07 FF 07 7F
F0 F0 C0 C0 FF F0 FE
00 00 00 00 C0 00 00
00 00 07 3E FF 00 00
1F FF C7 07 FF 07 7F
F0 C0 C0 C0 FF F0 FE
00 00 00 00 80 00 00
00 00 07 3E FF 00 00
1F FF C7 07 FF 07 7F
F0 C0 C0 C0 FF F0 FE
00 00 00 00 80 00 00
D7 7C 15 D8
CC 55 77 C9
8C 37 C0 DB
B9 C5 F0 9D
CA 53 A6 B9
9C 7C 0F DB
8A 45 09 9D
C9 BC 8C C8 CC 75 17 C6 52 47 C3 55 7C 26 57 C2 65 7C 70 F0 97 01 F0 78 D8 CD 8C B0
Fig. 2 RUN LENGTH ENCODING IS AN EFFECTIVE compression scheme for black and white images. In this example, four bit bytes representing how many white or black pixels replace the bitmap for a 2:1 compaction.
75 Ohm Termination
75
Video Source
Fig. 3 THE CORRECT WAY TO CONNECT several video monitors to a single source is to daisy chain them. Only the last monitor input should be terminated with a 75 ohm resistor. Never connect cables in parallel!
32.2
August, 1990
this scheme are known as the modified Huffman codes. These are in no way limited to ASCII . Any place you have symbols with different probabilities, you can use variable length codes to your advantage. Often, your message will first get analyzed, and a set of rules in the form of a dictionary will get sent as a prolog. Matching the coding to the message usually will give you the optimum compaction. Most modern fax machines use what is known as a Group III compression, which is based largely upon run length encoding. One source of circuits and info on Group III compression is American Micro Devices, especially their new AM95C71 video compression-expansion chip set. But the best way to compress a fax image is to flush it and use PostScript instead. Besides offering you a ridiculously better image quality, most PostScript superfax images transmit in one-tenth the time and will store in one-tenth the space. on the Apple IIe simply by the brute force coding of only frame-by-frame changes in successive pictures. But delta modulation really shines for sound synthesizers. To compact voice or music, you try to anticipate where the sound is going to be in the next sample, and then transmit only the value difference. This is known as linear predictive coding, and the sound toolkits on the Macintosh and the Apple IIgs include lots more detail. A typical compression is 2:1 or 8:3 , depending on the acceptable quality. I use an even sneakier form of delta modulation to speed up PostScript. Say you have a shipping label or a business letter with a fancy letterhead and a digitized signature. Instead of retransmitting the entire image each time, you will simply place the total image into your PostScript printer once and then print the first name and address. Youll then erase only the name and address, and then print up your second name and its address, repeating until your entire mail list or whatever is finished. The same idea works well with forms, or anywhere else a long makeready time can be shared among many different copies.
75 1.05 K
coax cable
video output #1
75
8 7 6 5
1K
video input
MAX457
OUT IN- IN+ V1 2 3 4
75 1.05 K
coax cable
video output #2
75 0.1F 1K
-5V
Fig. 4 THIS DUAL VIDEO DRIVER is handy any time you want to use more than one monitor with a VCR or a computer, especially with long cables.
32.3
Hardware Hacker
space. So, there are lots of possibilities here, over and above the obvious of trying to increase the number of tracks or the number of bytes per track. Knowing exactly how your disk space is organized can often save you bunches of bytes. As one ancient example, leaving off those last eight invisible bytes of an Apple II HIRES image saved you an entire DOS disk sector per picture. In ProDOS, a 511 character textfile requires only one sector, while a 513 character one needs three sectors, one for the first 512 characters, a second for the new index block now required in your switch from a seedling to a sapling file, and a third for the final data actual byte. In fact, you can save at least one sector anytime you can get your file size just under a multiple of 512. Full details in the Beneath Apple ProDOS from Quality Software. On the Mac, removing any unused resources from your System files can often free up surprising amounts of disk space. This becomes especially critical on earlier Mac versions with limitied RAM Similarly, knowing how your disk files are arranged and stored on any system can sometimes save you quite a bit of space. Especially on a disk that is nearly full and needs room for one or two more short files. Our first major compression chip breakthrough for this month is the IC-105 co-processor from Infochip Systems. This one is initially intended for the MS/DOS world and gives you an average of a 3:1 or higher compaction for disk files. And does so while remaining fully noiseless and lossless, and still allowing a full random access. Compression is real time up to 2 Megabytes per second, and the decompression blazes along at a 5 Megabyte per second rate. An adaptive algorithm is used that senses the type of data and creates optimum data dictionaries for each type. A Markoff scheme is used. Interestingly enough, you do not have to store these data dictionaries. They will automatically build themselves up during the decompression process. Image compression can go as high as 15:1. The initial cost is in the $200 range. The obvious first use for this chip is for tripling the disk storage on any laptop computer. It will be interesting to see how error tolerable this new technique is. This should end up a real winner if it performs reliably enough in the real world. atrocity out of the CD-ROM water. The obvious uses include genuine multimedia for personal computers, compact full animation on CD-ROM and for squashing those bandwidths needed for HDTV broadcasts to well under those currently being used for plain old tv. C-Cube does have a free technical info packet for you on their products, along with demo boards and software emulators. Projected pricing of their chip is in the $95 range.
Fig. 5 THIS NEW SMT GRABBER offered by Tektronix has a bendable snout. Free samples are available.
32.4
August, 1990
selected with a switch near the video input on the monitor or display. Note that you cannot simply put a bunch of monitors in parallel, since all of their loads will gang up and swamp your source, leaving you with a very washed out picture, poor sync, or even nothing at all. Many of the video output stages on personal computers are rather wimpy and dont take at all lightly to long cables, especially parallel ones. Instead, figure four shows you a dual video driver using the MAX457 . Each output is an independent 75 ohm source. You can cascade as many of these as you need to. To make up for the drop across the 75 ohm source resistance, the video gain is set to precisely two. Half of this develops across the source, and half across your terminating load. At this gain of two, the bandwidth is around 30 megahertz, so any high definition stuff can be handled as well as ordinary video. For local HDTV use, the open loop bandwidth is 75 Megahertz. A split +5, -5 supply is required. Sorry about that, but you do have to somehow suck out current on all your grounded sync tips. Supply current is under 40 mils for the pair of drivers. The normal output level is adjusted to a white level of 2 volts and grounded sync tips. This becomes a standard one volt video after the drop across the source resistor. Naturally, you must not rats-nest breadboard any 75 Megahertz video amplifier. Use a clean double sided printed circuit layout with extensive ground plane areas and a thorough supply bypassing, or instabilities and oscillation are nearly certain to result. Other uses do include driving A/D converters. Please let me know what unusual applications you can think of for this great new circuit.
AMATEUR RADIO ASTRONOMY RESOURCES Amateur Radio Astronomers 247 North Linden Street Massapequa, NY 11758 (516) 798-8459 Astrophysical Journal PO Box 37005 Chicago, IL 60637 (317) 753-3347 Bobs Electronic Service 7605 Deland Avenue Fort Pierce, FL 34951 (407) 464-2118 Jeff Lichtman 1425 Parkmont Drive Roswell, GA 30076 (404) 992-4959 Nasa Tech Briefs 41 East 42nd Street Ste 921 New York, NY 10017 (212) 490-3999 NRAO Observatory PO Box 2 Greenbank, WV 24944 (304) 456-2011 Science/AAAS 1333 H Street NorthWest Washington, DC 20005 (202) 326-6400 VLA Astronomy Site PO Box "O" Socorro, NM 87801 (505) 772-4011
of similar stuff. His free catalog also has a good no-nonsense intro. Jeff has published three classics in the field which include his fine Radio Astronomy Handbook, Solar Radio Astronomy, and his Microwave Radio Astronomy. All well done. A leading amateur astronomy club is apparently the Society of Amateur Radio Astronomers. Their $20 yearly membership includes a newsletter. Publications providing radio astronomy info in them include Science, the Astrophysical Journal, the NASA Tech Briefs, and the old IEEE Radio Astronomy Transactions. Two radio astronomy sites which welcome visitors are the NRAO observatory located in Greenbank, West Virginia, and that VLA site in New Mexico. The VLA (very large array) is sort of in my back yard, so stop in and visit when you get there. That VLA is situated within the Magdalena-Pietown-Quemado metro area. Just take the loop thruway past the theater and industrial districts. Then turn left at the fourth cow. The VLA is one of the most spectacular scientific installations anywhere in the world. What about ET watching? Because of all the potential kook aspects and professional credibility, SETI activities are not normally the main focus of most serious amateurs. Serious extra terrestial intelligence searching needs incredibly complex hardware and rigorous elimination of locally explained sources, not to men-
tion a gonzo supercomputer or two, a thorough understanding of the cross correllation functions and Fourier transformation, and a zillion acres of land in a low-interference locale. But you are certainly encouraged to try. Thats what hacking is all about. When you make contact, be sure and say Hi for me.
32.5
Hardware Hacker
32.6
Don Lancasters
Hardware Hacker
September, 1990
ts not at all obvious to me why we need all of the foot dragging, infighting, and squabbling going on today over HDTV high quality video display standards. It seems that several government agencies are now battling each other to win the coveted role of chief obstructionist. To me, it is entirely obvious that HDTV will use square pixels, will not have interlace, will make use of fully programmable, instead of hard-wired (single standard) receivers and displays, will be totally digital, will use real time JPEG compression, and will follow Japanese standards. It is also totally obvious to me that terrestrial broadcasting will serve a negligible to totally vanishing role in HDTV, while the computing, satellite, VCR , and cable uses will overwhelmingly dominate. And any intermediate or interim transition steps will prove to be a monumental waste of time and money, done by the wrong people for the wrong reasons. So, lets just ban the networks and the feds from any HDTV input whatsoever, and then get on with it. They are the enemy, not Japan. Our topics this month range from the ridiculous to the sublime
Perpetual Motion Audio Log Voltmeter Independent Research The Magnetocaloric Effect Refrigeration Breakthrough
common traits. None of them have ever attended an introductory college physics course, or else they seem to have slept through it. While all of them claim they "just cant find anything at all" on their idea, they studiously go out of their way to avoid doing any real or honest library research. As weve found out a few times in the past, any hardware hacker anywhere can instantly get the very latest scoop on anything by way of the Dialog Information Service. We will see more on this shortly. A disproportionate number of the perpetual motion buildees seem to belong to one particular religion that happens to be very big on faith and on self-reliance. Theres often a very heavy dose of paranoia, usually aimed at the patent office, a local university, all the oil companies, Detroit (which could not possibly suppress anything except quality or profits), an ex-boss, or else "them" in general. Almost always, the buildees think linearly instead of cyclically. Thus, while a power stroke of the repelling magnets or their freezing milk bottle
repelled each other, this gyroscope would accelerate. Or so I thought at the time. Very strangely, that gyroscope locked up instead of spinning. Seems it latched itself into a minimum reluctance field position and just sat there. These days, I guess I dont really understand why perpetual motion is desirable. Since unlimited free energy would hasten the entropic heat death of the planet, perpetual motion is both environmentally unconscionable and socially reprehensible. The first thing we should do to a successful perpetual motion machine designer is to stake them to an anthill. And leave them there until the next meeting of the steering committee. Nonetheless, perpetual motion is a fascinating topic. Several very good books on this subject are available from Lindsay Publications, while a few of the more interesting ongoing perpetual motion scams are available to you through H&A Industries or the Tesla Book Company. After working with quite a few of them, the perpetual motion buildees these days all appear to share several
S N
S
S S
N
N
Fig. 1 MY VERY FIRST PERPETUAL MOTION MACHINE. As the like poles of the magnets repel each other, they accelerate the rotor on a gyroscope. Sadly, it latched, rather than speeding up. But maybe if I used stronger magnets and better bearings
S
N
33.1
Hardware Hacker
Several of us folks around here at
Radio-Electronics editorial have now
(1) - If you cant find it, you aint lookin. (2) - Research is a continuous activity done on a total lifestyle basis. You cannot turn it on and off at will. (3) - Anticipate what you are going to research in the future and get started on it long ahead of any actual or possible need. (4) - Your own personal resource file should be your first and foremost starting place. Your begining file should fill at least eight cubic yards, and should expand from there. (5) - The free trade journals are far and away your most important external resource. A complete listing appears in Uhlrichts Periodicals Dictionary. Use those bingo cards! (6) - The Dialog Information Service at your local library can instantly find anything for you from anywhere at anytime. (7) - Other important library resources are the Science Citations Index, the Thomas Registry of Manufacturers, Interlibrary loan, the UMI reprint service, and the Encyclopedia of Associations. (8) - Form an extensive network of contacts outside of your circle of family, friends, or work associates. Clubs, helplines, and BBS systems are ideal for this sort of thing. (9) - Let your subconscious be your guide, starting from fundamental principles and guidelines. Go with the vibes. (10) - Quest tinajas. .
somehow gotten some silly ideas into our collective heads. For some unbeknownst reason, many of us presently feel that
(A) Neither matter nor energy can be created or destroyed, except by an atomic process. (B) Available energy always seems to convert itself from higher quality forms into lower and less useful ones. Not once have the dishes ever washed themselves. Nor have those pool balls ever re-racked by themselves. (C) Nearly all physical and electronic processes will end up producing unrecoverable low grade heat energy, usually through friction or electrical resistance. (D) Despite a house-senate compromise committee, congress is not expected to repeal the three laws of thermodynamics this session. Paraphrased, these three laws are (1) You cant win; (2) You cant break even; and (3) Yes, the dice are crooked, but it is the only game in town.
makes a lot of sense to them, they will usually ignore all the repetitive and cyclic energy supplying stages needed to get to that step. Theres also the cosmic cupcake syndrome, the few chips shy of a full board affliction, and the boy a whole flock of them flew over that time con-
cept. But we need not get into any of these here. Finally, there is the magic bullet. Their idea almost but not quite works. So, all well need to fix it are better gears, stronger magnets, a larger milk bottle, or a different rear axle ratio. Or more bucks for research.
N
ICE GADOLINIUM ROD
I guess one of the reasons some of us around here feel this way is that not once in the history of hardware hacking has even one reproducible counterexample to these silly ideas of ours ever gotten successfully and unarguably demonstrated. Naturally, you are free to agree or disagree with us as you wish. But if you disagree, we do make only one simple request: Provide us with an experiment that can be independently duplicated by disinterested outsiders which causes your effect to show up at least reasonably well. Then we will all believe.
S
Fig. 3 THE MAGNETOCALORIC EFFECT, very greatly oversimplified. Gadolinium and other rare earth alloys can absorb heat energy when in the presence of a magnetic field and release it otherwise. The efficiency can be as much as 40:1 better than mechanical cooling. Magnetic refrigeration is usable over an absolute zero to above room temperature range. Important first uses will likely be in cyrogenics, superconductivity, liquified gases, and for hydrogen fuel research.
So, what is the best way to research any topic? I dont know how many calls and letters I have gotten from people which live in such a "remote" area and will claim that "absolutely nothing" is available locally. Believe it or not, one of these letters was from Cambridge, MA and yet another was from Palo Alto, CA. Well, I have been sitting right here watching Gila Monsters on this sand dune smack dab in the middle of the
33.2
September, 1990
Upper Sonoran desert for nearly two decades now. Although almost everything I do is local (and much of it is done underground or in mountaintop wilderness areas), I have had zero problems whatsoever in handling top quality research on all kinds of very rewarding and well-paying topics. So do not give me any "remote" bull. Admittedly, my tiny and isolated town of 2400 now does have its very own symphony orchestra. But that is another story. Figure two lists a few of my key secrets to doing independent hacker research. The overwhelming reason you cannot find something is because you are not looking. You are instead going through several inept motions and keeping yourself busy, instead of taking obvious steps and handling all of them in a logical manner. Research is not an activity that you turn on or off. Instead, you put yourself in a continuous research mode in which you gather and collect everything, needed or not, or expected or not. Never mind which topic. Your subject does not matter in the least, because chance favors the prepared mind. Thus, your very own personal resource file is far and away the most important place to look, should any specific need come up. Set a minimum goal of eight cubic yards for your personal resource files. At least for a bare bones startup. Then let it grow from there. Your foremost outside resource should be all of the trade journals. I subscribe to over 400 of them. As we have seen in the past, all fields do have all their own private technical magazines which are intended for a select group of insiders. Most of these are free, provided you tell them what they want to hear on their qual cards. Many include bingo cards, annual directories, and tech info. Naturally, youll circle everything even remotely useable on the bingo cards. If in doubt, circle it. If you do not personally rent the largest box in your local post office, youve missed the point here completely. Electronic trade journal examples include EE Times, Electronics, EDN, Electronic Products, Electronic Design, and dozens more. Well look at these names and numbers in some future sidebar. For now, you go to the library and view the Hackers Holy Grail, that is otherwise known as the Uhlrichts Periodicals Dictionary. This single hardware hacking resource is far and away more important than all of the rest put together. Use and enjoy. The second most important outside Hardware Hacking research resource is the Dialog Information Service. Its also at your local library. For a small fee, Dialog will instantly research anything, anywhere, anytime. As an example, well shortly be looking at magnetic refrigeration. To get from knowing virtually nothing on this topic to having eighteen of the most recent worldwide key abstracts in hand took me seven minutes and cost me a total of $27.57. Theres lots of other goodies in the library, should you snoop around in enough nooks and crannies. In general, the least valuable things in any library are its takeout books. Theres that free Interlibrary Loan Service which lets you pick up anything from anywhere, and the faster UMI service which could get you any reprint of anything provided you know the exact title, journal, and pages. For much less than Dialog, so long as you dont mind waiting a few days. Theres also a Thomas Registry of Manufacturers that lists who makes everything, but Ive found this to be of limited utility. Also, check into the Encyclopedia of Associations, and, if you cant locate Uhlrichts, then the International Standard Periodicals Dictionary is almost as good. Another library favorite of mine is
Magnetocaloric Effect in Strong Magnetic Fields A.M. Tishin, Cyrogenics (UK), February 1990, v30 #2, pp 127-136. Magnetocaloric Effects in Rare Earth Magnetic Materials A.S. Andreenko, et. al., Soviet Physics (USA), August 1989, v32 #8, pp 649-664. Magnetocaloric Effect in Thulium C.B. Zimm, et. al., Cyrogenics (UK), September 1989, v29 #9, pp 937-938. Magnetic Refrigeration Superconductivity Industries, Spring 1989, v2 #1, pp 32-41. Magnetocaloric Effect and Refrigerant Capacity of Tb-Dy Alloys S.A. Nitkin, et.al, Physics Status Solidi (East Germany), May 1989, pp 117-121. Magnetic Refrigerator for Superconducting Magnets at 1.8K V.A. Altov, et.al., ICEC 12 (UK), 1988, pp 635-640. Magneto-thermal Properties of Sintered Gadolinium E. Gmelin, et. al., ICEC 12 (UK), 1988, pp 432-436. Determination of the Cooling Capacity of Magnetic Refrigerants S. Nikitin, et. al., Soviet Technical Physics Letters, April 1988, v14 #4, pp 327-328. Magnetic Refrigerator T. Hashimoto, Refrigeration (Japan), 1988, v63 #733, pp 1189-1201. Magnetic Field Changes in the Entropy of Europium Sulphide P. Bredy, et.al, Cyrogentic (UK), Sept 88, v28 #9, pp 605-606. Magnetothermal Conductivity of Er-Al for Cyrogenic Applications C.B. Zimm, et.al., Journal of Applied Physics, 15 April 1985, v57 #8, p4294-4296. Adiabatic Temperature Changes in Ferromagnetic Intermetallic Compounds C.B. Zimm, et.al., Journal of Applied Physics, 15 April 1985, v57 #8, p3829. Magnetic Refrigeration T. Hashimoto, et. al., Solid State Physics (Japan), March 1985, v20 #3, p161-175. Characteristics of Magnetocaloric Refrigerants below 20K T. Hashimoto, et.al., ICEC 9 (Japan), May 1989, pp 26-29. A Composite Material for Magnetic Refrigeration Using Internal Heat Transfer. B. Daudin, et.al, Cyrogenics (Great Britian), September 1982, v22 #9, pp 439-440. Magnetic Refrigeration from 10K to Room Temperature T. Hashimoto, et.al., Cyrogenics (Great Britian), November 1981, v21 #11, pp 647-653. T-S Diagram for Gadolinium Near the Curie Temperature S. Benford, et.al., Journal of Applied Physics, March 1981, v52 #3, pp 2110-2112. The Magnetocaloric Effect in Dysprosium S. Benford, et.al, Journal of Applied Physics, March 1979, v50 #3, pp 1868-1870.
Fig. 4 A FEW OF THE RECENT PAPERS on magnetic refrigeration and the new magnetocaloric effect.
33.3
Hardware Hacker
the virtually unknown Science Citations Index. Unlike all the others, this one lets you move forward through time, rather than back into older and older material. This works by listing who put whom into their end-of-paper bibliographies. For instance, any competent new technical paper on active filters must reference Sallen and Key. Anything new on cold fusion absolutely must list Pohns and Fleishman. Anything new on unfocused solar collectors simply must cite Winston, and so on. If they dont, then they arent worth reading anyhow. Simply shove any of these names through the index, and you generate all of the newer papers in the field. After a while, new author names will start cropping up and repeating. You then use the avalanche effect to find the latest and the best, just by starting with one or two ancient authors. And do not ignore the librarys kiddie, young adult, or popular press books. Excellent, understandable and readable backgrounds can be easily picked up in the Doubleday Science Series, or the Life Science Library. Beyond the library, youll want to collect the specialty direct mail books catalogs. Weve covered this resource in depth in a previous column and contest. More details in my Hardware Hacker II reprints. Lets see. What else can you do? Youll definitely want to set up some sort of an extensive personal network that involves people strictly outside of friends, family, or work associates. Obviously, my help line works like a champ here. Electronic bulletin board systems are another great route to networking. So are clubs. Your own personal experiments can very much clarify any topic, as can teaching a class about it. The purpose of research is to get the effect you are after to show up reasonably well in as simple and as cheap a way as possible. But stay in school forever. While theres lots of possibilities here, the best Ive found are local community college courses, and that self-study material from Heath. Finally, simply let things gel. Take Bowseretta up the mountain. Quest a tinaja. Map that terminal crawlway. Any field has an order and a flow to it. Often in directions which "they" dont care to admit. Start with a few fundamentals, think about it for a while, and a pretty fair picture of the rest may fall in place without much in the way of conscious intervention. Remember that sincerity is everything. Once youve got that faked, all else follows. technological breakthrough, one that is eminently hackable, besides being a sure fire winner for a research topic, school paper, or science fair entry. Only instead of me doing all the work, lets try doing it together. See how much you can improve your research skills along the way. Just show me an easily done and Radio-Electronics compatible method to demo the magnetocaloric effect described below, at room or lower temperatures. Or else add in any way (patents, papers, articles, data sheets, etc) to our ongoing magnetic refrigeration dialog below. Therell be all the usual Incredible Secret Money Machine book prizes, along with an all expense paid (FOB Thatcher, AZ) tinaja quest going to the very best of all. As usual, send your written entries directly to me here at Synergetics, rather than on over to Radio-Electronics editorial.
Magnetic Refrigeration
Theres apparently a brand new way to cool things that is just turning the corner from laboratory work to preliminary product development. If what has happened so far is to be believed, it should completely blow away many of the traditional cooling schemes, particularly for very low temperatures. This genuine new breakthrough is called the magnetocaloric effect, and I have grossly oversimplified it in my figure three. The latest key papers appear in the listings of figure four. Basically, if you take critical rare earth elements or their alloys, they will absorb heat when magnetized and release heat otherwise, acting as a heat pump. At least over some temperature ranges and over specified magnetic field strengths. Gadolinium is one popular material. Heat transfer operations take place in and around the Curie Point. Most magnetic materials lose many of their properties should they exceed their Curie Point temperature. The magnetocaloric effect can be tuned over a range of absolute zero to above room temperature. Efficiencies which are as much as 40:1 better than mechanical refrigeration have been bandied about. Yes, the effect can be done using no moving parts. Obvious applications for magnetic
+ DC OUT
10F
10F +
NE604
CAP GND MUT VCC RSO MAO UAO QIN
1 2 3 4 5 6 7 8
4.7K
+ 33F +6 VDC
100K
0.1 F
Fig. 5 AN AUDIO VOLTMETER having an range of 0 to -80 DBM. Use this one to calibrate microphones and speakers, or as a receiver "S" meter. Output is 0.5 volts at -80 DMB and 5 volts at 0 DBM. Sensistivity is 10 microvolts.
33.4
September, 1990
refrigeration include cyrogenics and superconductivity, the production of liquid gases (especially hydrogen as a fuel), and as Freon replacements for traditional room air conditioners. A few sources for gadolinium and its related rare earths are shown in our Rare Earth Resources sidebar. This is all so new and so hot (Uh better make that so cold) that I dont have too much for you beyond these key papers of figure four. One very detailed and thorough but extremely expensive report on this magnetic refrigeration is now available through Technical Insights. Much more on all this when I actually get the papers and can do some hacking of my own. UPDATE: More info on magnetic refrigeration does appear in Hardware Hacker #35.
RARE EARTH RESOURCES Aesar Box 8247 Ward Hill Haverhill, MA 01835 (800) 343-1990 Alfa Products PO Box 8247 Ward Hill, MA 01835 (800) 343-0660 Cerac PO Box 1178 Milwaukee, WI 53201 (414) 289-9800 ESPI 5310 Derry Agoura, CA 91301 (800) 638-2581 Fisher Scientific 711 Forbes Avenue Pittsburgh, PA 15219 (412) 562-8300 Goodfellow 301 Lindenwood Drive Suite 1 Malvern, PA 19355 (800) 821-2870
An Audio Voltmeter
Signetics has an intriguing cellular radio log amplifier chip in their NE604 and described in their new ap-note AN-1991 . Figure five shows you how to use this chip to build an audio level meter usable over an incredible 0 to -80 DBM range. It accepts audio levels as low as ten microvolts, and should be useful for speaker and microphone testing, as well as for a receiver "S" meter. Reasonably priced and wide range log amps are a rarity, so this is a welcome chip. Be sure to let me know what other uses you can come up with for this gem.
high at $20 per year for what should be a free house organ. The latest issue includes info on solid state visible laser diodes, pagers, phaselock loop circuits, and oscilloscope tubes. The bargain of the month has to be Jerrycos stock #79106 copy machine chassis. This is a wide dual-bin input feeder and registration assembly for $10.50 . As is, its a great electromechanical treasure trove. With some stupendous hacking effort, you might convert this into a fake Kroy Kolor machine, an automatic dual bin laser feeder, or find some unique robotics applications for it. Arlin Shepard of Lazer Products tells me he has at long last found a few more Canon fusing units. We looked at these several columns ago for Kroy Kolor, lamination, and for dry film printed circuit bonding. Arlin is also my favorite source of recharging supplies. He offers toners, wiper pads, seals, pixie dust, and lots more. As weve seen in the past, you can reduce your laser printer toner costs by 20:1 or more by doing your own cartridge refilling. Adobe Systems has released their new black book, otherwise known as the Adobe Type I Font Format. This has all of the previously secret insider stuff on their encrypting, compacting, and the hinting of high quality typography. I do have a few copies on hand here if you need one. Our free mechanical samples this month include Alumilite, an easily castable urethane that sets in three minutes. It easily holds small details. And the Cycle-Flex mechanical drive cables and fittings from CMA. Rounding out the free samples for
today are those Bend-Flex flexible printed circuits from Rogers and the Conpad elastomeric battery clip replacements from Stockwell. A High Current Transistor Design Guide wall chart is available through PowerTech. And Rochester Electronics offers a new inventory guide to discontinued semiconductors. Tec Spec is an outstanding hacker newsletter chock full of easy to build and hands-on stuff, including Van De Graff generators, visible lasers, and do-it-yourself vacuum pumps. A free sample copy is available. UPDATE: Sadly, Tec-Spec ceased publication in August of 1992. Another special interest newsletter is the ongoing Tesla Coil Builders Association quarterly. This one is $20 per year. Turning to my own products, I do have complete autographed sets of all of my book-on-demand published Hardware Hacker II reprints for all my Radio- Electronics columns here waiting just for you. And, for more information on self-directed research, be sure to check into my Incredible Secret Money Machine. Finally, I do have a new and free mailer for you which includes dozens of insider hardware hacking secret sources. Write or call for info. Our usual reminder here that most of the items mentioned appear either in the Names and Numbers appendix or the Rare Earth sidebar. As always, this is your column and you can get technical help and offthe-wall networking per that Need Help? box. The best calling times are weekdays 8-5 in Mountain Standard Time. Lets hear from you. 3
33.5
Hardware Hacker
33.6
Don Lancasters
Hardware Hacker
October, 1990
Using SAW Devices Designing Active Filter Digital Signal Processing Switched Capacitor Filters Unique Digital Thermometer
and the peakedness of the response at or near resonance. The tuning dial on an AM radio is an example of a bandpass filter. A second order high pass filter is shown in figure 1-C . Here, the high impedance of the series capacitor and the small impedance of the shunting inductor attenuates the very low frequencies, creating a double whammy attenuation rate of +12 decibels per octave. At higher frequencies, the capacitors reactance is low and the inductors is high, freely passing the highs without attenuation. At the corner frequency, you will once again get a peaked, smooth, or drooping response depending on your L / C ratio and its damping factor. The bass control on a hi-fi is one example of a high pass filter. Actually, there is no such thing as a true electronic highpass filter, since one of these would also have to pass microwaves, heat, light, and X-rays. At very high frequencies, the circuit strays (such as a capacitor self resonating on its own leads) can alter the high end performance. High pass filters also do tend to be "noisy", since they freely pass all the harmonics of all supposedly rejected waveforms. Compared to a low pass filter, which carries out one or more integrations, a high pass filter does
ur no-charge RE Hardware Hacker helpline sure gets a lot of queries on filters and filtering. So, for this month, I figgured we might go over several filter fundamentals, especially for deciding exactly what kind of filter should get used where
Passive Filters
An electronic filter is some frequency selective network that favors certain frequencies at the expense of others. Filters are normally used to strengthen wanted signals while you try to reject unwanted ones. Traditionally, filters were built by using combinations of inductors and capacitors. These are called passive filters, and three examples appear in figure one. In 1-A , we have an example of a low pass filter. At lower frequencies that inductor appears as a piece of straight wire and the capacitors reactance is extremely high. So dc and lower audio frequencies are readily passed. At very high frequencies, the inductor looks like a high blocking impedance and the capacitor looks like a very low shunting impedance, so all the high frequencies will get strongly attenuated. The one-two punch of this example makes this a second order filter. At low frequencies, the response is flat. At high frequencies, the response will quarter with each doubling of your frequency, falling off at a -12 decibel per octave rate. Just what will happen at the corner frequency with this circuit? That all should depend upon your ratio of the inductance to capacitance. Should you seek out the smoothest possible response, you can adjust the ratio of L to C so that your response ends up precisely 3 decibels down, or roughly 70 percent of the amplitude at the corner frequency. This -3 decibel point could also be called the cutoff frequency. Such a low pass filter with the smoothest possible passband is known as a Butterworth filter. The L/C ratio of this circuit is also
known as the damping. If you end up critically damped, you should get the smoothest possible response. If you are underdamped, you get a rising or a peaked response at your corner frequency. If you are overrdamped, you get a very droopy result. A simple treble control on a hi-fi is an example of a low pass filter. A bandpass filter appears in figure 1-B . At rather low frequencies, your capacitor provides a high reactance, and your response increases at a +6 decibel/octave rate. At the resonance, otherwise called the center frequency, both the reactances cancel, giving you unity gain. At the higher frequencies, the inductor provides a high series reactance, and the response decreases at -6 decibels per octave. The ratio of the inductance to the capacitance sets the damping, which in turn will set the sharpness of your resonant peak. Since very low damping values are most often involved, a factor known as the Q is used instead. This Q, or "quality" factor ends up as the inverse of the damping. The Q is also the bandwidth of the center peak between its -3 decibel points. Note that, no matter how high the Q , the slopes at very low and very high frequencies will stay at +6 and -6 decibels per octave. Your choice of Q only determines the narrowness
34.1
Hardware Hacker
They also lack gain and do require carefully controlled source and load impedances. So, youll want to avoid passive filters at all costs. Unless you happen to be working with extremely low signals, at some rather high power levels, or at very high frequencies where nothing else will do the job. The usual way of filtering stuff today involves
39K
39K x (2-d)
10K IN 0.16
10K
+ OUT 0.16
Active Filters
(A) SIMPLE SECOND ORDER SALLEN-KEY SECTION
39K
6.2K
39K
47K
10K IN 0.16
10K
+ 0.16
10K
10K
+ OUT -
0.16
0.16
Fig. 2 TWO SALLEN-KEY LOW PASS active filters. Most any premium grade of op-amp can be used. Components should be 5% or better.
one of more differentiations, or slope extractions. This also can add to your overall noise. Fancier filter responses are picked up by using additional inductors and extra capacitors to increase the order of your filter. Unfortunately, you can not just "stack up" passive sections. Each in turn has to play a specific and very often a non-obvious part in your overall desired response.
My favorite passive filter book is Weinbergs ancient Network Analysis and Synthesis (McGraw Hill, 1962). The Radio Amateurs Handbook also has loads of good passive filter design info in it, as do most college level circuit or network texts. These days, youll usually try to avoid passive filters like the plague, since they are bulky, expensive, and hard to redesign, calibrate, or adjust.
Fig. 3 THIS SINGLE AMPLIFIER MULTIPLE FEEDBACK bandpass active filter can only be used for Q values of ten or less.
With an active filter, you can use combinations of resistors, capacitors, and operational amplifiers to fake the response youd get from a passive LC filter. Some energy from your power supply is used to substitute for the energy normally stored in an inductors magnetic field. Important advantages of the active filters are that there are no inductors involved, they are easy to design, and they are easy to tune. Because the active filters can also provide buffering and gain, they are far less sensitive to source or load impedances than passive filters. While there have been many different active filter design methods in the past, only two have survived and remain popular today. These are the everyday Sallen-Key single op-amp filters and higher performance State Variable filters which need three or four op-amps for each of their second order sections. Figure 2-A shows you the second order Sallen-Key low pass filter, normalized to a 10K impedance level and a 1 kHz cutoff frequency. The original horses mouth on this was A. Sallen and Keys A practical method of designing RC active filters from the IRE Transactions on Circuit Theory, March 1955, pp 74-85. Their prototype Sallen-Key filters were intended for use with cathode followers, the vacuum tube precursor to a transistor emitter follower. These were designed to work with a unity or slightly lower gain. Which forced you to select weird ratios for capacitors and caused an interaction between the corner frequency and your desired damping. Many years ago, I played around with the Sallen-Key math in detail and came up with a twist that made these filters far simpler to design and
34.2
October, 1990
actually use. I named this the Equal Component Value Sallen-Key filter. In figure 2-A , both capacitors are of identical size and set only frequency. And both resistors are also identical size and also set only the frequency. The op-amp gain independently sets the damping to a chosen value. To scale your frequency, you just increase either resistor pair or either capacitor pair to lower frequency, and vice versa. To scale the impedance, proportionately raise the resistors and lower your capacitors. Full details in my Active Filter Cookbook. Usually, you combine, or cascade second order sections to build up any fancier filter of higher order. By very carefully selecting the damping value and corner frequency of each second order section, you could produce an overall filter response of most any desired complexity. Some of these fancier responses do have specific names. Ferinstance, a Bessel filter will give you the flattest possible time delay at the expense of the amplitude falloff. As weve just seen, a Butterworth filter could give you the smoothest possible amplitude response. If you are willing to allow some lumps in the passband, you end up with a Chebycheff filter. Popular Chebycheff filters can get built up with one, two, or even three decibels of passband ripple. Finally, the Cauer or Elliptical type filter will also add "zeros" to your stopband, giving you the fastest possible falloff just outside the passband. One thing which many beginning hackers miss: If your filter is made up of cascaded and isolated second order sections, the only difference between all of these filter types lies in how you select your damping and the corner frequency of each section. Thats all. Thus, once you have a filter built, all you need do is change the damping and the corner frequency for each section to get any desired result from Bessell on through Chebycheff. As with passive filters, you do not simply pile up identical second order sections. Instead, you must carefully select your damping and corner frequency of each section to properly interact with the others to get the final desired response. Figure 2-B shows a fourth order Butterworth low pass filter having a 1 kHz cutoff frequency with a -24 decibel per octave rolloff. Note that the two sections are not identical. Note also that a dc return path to ground absolutely must be provided back through the input source for a low pass active filter to work. It has been pretty much proven to everyones satisfaction that you can not build up a high Q active bandpass filter using only a single op-amp. The impedance levels, interactions, and gain requirements all gang up on you as you raise the Q. Figure three shows you the best of a sorry lot of single op-amp second order bandpass filters. This one gets called a multiple feedback circuit. It is only recommended for Q values of 10 or less. Thus, while useful for such tasks as an equalizer or a psychedelic lighting filter, it is strictly limited to lower Q uses. The higher performance workhorse second order active filter is called a State Variable Filter and can be built using three or four op-amps as shown you in figure four. This one has three separate lowpass, bandpass, and highpass outputs, is very well behaved, and can work with very high Q or low damping values. A fourth op-amp can be added to the circuit shown to make the gain independent of the damping. Finally, figure five shows you a Sallen-Key single op-amp highpass circuit. This one is just the lowpass circuit of figure two "insided out" to get the complementary response. Active filters get restricted by the frequency response, the slew rate, the distortion and noise floor of the op amp used. They are best used at audio and lower video frequencies having signals which are neither very small nor excessively big. Quality op-amps for active filters are available from Linear Technology, Burr-Brown Inc., Analog Devices, Maxim, National and PMI. By one of those utterly astounding coincidences that seem to infest this column, Ive written an Active Filter Cookbook that somehow has gotten up to its fifteenth printing. It does include everything you need to build all your own real-world active filters. You could write or call me for an autographed copy or two.
10K 10K IN
3.3K
Fig. 4 THIS STATE VARIABLE ACTIVE FILTER gives lowpass, bandpass, and highpass outputs and is best used for high Q or low d applications.
34.3
Hardware Hacker
properly switched capacitors, you can favor selected frequencies and reject other ones. The big advantages for switched capacitor filters are their electronic tunability and minimum cost. In the past, switched capacitor filters were noisy and often produced significant distortion. Todays circuits are better. They still do not handle extremely small input signals very well. Important uses for these switched capacitor filters are modems, touchtone detection, telecom in general, disk drives, CD players, industrial instrumentation, and data acquisition. Several large suppliers of switched capacitor filters now include Linear Technology, Maxim, and Exar. All three sources have lots of data sheets and ap notes available. A digital signal processor is just a specialized type of microcomputer. To do a filtering using digital signal processing, you first analog-to-digital (A/D) convert your input signal into strings of numbers in memory. Then you use digital techniques to modify those numbers in memory. Finally, you digital-to-analog (D/A) convert your filtered results. While this may sound like a real runaround, these days you can digital signal process in one single and reasonably priced chip. There are many big advantages to DSP filtering. One big advantage is that you can easily create filters that can be extremely difficult or impossible to handle with passive or active analog techniques. Examples include "brickwall" filters having near-infinite response slopes, and filters with a constant phase or a controlled group delay. Another major advantage is a full programmability. Nothing changes except several software words if you want to completely change what your filter does or how it does it. Most of those DSP chips are really specialized microprocessors. Usually they have powerful internal firmware commands which can let you rapidly multiply, shift, and add. They also usually provide specialized functions such as barrel shifting and zero time testing and branching. Fixed point and floating point processors are both available today. DSP is obviously limited to lower frequencies unless you do not require real time results. You also have the usual aliasing problems and the A/D quantization noise limitations to cope with as well. Important DSP chip uses include geophysics, biological research, for speech synthesis and recognition, radar, and electronic music. Just about all of the latest personal computers and synthesizers have gone the DSP route for most of their sound capabilities. The main manufacturers of DSP chips do include Analog Devices , Texas Instruments, Motorola, and TRW. All of these folks do have data books and detailed ap notes available. Surface Acoustical Wave, or SAW filters are a specialized type of low cost filter which is quite popular in television and cable systems. These consist of a piezoelectric disk which has an acoustical transmitter and an acoustical receiver on it. Unlike your usual audio devices, these acoustical waves do involve sound waves whose frequency typically lies in 40 to 400 Megahertz range. Special finger-like contacts can be placed between the transmitter and receiver. These fingers cause constructive and destructive interference patterns that give the SAW device a specified response. Typical uses are for TV vestigal sideband filtering, for various descrambling circuits, and to set your IF amplifier passband in tv receivers. The big advantages of SAW filters are that they are very cheap, accurate, stable and need no tuning. Sadly, the SAW filters dont really hack all that well, even though you can readily find them for a buck each surplus. Any SAW filter handles one and only one specified job, done in a carefully specified manner at one specified frequency. And into one specified load from a specific source. They are not in any way tunable or adjustable. What you got is what you get. And your own custom SAW filter involves a very steep setup charge, comparable to that of a full custom integrated circuit. One leading supplier for the SAW filters is Plessey Signal Technologies, while bunches of surplus sources do advertise in Nuts and Volts. Let me know if you need any more details or use circuits for any of these alternate filter technologies. This months resource sidebar is on active filters and shows you many of the names and numbers you will need to get started.
A Digital Thermometer
Radio Shack now has a amazingly sophisticated digital thermometer that sells for less than twenty bucks. It is part number 277-123. This beauty runs off a single AA battery and displays four digit LCD temperature in either Farenheit or Centigrade over a -40 to 122 F range. Incredibly, there is a recording max-min feature and a settable alarm and controller capability. Both on-off levels and a 2 kHz piezo tone do get provided as outputs. You have a choice of a one second or a fifteen second update for your display. Total operating current is in the microamp range. Their out-of-the-box accuracy is only a degree or so Farenheit, but this
20K
20K x (2-d)
0.16 IN 10K
0.16
+ OUT 10K
Fig. 5 THE SALLEN-KEY HIGH PASS second order filter section is simply the lowpass circuit of figure 2-A insided out. Values are for 1 Kilohertz.
34.4
October, 1990
circuit seems stable enough that you just might custom calibrate it for such things as a hot tub controller or for cave stream temperature logging. The unit-to-unit variations are also within a degree. The resolution is a tenth of a degree and appears real. Although not waterproof, this gets easily fixed with a tube of silicon bathtub caulk. As received, the device displays Centigrade at fifteen seconds per update. To select Farenheit with a one second update, just jumper pins 13, 14 and 16 on the edge connector. Well probably look at this gem in more detail in a future column. For now, lets make a contest out of it. We will have all the usual Incredible Secret Money Machine book prizes, along with an all-expense-paid (FOB Thatcher, AZ) tinaja quest for two going to the very best of all. Either (A) show me a new and unusual application for a miniature and micropowered max-min thermometer; or else (B) show to me some simple and clean way to get all of the digital data out of the 277-123 into a remote computer. As usual, be sure to send your written entries directly to me, and not to Radio-Electronics editorial.
ACTIVE FILTER RESOURCES Analog Devices One Technology Way Norwood, MA 02062 (617) 329-4700 Burr-Brown 6730 South Tucson Blvd Tucson, AZ 85706 (602) 746-1111 Exar Box 49007 San Jose, CA 95161 (408) 732-7970 Linear Technology 1630 McCarthy Blvd Milpitas, CA 95035 (408) 432-1900 Maxim 120 San Gabriel Drive Sunnyvale, CA 94086 (408) 737-7600 Motorola Box 20924 Phoenix, AZ 85036 (800) 521-6274 National 2900 Semiconductor Drive Santa Clara, CA 95051 (408) 721-5000 Nuts & Volts Box 1111 Placentia, CA 92670 (714) 632-7721 Plessey Signal Technology 9630 Ridgehaven Court San Diego, CA 92123 (619) 571-7715 PMI 1500 Space Park Drive Santa Clara, CA 95052 (408) 727-9222 Synergetics Box 809 Thatcher, AZ 85552 (602) 428-4073 Teledyne Semiconductor 1300 Terra Bella Avenue Mountain View, CA 94043 (415) 968-9241 Texas Instruments PO Box 1443 Houston, TX 77001 (800) 232-3200 TRW Box 2472 La Jolla, CA 92038 (619) 475-1000
An Overview of SCSI Concepts is a great new manual available free from Peer Protocols. Three off-the-wall odd magazines include Extraordinary Science, Surplus News, and Science Frontiers. Free software demos this month include that H8SX calculator disk from Hewlett-Packard, and the 945 heater controller demo from Watlow. A free mechanical sample of GUR is available from Hoechst. This is a new plastic which is tough, cheap, self-healing and slippery. A pair of data books and a postcard samples order form from Maxim includes Analog Product Highlights and their New Releases Data Book.. From PMI, a humongous Analog Integrated Circuits Data Book #10. Four pounds and several thousand pages. Included are all of the old SSM (Solid State Music) chips. The latest flyer by All Electronics includes $6.50 strip chart recorders and $15 COSMAC microcomputer trainers. For those of you who came in late, the CDP1804 COSMAC archi-
tecture gave totally new meaning to the term bizarre. A most wondrous beast for sure. Turning to my own products, for the fundamentals of digital integrated circuits, check into either my TTL Cookbook or CMOS Cookbook. And for those of you interested in that PostScript language or in Book-onDemand publishing, our ongoing RoundTable is doing fine on Genie. By the time you read this, over 1000 free downloads should be available. Finally, I do have a new and free mailer for you which includes dozens of insider hardware hacking secret sources. Write or call for info. Our usual reminder here that most of the items mentioned appear either in the Names and Numbers appendix or in the Active Filters sidebar. As always, this is your column and you can get technical help and offthe-wall networking per that Need Help? box. The best calling times are weekdays 8-5 in Mountain Standard Time. Lets hear from you. 3
34.5
Hardware Hacker
34.6
Don Lancasters
Hardware Hacker
November, 1990
ome of you hardware hackers seem to make a big deal out of getting one or two pieces of newer electronic parts. As a general rule, if a chip house has a stiff minimum order, they often will be rather generous with free evaluation samples. All you usually have to do is ask for them in a professional manner. Other times, some other hacker will recognize a need and fill it for you. For instance, Mike Giamportone of his Thumb Electronics is stocking the LSI melody chips for you at $1.50 each. Many of the other great LSI hacker chips are available in small quantities from Belco Electronics. My new PostScript BBS is already going great guns on Genie. Yes, all of the "lost" Guru columns are here, as is much of the Midnight Engineering stuff. Youll find my insider hacker secrets brochure available as library downloads #112 and #113 . You can also reach me here for email and such as SYNERGETICS. All this on Genies PSRT, and short for the PostScript Roundtable. For local connect info, call Genie (voice) at (800) 638-9636. While I was wandering around a few of those other RoundTables on Genie, I did find a great radio and electronics board up at M345 . These folks are heavily into monitor and scanner listings, satellite stuff, unusual receiver circuit hacks, general communications, and bunches more. They have thousands of downloads available for you. Lets start off with some unique circuits known as
Using Vortex Coolers Polyphase Generators Walking Ring Counters Magnetic Refrigeration Electronic Trade Journals
ring counter, the following sequence is generated 00000 10000 11000 11100 11110 11111 01111 00111 00011 00001 0 0 0 0 0 (repeats) Since we repeat after ten counts, we apparently have built a decade, or divide-by-ten counter. In fact, it turns out that the count length of a walking ring counter is normally twice your number of stages in use. Thus, three D-flops give you a divide-by-six, four a divide-by-eight and so on. This is a synchronous counter, since all stages are clocked simultaneously. There are none of the glitches that get produced by a ripple counter. Note that there are no intermediate frequencies. You have your input frequency and an output square wave of frequency input/2n, where "n" is your number of stages. Which makes the walking ring counter ideal for very high frequency use, especially as UHF and VHF prescalers. Another unique property is that all of the counter states can be decoded using only a single two-input NAND
some strange reason, Apple does not appear to be shouting this fact from their Cupertino rooftops. Since no good deed ever will go unpunished, elegant simplicity will always get ruthlessly stomped upon whenever and wherever it occurs in the real world. One older and classic example of elegant simplicity is the Walking Ring Counter. This is a hardware circuit or a software routine which generates a unique count sequence in a simple, unusual, and quite sophisticated manner. When used in the right place at the right time, The walking Ring Counters can provide very clean solutions to electronic problems that otherwise are often sloppily or klutzily done. Walking ring counters are also sometimes called Johnson Counters. To build a walking ring counter, take a string of type D flip flops or adjacent bits in a software word. Cascade the flip flops as an ordinary serial shift register, but connect the Q output of the last stage back to the D input of the first stage. Then you clock all of the flip flops from a common source. Figure one shows you a three stage walking ring counter well be looking at shortly in more detail. For now, pretend this one is a five stage counter instead of being only three stages long. Further, assume we have reset the counter to its 00000 state. As you clock a five stage walking
240
120
Fig. 1 A THREE PHASE REFERENCE GENERATOR using a walking ring counter. The three output square waves produced are at one sixth the clock frequency and are precisely shifted by 120 and 240 degrees.
35.1
Hardware Hacker
divide-by-64 gets old in a hurry. Thus, walking ring counters are rarely used beyond a divide-by-twelve. Any flip flop counter of any length can get into 2^n possible states. Thus, a five flip flop walking ring counter can get into 32 possible states, only ten of which are legal and allowed. If you inadvertently get into one of those "wrong" count states, you can get into a disallowed sequence that can cause you lots of grief. As a ferinstance, the three stage, divide-by-six counter of figure one normally provides this sequence 000 100 110 111 011 001 0 0 0 (repeats) But if you ever get into the 0 1 0 sequence, heres what results
Fig. 2 ELECTRONIC DIE uses a walking ring counter in an unusual 1-3-5-6-4-2 sequence to greatly simplify the spot decoding.
+V
D C
Q _ Q
D C
Q _ Q
4,5, or 6 if low D C
not 1 if low
odd if low
6 if low
Q _ Q
TO SECOND DIE
010 1 0 1 (repeats) All of a sudden you now have a divide-by-two counter instead of the divide-by-six. Which could really be bad news. Thus, you somehow have to guarantee that a walking counter never gets into any disallowed states!. Resetting the counter or else preloading your software registers are two good ways to prevent disallowed sequences. A sledgehammer cure is to always force internal 1 states whenever the two ends of the counter are also both set to 1. For instance, you could AND decode the outside Q outputs and use these to preset all the internal stages to their 1 state. This will quickly find any disallowed sequence and repair it on the fly. Lets look at three uses for walking ring counters. I cannot believe how many really dumb circuits have been published in the trade journals that were intended to generate three phase power control signals. As figure one shows us, all you really need is a divide-by-six walking ring counter. Provide six times your desired frequency, and you get three square waves out, all phase shifted by precisely 120 degrees. You can obviously change the total number of stages for a quadrature output, for quad phase, five phase, etc.
gate per state. Further, a full decode uses all of the Q and Q outputs in a balanced and symmetrical manner. Back in the days when you worried about the fanout of a logic block, this was a really big deal. Ill save the details on this for you to ponder over on your own, but most states decode using the Q output of one stage NAND ed with the Q output
IN-PHASE SINEWAVE OUT
of an adjacent one. Again in a very clean layout that is superb for high frequency use. What is wrong with the walking ring counter? Elegantly simple or not, it does have two bad habits. The first is that you need half as many flip flops or word bits as your count modulo. Using five flip flops to divide by ten seems no big deal. But using 32 of them for a
QUADRATURE SINEWAVE OUT
31K
22K
31K
31K
22K
31K
8X INPUT CLOCK
D C
Q _ Q
D C
Q _ Q
D C
Q _ Q
D C
Q _ Q
Fig. 3 QUADRATURE SINEWAVES which are phase shifted by precisely 90 degrees are easily generated with a walking ring counter. The uses include single sideband communications and synchronous demodulation.
35.2
November, 1990
More details on all of this in my CMOS Cookbook. Another great place to find lots of really klutzy circuits is in an electronic dice game. Figure two shows you the elegant simplicity of a walking ring die. To understand this circuit, you first define your die spot count sequence as 1-3-5-6-4-2 . This gives you a "free" even-odd decoding for the middle spot, and a "free" 4-5-6 decoding for two diagonal outside spots. Use a NAND gate to decode "not 1" for the other two diagonal outside spots. Use an AND gate (or else two diodes) to decode "six" for those two outside middle spots. A medium frequency gated oscillator clocks the die, and a second die can be cascaded off the first one. Finally, figure three shows you a digital sinewave generator that produces a pair of sinewaves frequency shifted precisely by 90 degrees. This circuit is based on stuff we looked at back in the January 89 RE . See my Hardware Hacker II for details. Ive purposely left the part numbers off all our figures this month, to encourage you to experiment on your own. Your best choices will be such chips as a 4013 or 74HC74 dual Type D flip flop. Most any parallel out shift register could also be used, as could software shift commands applied to any old alterable byte. For our contest this month, either (1) Show me an unusual use for a walking ring counter, or else (2) show me some other circuit, tool, or technique that is elegantly simple. There will be all of those usual Incredible Secret Money Machine book prizes for the dozen top entries, with an all expense paid (FOB Thatcher, AZ) tinaja quest going to the best. temperatures. The best paper Ive seen so far is the Magnetocaloric Effects in Rare-earth Magnetic Materials, authored by A.S. Andreenko and crew in Soviet Physics Usp 32 (8), August 1989. Since an all solid state cooler would involve switching extremely strong magnetic fields, the chances are that some motion will be involved in most of the practical magnetic coolers. Figure four shows you a shuttle type cooler. Gadolinium is moved to an area free of a magnetic field where it picks up heat from the object to be cooled. It is then moved into a very strong magnetic field where it dumps energy to a suitable heat sink. This cycle is repeated at an optimum rate for the quantities of heat involved. Rotary coolers and Sterling Cycle coolers have also been proposed. See the above paper for more details. And let me know any new resources you run across on this. like a perpetual motion machine, it is quite real and rigorously obeys the laws of thermodynamics. As figure five shows us, a vortex cooler is nothing but a very specially designed nozzle. The input is connected to a source of compressed air, as is commonly found in most production areas. There are two outlets. Hot air comes out the hot end and cold air comes out the cold end. Leading manufacturers of vortex coolers are Vortec and Exair. They both have lots of catalogs, cassette tapes, data sheets, and aplication notes available. How does it work? The air enters tangentally at faster than the speed of sound and forms a rapidly spinning hollow vortex cylinder having a very high angular velocity. When it gets to the hot end, some of it is released to ambient, and the rest of it spirals back inside the spinning hollow inlet air vortex cylinder. Now for the tricky part. The inlet air and the outlet air are both spinning at the same radial velocity, since they are in contact with each other. But the momentum of the inside air has to be less since it has a smaller radius of rotation, and since angular momentum is inversely proportional to the spinning radius.
Vortex Coolers
This is another type of "gee whiz" type of cooler that is useful for cooling electronic enclosures and doing such specialized things as preventing any needle breakage in heavy duty sewing machines. A vortex cooler has zero moving parts. While it looks and feels
HEAT SINK TO AMBIENT VERY STRONG MAGNETIC FIELD ZERO MAGNETIC FIELD
(A) Gadolinium block picks up heat and has its own entropy raised at the heat source. The block is then moved.
(B) Shuttled block releases its own heat energy and has its entropy lowered at the heat sink. The block is then returned.
Fig. 4 THE SHUTTLE TYPE MAGNETIC REFRIGERATOR moves a block of Gadolinium between a heat source in a magnetic field free area and a heat sink in a quite strong magnetic field area. This type of heat pumping can end up extremely efficient, especially at cyrogenic temperatures.
35.3
Hardware Hacker
air supply. And your air has to be extremely clean and extremely dry if the cooler is not going to freeze or jam. These can be noisy. And the pricing in single quantities often approaches $100 for industrial units. Question: Could you put a vortex cooler in a car vent? Lets have your thoughts on this.
ELECTRONIC TRADE JOURNAL RESOURCES Appliance 1110 Jorie Blvd, Oak Brook, IL 60522 (708) 990-3484 Circuit Design 1790 Hembree Road Alpharetta, GA 30201 (404) 475-1818 Circuits Manufacturing 500 Howard Street San Francisco, CA 94105 (415) 397-1881 Defense Electronics 6300 S. Syracuse Way, Englewood, CO 80111 (303) 220-0600 Design News 275 Washington Street Newton, MA 02158 (617) 964-3030 EDN 275 Washington Street Newton, MA 02158 (617) 964-3030 EE Times 600 Community Drive Manhassat, NY 11030 (516) 365-4600 Electronic Comp. News One Chilton Way Radnor, PA 19089 (215) 964-4345 Eletronic Design 611 Route #46 West Hasbrouck, NJ 07604 (201) 393-6060 Electronic News 7 East 12th Street New York, NY 10003 (212) 741-4230 Electronic Pack & Prod 1350 East Touhy Avenue Des Plaines, IL 60018 (708) 635-8800 Electronic Products 645 Stewart Avenue Garden City, NY 11530 (516) 227-1300 Electronics 611 Route #46 West Hasbrouck, NJ 07604 (201) 393-6060 Machine Design 1100 Superior Avenue Cleveland, OH 44114 (216) 696-7000 Measurement & Control 2994 W. Liberty Avenue Pittsburgh, PA 15216 (412) 343-9666 Motion PO Box 6430 Orange, CA 92613 (714) 974-0200 PCIM 2472 Eastman Avenue Ventura, CA 93003 (805) 658-0933 PCNetwork 30 Moran Street Newton, NJ 07860 (201) 383-2296
Because the momentum must be preserved, energy is transferred from the inside spinning air cylinder to the outside one. This cools the cold end exit air as it heats the inlet air. Full details appear in the Vortec application notes. There is usually an adjustment screw that lets you vary the ratio of hot to cold exit air. One setting gives you
the lowest possible exit temperature, which routinely can go as low as -40 degrees. A different setting gives you the most efficiency and the maximum cooling at a higher exit temperature. The air consumption and the overall efficiency are acceptable for most current uses. What are the limitations to vortex cooling? Obviously, you do need an
Fig. 5 A VORTEX COOLER has no moving parts, yet easily can provide an air stream as cold as -40 degrees. The key secret is a pair of coaxial vortex streams. In order to preserve angular momentum, heat energy gets transferred from the cold output stream back to the input air stream.
35.4
Don Lancasters
Hardware Hacker
November, 1990
id I really do that? Well, I will admit I thought I had made an error once, but it turned out I was mistaken. At any rate, most of the op-amps in my October 90 RE Hardware Hacker column on active filters seem to be upside down. All opamps in all figures should have their (-) or inverting input on the top and the (+) or non-inverting input on the bottom. Figure three is correct; figures two, four, and five do need revised. All of the illustrations in my Active Filter Cookbook are correct. Sorry about that. Since all of my technical illustrations are submitted as "shoot-and-go" camera ready PostScript, I did it all by myself. As all of you long-term Hardware Hackers already know, all the artwork you see in this column gets done with the AppleWriter word processor with my Apple IIe . No scanners or artwork generation packages of any kind are used; everything gets done directly with "raw" PostScript. That includes all the schematics, pictorials, charts, text boxes, all printed circuit layouts, and all of the perspective and isometric stuff. Our superb new PostScript PSRT RoundTable (M835) is going great guns on GEnie, and youll also find some advance Hardware Hacker and Midnight Engineering info there for your downloading. You could also contact me there via email by using my [SYNERGETICS] address. You can give GEnie a voice call at (800) 638-9636 for connect info. They recently have reduced their rates and have expanded their many services. Nearly 100,000 library downloads are now available in one of the finest and highest quality shareware and public domain collections in the world. Just about all computers do get supported, major or minor. My thanks to all of you perpetual motion builders for your thoughtful suggestions on how to shield, time, or rearrange all the magnets to get my energy producing gyroscope (Sept. 90 RE) to run, rather than simply latching
Discrete Cosine Transforms Digital Sinewave Generator More on Video Compression Santa Claus Machine Update Audiovisual Switcher Circuit
rather acrylic aftertaste to your roast beef sandwich. But the hacking opportunities here do boggle the mind. Firstoff, todays machines seem obscenely overpriced. Second, the best possible prototyping method clearly does not exist in any of the crude approaches in use today. Third, and finally, there are all sorts of service bureau, rental, and franchise opportunities emerging. There is no reason whatsoever why any hacker $200 solution to desktop prototyping should not exist. And, needless to say, all the folks here at Radio-Electronics editorial will pay very well for a hacker friendly one. The reasons to be desktop prototyping are obvious. Just like you can now instantly produce a Book-onDemand home published volume, you
up. And an extra special thanks to the few of you that sent me all of those super-strong magnets to try. Theres lots of goodies this month. Lets start off with a
4.7K 10F right audio left audio video 20F 75 right audio left audio video right audio left audio video right audio
(A)
(I)
(B)
right audio left audio video
(II)
(C)
20
19
18
17
16
15
14
13
12
11
LO1 LI1 LI2 LI3 L02 R02 RI3 RI2 R1I RO1
HAA11508
VI1 GND VI2 CT1 VO1 VCC VI3 CT2 VO2 MUT
1 2 3 4 5 6 7 8 9 10
(inputs)
+12V CTRL1 CTRL2 MUTE
(outputs)
Fig. 1 THIS TRIPLE AUDIOVISUAL SWITCHER uses the Hitachi HAA11508 to simultaneously select video and two channel stereo audio from any one of three sources and then route them to one of two selected outputs.
36.1
Hardware Hacker
system uses a tank of an uv-sensitive photopolymer. An elevator sits in the tank and is initially just below the surface. Their ultraviolet laser will selectively harden the photopolymer, hardening it out onto the elevator. Then the elevator gets lowered a few mils and the process repeated. Pancake style, a composite object gets built up on a layer-by-layer basis. Compound curves and tricky undercuts that can be impossible to do the old way are easily done. While the uv photopolymer used is normally both expensive and hard to get, a possible hacker substitute could be the Merigraph photopolymer used for rubber stamps, and available by way of Grantham or R. A. Stewart. A second generation visible light photopolymer is getting developed at Dupont, and one major system which uses these is Quadrax. Advantages of the visible photopolymer include its lower viscosity, the ability to use a cheaper laser, and (with proper safety precautions) being able to see what you are doing. Because of the lower viscosity, you can also raise the liquid level, rather than actually moving what you have already produced. While these liquid vat methods do work, they are limited in accuracy to several mils, are rather costly, and end up severely restricting your choice of materials. A fresh, different and totally dry approach to desktop prototyping gets used by DTM. Here a granular powder of certain waxes or any of a number of plastics is selectively hardened by an infrared laser in a process known as sintering. Sintering simply melts the surface of the granules so they stick together. In addition to being cleaner and simpler, you have a wide variety of possible material options, including real metals. Many powders can be reground and recycled, and they are far cheaper than the photopolymers. Lets see. A few loose ends. Other approches to Santa Claus Machines include photochemical machining of thin layers and various other "gasket" techniques that get bonded together to produce composite objects. One think tank doing active instant prototype development is Batelle, while conventions and seminars are sometimes
CTRL1
CTRL2
OUT I
OUT II
0 0 1 1
0 1 1 0
A A B B
A C B C
Fig. 2 THE TRUTH TABLE for the triple audio switcher. Logic signals on the control inputs decide which input gets selected. An optional MUTE input turns off the audio only on output II.
should eventually be able to create virtually any object on a desktop. At pretty much the same cost and speed. All at development expenses and a time-to-market that are only a tiny fraction of traditional machine shop approaches. And you should even be able to fake teleportation, simply by sending the replication codes to a remote machine over a satellite, modem, or fax line. Or, someday, a Playmate of the month on CD ROM. Yes, the real thing. No holograms need apply. Our Santa Claus Machine resource sidebar for this month can show you where to go for more information on this exciting new field. There are several dozen approaches to desktop prototyping, but those with the most commercial success so far are
limited to lower end CAD/CAM, ultraviolet stereolithography, visible light stereolithography, and laser sintering. Lets look at each in turn The low end CAD/CAM can involve miniature computer controlled milling machines or lathes, and the parts are produced in the usual way by using a cutting program. Today, most of these machines are aimed at the educational market. Sources include masterCAM, Roland Digital, and Light Machines, while others do advertise in Industrial Education and School Shop. One material often used is a special machinable wax now available from Freeman Supply. This product is low cost, does not wear your tools, and can be recycled. The pioneer in ultraviolet stereolithography is 3-D Sytems. The 3-D
A Fast Algorithm for the Discrete Cosine Transform C.H. Smith, IEEE Transactions on Communications Sept. 1987 pp 1004-1009. Image Compression for High Speed Network Transmission C. Ciracia, Circuit Cellar Ink, Aug/Sep 1990, pp 19-26. Teleconferencing K. Rao, Van Nostrand Reinhold Company, 1985. Digital Coding of Waveforms N. Jayant, Prentice Hall, 1984. A New Wave in Applied Mathematics B. Cipra, Science 24 Aug 1990, pp 858-859. Survey of Adaptive Image Coding Techniques A. Habbibi, IEEE Transactions on Communications, Nov 1977, pp 1275-1284. Advances in Picture Coding H. Mussmann, Proceedings of the IEEE, April 1985, pp 523-548. Progressive Transmission of Gray-Scale and Binary Pictures K. Knowlonton, Proceedings of the IEEE, July 1980, pp 885-896. Predictive Coding Based on Efficient Motion Estimation R. Srinivasan, IEEE Conf. Communications, May 14 1984, pp 521-526. DCT Processing of NTSC Composite Video Signals A. Ploysongsang, IEEE Trans. on Communications, March 1982, pp 462-479.
Fig. 3 A FEW OF THE RECENT PAPERS on video image compression and the DCT discrete cosine transform.
36.2
November, 1990
done by the CAD/CIM Roundtable. One of many souces for traditional CAD/CAM support software appears to be Control Automation. One thing thats totally obvious to me the PostScript language will play a major role in the future of desktop prototyping. Obvious reasons here are PostScripts nearly total device independence, its incredible graphical performance (especially in all three dimensions), and its ability to let your favorite and el-cheapo word processor completely and totally blow away virtually any of todays costly CAD/ CAM application packages. For our contest this month, just add to our Santa Claus Machine dialog in some useful way. There will be all of the usual Incredible Secret Money Machine book prizes going to the dozen or so top entries, with an all expense paid (FOB Thatcher, AZ) tinaja quest for two going to the very best of all. As usual, please send your written entries to me here at Synergetics, instead of to Radio-Electronics. All the inputs and all outputs are supposed to get capacitor coupled. Since the switching times and all the overlaps are not specified, this chip appears mainly intended for use in static switching or source selection. Instead of being applied to chroma key or picture insertion applications. Naturally, for any higher quality switching, full double sided pc boards and lots of ground plane are an absolute must. zero, leaving a sparse data set. Put another way, what you dont know (or see) wont hurt you. The sparse data set then gets further compacted using Huffman or similar techniques. The net result is a video compression in the 30:1 range without significant visual degradation. The reasons for compression is to reduce your disk storage time and your communication time. Far more importantly this sort of data compression is essential to squash a HDTV high definition video image on down into reasonable bandwidth channels. So knowing and understanding video compression is an essential part of the new multimedia revolution that is combining video and computing. Two sources of DCT chips include C-Cube and SGS, while some useful (but very much slower) do-it-yourself software routines are shown in the Ciarcia article.
ML2035
VSS SCK SID LAT
1 2 3 4
Fig. 4 A DECEPTIVELY SIMPLE DIGITAL SINEWAVE GENERATOR. This $8 chip does a full 21 bit phase addition and a 512 level table lookup for an outstanding quality. Resolution is one hertz over a 1 Hertz to 25 kHz range!
36.3
Hardware Hacker
As a second contest for this month, just tell me something different you would do with this new ultra-cheap, ultra-simple, and ultra-precise digital sinewave generator.
SANTA CLAUS MACHINE RESOURCES Batelle 505 King Avenue Columbus, OH 43201 (614) 424-7782 CAD/CIM Roundtable 1050 Commonwealth Avenue Boston, MA 02215 (617) 232-8080 Control Automation 2350 Commerce Park Dr NE, #4 Palm Bay, FL 32905 (407) 676-3222 DTM 1611 Headway Circle B2 Austin, TX 78754 (512) 339-2922 DuPont 1007 Market Street Wilmington, DE 19898 (302) 774-1000 Freeman Supply 1246 West 70th Street Cleveland, OH 44102 (800) 321-8511 Grantham/Polly-Stamp 418 Central Avenue NE East Grand Forks, MI 56721 (218) 773-0331 Industrial Education 1895 Crooks Road S135 Troy MI 48084 (313) 649-4900 Light Machines 669 East Industrial Drive Manchester, NH 03103 (603) 625-8600 masterCAM 2101 Jericho Turnpike New Hyde Park, NY 11040 (516) 328-3970 Merigraph/Hercules 300 East Shuman, Ste 260 Naperville, IL 60566 (800) 323-1832 Quadrax 300 High Point Avenue Portsmouth, RI 02871 (401) 683-6600 Roland Digital 7200 Dominion Circle Los Angeles, CA 90040 (213) 685-5141 School Shop Box 8623 Ann Arbor, MI 48107 (313) 769-1211 RA Stewart 641 South Palm, Unit H La Habra, CA 90631 (213) 690-4445 3-D Systems 26081 Avenue Hall Valencia, CA 91355 (805) 295-5600
It can easily produce any sinewave from 1 Hertz up to 25 Kilohertz in one hertz steps, and its serial interface mates beautifully with virtually any computer or microcontroller. Figure four shows the extremely simple circuit. A split supply of +5 and -5 volts is needed. The output frequency resolution will be fout = fxtal / 8,388,608 Thus, you would use the 8.388608 Megahertz crystal to get a one Hertz resolution. Or you could apply any external frequency up to 12 MHz. Your output sinewave frequency is determined by the 16-bit digital word you last selected. For instance a digital "1" would get you one hertz, while a "5623" (or a hex $15F7) will get you 5.623 kiloHertz. Your output is a clean and low distortion sinewave of around 12 volts
peak to peak amplitude. In order to save on package pins, your digital control word is entered serially. There are three pins involved. Their SID or serial input data pin accepts one data bit at a time, least significant bit first. The SCK or serial clock accepts each new data bit on its rising edge. Finally, the LAT or latch stores the input serial data stream on its falling edge. This should interface beautifully with the game paddle port of an Apple IIe, or anywhere else you have three I/O lines available from a computer or microcontroller. Theres also a ML2036 in a larger package with some extra features at slightly higher cost. One obvious tip: You do input the frequency in binary or in hex, and not in decimal! Dont laugh. This is a very common hacker mistake.
36.4
Don Lancasters
Hardware Hacker
December, 1990
s most of you already know, we now provide a no-charge Hardware Hacker helpline for you as is shown in the book appendix. Here I can try to show you all the products, resources, publications, and consultant services that might be helpful to you. But every third day or so I get a call which starts off with "I cant tell you what Im now working on or what it does..." and then goes on to request something totally absurd. For instance (and no way could I have made these up), a phase detector with a 143 decibel dynamic range, a fishtank full of hot mercury, a microwave traveling wave tube made from a fluorescent lamp, or a matched quad of high power zener diodes. Or a slow blow fuse to get around a circuit that had an obvious short in it. To me, paranoid secrecy on any new idea is both asinine and monumentally stupid. Paranoid secrecy obviously diverts time and energy from improving your product, besides being the inherently evil opposite of what shared hardware hacking is supposed to be all about in the first place. Paranoid secrecy also largely prevents the multi-step, multiperson process that is the essential core of useful product development. Paranoid secrecy tends to grossly overvalue ideas. Back in the golden age of hardware hacking, well thought out ideas were once worth as much as a dime a dozen. These days, of course, ideas are worth less than ten cents a bale in ten bale lots. It is only when those ideas are converted into actual final products being put to work or play by dozens of happy end users that they gain any value at all. It aint creative unless it sells. Paranoid secrecy discourages the thorough third-party testing of your working models and ongoing beta development prototypes. Finally, paranoid secrecy virtually eliminates any possibility of selling your product. By far one of the hardest lessons Hardware Hackers have to learn is that others must come to you
Audio Level Meters A cold Fusion Update Alternates to Patenting Understanding Decibels Wholesale Surplus Sources
I recently received yet another sad case for my overflowing patent victim files. But this ones so apallingly bad that it forms a near perfect example of everything not to do. This new patentee was wondering why the Fortune 500 companies were not beating his door down, and why he couldnt even get anyone to promote his idea. Well, for openers, his patent clearly should never have been issued. On the patent side, it had obviously violated a dozen or more fundamental optical control patents held by a major corporation. On the non-patent side, it describes an obsolete concept totally blown out of the water by an industry standard 1952 technical paper, and thoroughly discussed (and dismissed) in a 1973 student textbook.
for your product to be successful. You cant sell to them; they must come to you. Always. So, by all means give me a helpline call if you are willing to tell me what you are doing and are willing to share it with others. On the other hand, if you want to attempt to lay a paranoid secrecy fantasy on me, Id suggest you try one of those 900 number kinky sex lines instead. They would be in a far better position to help you out.
Patents Again
As weve seen a number of times in past columns, any Hardware Hacker involvement with patents in any way, shape, or form, is absolutely certain to cause you a net loss of time, energy, money, and sanity.
1. Do not ever call yourself an inventor or behave like one. To do sets you up for interminable scams. Instead, you will want to become a product development engineer or a prototyping house. 2. Totally avoid any and all contact with anything even remotely patent related. In any way, shape or form. Do so religiously. 3. Dont bother creating anything in any field in which you are not evenutally certain to become an expert. An expert who is thoroughly familiar with the technical literature, the history of the field, the marketing realities, the insider trade journals, and the mainstream tools and techniques in use. 4. Publish all your key secrets and ideas in a major magazine, leaving out no detail, and omitting no insider secrets. This immediately can generate positive cash flow for you and safely tucks all your ideas away in the public domain, preventing most others from attempting to patent them. This also exposes your new ideas to the widest possible audience. 5. Try to set up some royalty arrangement with a small to medium firm in some position to market and distribute your invention. The tricky part: They must come to you, and never vice versa. That is why it is super important to publish your ideas and creations and expose them as widely as possible. 6. Your main defense against getting ripped off in any royalty setup is the expectation that you will be delivering newer and better stuff in the future. 7. Use the shotgun technique. There is no way that one single idea or product will hack it. To survive in this game, youll need hundreds or even thousands of new ideas and concepts working for you on a total lifetime and total lifestyle basis. 8. Expect to be ripped off. That way, it will be far less of a rude surprise when this inevitably happens to you. And far less of a stress trip. 9. Be realistic. You dont create things to get filthy rich. You create things because you like to create things and have some compelling desire or need to do so. As long as there are enough nickels to keep going, that is all that should really matter.
Fig. 1 SOME THOROUGHLY TESTED and hacker friendly alternates to patents and the patenting process.
37.1
Hardware Hacker
end user or builder of this product ever have been so much as contacted, let alone consulted. I sorely wish this was an isolated example, but it is now one of hundreds in my victim files that collectively have cost all you hackers many millions of dollars. Not once have I ever seen any example of anyone who has profited in any way from any patent involvement from within a small scale hardware hacking environment. So what are the alternatives? Figure one shows you some patent avoidance guidelines that may help you. Weve plowed some of this ground pretty well in the past. But one thing Id like to emphasize here is that you can never develop and market an idea as an outsider. To develop and market something thats useful, you must be thoroughly familiar with that fields technical literature; its history; all its marketing realities; the mainstream tools and techniques; and, above all, the insider trade journals involved. For instance, there is absolutely no way you should be writing forest fire simulation software if you have never sharpened a Pulaski. Working with active filters is pointless if you have never met Sallen and Key. And developing unfocused solar collectors without Winston is unthinkable. I once had a long talk with a retired patent examiner who carefully went back over his thirty years experience and tried to find out just how many individuals acting alone were successful with a strategy of "patent an undeveloped and unproven idea as an outsider and then try to sell it somewhere." The answer: Not one ever! For more on patents and patenting, see The Case Against Patents in the December 1990 Midnight Engineering (now in GEnie PSRT library as #162), or read my Incredible Secret Money Machine book.
60 Decibels is a 1000:1 amplitude gain. 40 Decibels is a 100:1 amplitude gain. 20 Decibels is a 10:1 amplitude gain. 18 12 10 6 Decibels is an 8:1 amplitude gain. Decibels is an 4:1 amplitude gain. Decibels is roughly a 3:1 gain. Decibels is a 2:1 gain.
3 Decibels is about a 30 percent gain and the double power point. 2 Decibels is a 20 percent amplitude gain 1 Decibels is a 10 percent amplitude gain. 0 Decibels is a unity amplitude ratio of 1:1. -1 Decibels is a 10 percent drop. -2 Decibels is a 20 percent drop. -3 Decibels is about a 30 percent drop and the half power point. -6 -10 -12 -18 Decibels is half amplitude. Decibels is around one-third amplitude. Decibels is quarter amplitude. Decibels is eighth amplitude.
-20 Decibels is a 10:1 attenuation. -40 Decibels is a 100:1 attenuation. -60 Decibels is a 1000:1 attenuation.
While the idea purports to be a micropower one, it requires the continuous application of very precisely regulated high energy levels. While it is supposed to be microminiature, a complex mechanism is involved. And multiplication is introduced at crucial
points in the circuit where outstanding linearity is an absolute must. Despite it being something that any of you could easily beat out on the kitchen table, it is obvious that the idea was never tested in a final use application. Nor could any possible
+12V 1K (typ)
KAA2281
AIN AFB A16 A11 A6 A3 A0 GND
1 2 3 4 5 6 7 8
1K (typ) +12V
Fig. 3 A STEREO AUDIO LEVEL METER using the Samsung KAA2281. Levels displayed are -16, -11, -6, -3, and 0 decibels.
37.2
December, 1990
popular ways of dealing with any electronic or physical quantity, linear and log, short for logarithmic. The linear scale has equal steps everywhere in it. Obvious examples include a ruler where your distance between 1 and 2 is exactly the same as the distance from 8 to 9. The channels on the AM radio dial are also linearly spaced with frequency, each being 10 kilohertz above or below its nearest neighbor. Now linear sounds great, and it is the only way to go if you want to prevent any interaction or any multiplication or distortion should two or more quantities of something both be present at once. But, if you attempt to go too far with a linear scale, one end or the other ends up far too cramped or way too spread out. Put another way, the dynamic range of a linear scale or system is often limited. The log system instead works with constant percentage scales. Equal anythings are close together on the "low" end and further apart on the "high" end. For instance, the notes on a musical instrument are usually a tad under six percent above or below one another. Or, more precisely, we are dealing with the twelfth root of two, or 1.059545, since the frequency of your notes doubles as you go up a twelve note octave, or 2:1 frequency. Log systems will often have a much higher dynamic range than the linear ones. And theres lots of places where "equal change" is more important than "equal steps." For instance, rank beginners are often mystified why standard resistors do not simply go 1.0, 2.0, 3.0, 4.0 , instead of 1, 2.2, 3.3, 4.7, 6.8. Obviously, we are dealing with a log scale here. First, because youll normally want to cause a constant percentage change when you alter a resistor in an electronic circuit. And secondly, because stocking different resistors in one Ohm increments up in the megohm range quickly would get out of hand. Some physical systems are inherently logarithmic. In fact, anything linear over a wide range in the real world is usually the exception rather than the rule. The ear is a log system, allowing it to deal with everything from a whisper to a shout over an incredible dynamic range. And since radio began with audio, it is convenient to employ log systems to describe the differences between powerful transmitted signals and weak received ones. All of which leads us around to decibels. Decibels are simply one convenient way of measuring things on a log scale with a potentially wide dynamic range. Since youll usually measure voltage rather than power, most hackers most of the time go with voltage decibels rather than the much rarer power ones. A reading of so many decibels tells you the relative strength between two quantities. For instance, a one decibel change is around a ten percent change in amplitude. It is also roughly the smallest amplitude change you can usually pick up with your ear. A two decibel change is around twenty percent, and a three decibel change is around thirty percent. More specifically, when one signal is at a -3 decibel level from another, (or "three db down"), you will be at 0.707 relative amplitude. Now, since power is related to the square of your circuit voltage, and since 0.707 squared is 0.5, being three decibels down also is at half power from whatever your 0 decibel level happened to be. Hi-fi components normally have their frequency responses specified between their upper and lower -3 decibel points. Put another way, this is the range over which the available power remains over half of what you would get in the "middle" of whatever frequency range youre measuring. Moving right along, a ten decibel change is roughly a 3:1 amplitude ratio, twelve decibels is one quarter, eighteen decibels is one eighth, and twenty decibels equals one tenth. One hundredth is forty decibels, and one thousandth is sixty decibels. The 143 decibel phase detector asked for above is the ratio between one microvolt and fourteen volts, an almost impossible dynamic range to deal with gracefully at reasonable cost in the real world. Ive gathered some of these "rule of thumb" decibel ratios into figure two for your reference. Once again, most all of your decibel measurements are relative and will always refer to two different levels in two points in your circuit. Decibels usually answer the question "How strong is the present signal or response compared to another one?" How strong is a six decibel signal? Somewhere between a femtowatt and a gigawatt, at least sometimes. Maybe. It all depends. All we know for sure is that this particular signal is twice as strong in amplitude as some other one at some other time or place. For an amplifier with a 20 decibel gain, we know that the output will be ten times as strong as the input amplitude, at
KAA2283
AIN AFB A16 A11 A6 A3 A0 GND
1 2 3 4 5 6 7 8
Fig. 4 A MONO AUDIO LEVEL METER can be built by using both halves of the Samsung KAA2283. Levels displayed are -18, -16, -14, -12, -10, -8, -6, -4, -2 and 0 decibels. Sensitivity is from 0.1 to 0.9 millivolts.
37.3
Hardware Hacker
least over the intended working levels of your circuit. Audio engineers will sometimes refer decibels to a very specific voltage level over a specified impedance. They will then use the term DBM to specify such an absolute level. One milliwatt into a 600 ohm line is one possible absolute DBM level. Here you are comparing the ratio of your signal against the standard reference, rather than another signal. But the stereo, hi-fi, and the home recording people have long since bastardized the meaning of the term DBM . Today, unless you are certain you are dealing with the professional studio audio standards, an apparent 0
DBM is just the maximum allowable voltage level at some point in some circuit above which the distortion levels are no longer acceptable. Thus, while your VU , or volume units display on a receiver or recorder does in fact measure some absolute voltage level, the chances are your particular VU units have nothing to do with anyone elses. All they are is a reference level above which distortion will start to become unacceptable. Samsung has a bunch of level meter driver chips that are simple and easy to use. Note that these all differ from a "bargraph" driver in that the bargraph works in linear steps and the level meter driver works in the log steps.
Electrochemically induced nuclear fusion of deturium. Fleishmann, Pons, Hawkins, vol 261 pp 301-308. Examination of nuclear measurements in cold fusion experiments. Abriola, et.al, vol 265 pp 355-360. Observations of the surface composition of palladium cathodes after electrolysis. Mebrahtu, et.al, vol 267 pp 351-357. Electrochemical incorporation of lithium into palladium from aprotic electrolytes. Dalard, et.al, vol 270 pp 445-450. Production of tritium from electrolysis at a palladium cathode. Packham, et.al, vol 270 pp 451-458. A long-term calorimetric study of electrolysis using palladium cube cathodes. Armstrong, et.al, vol 272 pp 293-297. Tritium production during the cathodic discharge of deuterium on palladium. Chene, et.al, vol 280 pp 199-205. Electrochemical fusion: a mechanism speculation. Lin, et.al, vol 280 pp 207-211. Tritium separation during heavy water electrolysis: implications for cold fusion. Corrigan, et.al, vol 281 pp 305-310. Experimental investigations of electrolysis using palladium cathodes. Zahm, et.al, vol 281 pp 313-319. In-situ measurements of deuterium uptake in palladium using diffractometry. Bennington, et.al, vol 281 pp 323-328. Surface and electrochemical characterization of Pd cathodes after charging. Ulmann, et.al, vol 286 pp 257-263. Morphological difference between hydrogen-loaded and deuterium-loaded palladium. Rolison, et.al, vol 287 pp 375-384. Local heat effects by electrolysis of heavy water. Ratkje, et.al, vol 273 pp 269-273. Energy balance in the electrolysis of water with a palladium cathode. Rock, et.al, in press. A critical analysis of electrochemical nuclear fusion experiments. Kreysa, et.al, vol 266 pp 437-450. Hydrogen and oxygen recombination in calorimetry of water electrolysis. Cunnane, et.al vol 269 pp 163-174. Experimental investigation of thermal and radiation effects by deuterium discharge. Chemla, et.al, vol 277, pp 93-103. Energy balance of electrolysis with a palladium cathode, Part I, Theoretical relations. Balej, et.al, vol 278 pp 85-98. Energy balance of electrolysis with a palladium cathode, Part II, Experimental results. Divisek, et.al, vol 278 pp 99-117. Absorption of electrolytic hydrogen and deuterium by Pd: cyanide adsorption. McBreen, et.al, vol 287 pp 279-291. Calorimetry of the palladium-deuterium-heavy water system. Fleischmann, et.al, vol 287 pp 293-348.
Figure three shows you the KA2281 which offers internal trip levels of -16, -11, -6, -3, and 0 decibels and gives you two stereo channels that can power five lamps each. Note that some commercial units will cheat and assign a pair of light emitting diodes to each output, giving you the illusion of ten steps when you really only have five. Figure four shows you the similar KA2283 which uses levels of -8, -6, -4, -2, and 0 decibels instead. This time, Ive shown you how to attenuate one side to make a single channel ten lamp display. You might want to make your lower level LEDs green, a middle one yellow, and use reds for the upper ones for an additional effect. In this particular circuit, less than 100 millivolts of input lights nothing and more than 900 millivolts lights everything. You can easily change the input attenuators to pick up other operating ranges. Note that the bottom half of the circuit has to attentuate ten decibels more than the upper one to keep all of the steps equal.
Fig. 5 SOME KEY PAPERS ON COLD FUSION that have recently appeared in the Journal of Electroanalytical Chemistry.
37.4
December, 1990
how Model Railroader is one outstanding source of unusual tools and techniques. They also occasionally get into casting compounds and photochemical machining. I think that MR has by far the finest technical writing of any publication anywhere ever on any level, and should be mandatory reading for any tech writer. Some day, I hope to eventually be able to write that good. Maybe with some more experience Another magazine which recently caught my eye was Radio Control Modeler. Again, lots of unusual tools and techniques. Im wondering what happens when you use what these folks call a heat sealing iron for Kroy Color or for circuit board dry resist laminating. The R/C people are also into amateur television stuff. But what I noticed the most (and this is probably ancient history to a R/C buff), is that these hobbists now are building models at one quarter scale. No, not quarter inch scale, but quarter scale. Something like a P-51 Mustang model can run a wingspan of 110 inches or so. And the models are infinitely cheaper than a real airplane. A few hundred dollars (less with lots of hands-on effort) and you are in the air. Quickly and cheaply. Now, any quarter size aircraft can hoist a respectable payload and carry it an acceptable distance. At costs that are ridiculously lower and far less regulated than "real" aviation. Uses? For video obviously. But less obvious would be such things as cave hunting, either by the regular stereo photography for sinkhole searches in difficult terrain, or with infrared film to attempt to find "near" entrances on bitter cold days. So for this months contest, tell me what you would do with an aerial platform that can handle a respectable payload. Lets go well beyond chopping holes in the Goodyear blimp (as was recently done), the nudist colony jokes, and drug deliveries. Limit all your responses to largely legal and purportedly more noble endeavors. Therell be all the usual Incredible Secret Money Machine book prizes for the top dozen entries, along with an all-expense-paid (FOB Thatcher, AZ) tinaja quest for two going to the very best. Be sure to send all your written entries directly to me at Synergetics.
WHOLESALE SURPLUS SOURCES Ace Electronics 3210 Antoine Houston, TX 77092 (713) 688-8114 ACP Components 1317 East Edinger Santa Ana CA 92705 (714) 558-8822 All Components 5999 Summerside #208 Dallas, TX 75252 (800) 367-7312 American Design Box 220 Fairview, NJ 07022 (800) 776-3800 Braniff Industries 91 South Ottawa Dixon, IL 61021 (815) 288-4500 Calcera PO Box 489 Belmont, CA 94002 (800) 257-5549 H&R Enterprises 21521 Blythe Street Canoga Park, CA 91304 (818) 703-8892 Japan Electronics 15138 Valley Blvd Industry, CA 91744 (818) 369-5000 J&C Electronics Box 630 Salem, MA 01970 (800) 447-7014 Jerome Industries 8737 Shirley Avenue Northridge, CA 91324 (818) 993-1393 Leo Electronics 22353 S. Western #201 Torrance, CA 90501 (213) 212-6133 Micro-C 11085 Sorrento Court San Diego, CA 92121 (619) 552-1213 New England Circuits 292 Cabot Street Beverly, MA 01915 (800) 922-NECS Oracle Electronics 1355 Adams Court Menlo Park, CA 94025 (415) 326-2660 Search Electronics 228 East Atara Monrovia, CA 91016 (818) 359-3350 Southwest Memory 3939 Beltline Road #775 Dallas, Texas 75244 (214) 484-1595 Surplus Traders Winters Lane Box 276 Alburg, VT 05440 (514) 739-9328 World Wide 18 Stern Avenue Springfield, NJ 07081 (800) 222-6268
Microprocessors Inc. Derf Electronics 24000 South Peoria 1 Biehn Street New Rochelle, NY 10801 Beggs, OK 74421 (918) 267-4962 (800) 431-2912
The foremost rule in dealing with these folks is no messing around. The preferred contact means is by FAX , and most quotations are only good for twelve minutes or so. Quick cash sales are the norm. As with any surplus, items quickly go in and out of popularity and may vanish entirely. Prices can suddenly skyrocket should the non-surplus lead times lengthen. Quality also can range from absolutely first rate to useless floor sweepings. Buyer beware.
37.5
Hardware Hacker
37.6
Don Lancasters
Hardware Hacker
January, 1991
Low cost visible lasers Electric motor resources Induction motor controls New wavelet math theory Dual digital potentiometer
selectively switch rotor windings in and out so that one or more windings is constantly getting attracted to the fixed stator field. As your armature rotates, new windings get switched in by the commutator, so that a more or less continuous attraction, and thus a rotary motion gets produced. The same universal motor works almost as well with ac, except that sometimes you have strong fields and sometimes weak ones as the current alternates. Usually, the mechanical inertia of your load will more or less average all these variations into continuous rotary power. Even with AC, you still always have opposite poles attracting one another. They just happen to alternate their polarity 120 times a second. But they always change together. Note that you could reverse your universal motor by placing a DPDT switch between the commutator and the stator coils. But this can cause excessive wear if the motor brushes were not designed for two-way use. As your universal motor slows down, its torque will increase. This tends to add to the motor stability. As you slow down, the available "twist" increases to speed you back up, and
et us start off this months column with a very hackable circuit opportunity that has not quite seen the light of day. At least not yet
A motor usually will consist of a stationary part called a stator and a moving part known as a rotor. Another name for the rotor is an armature. If it gets necessary to physically transfer current to your rotor, either brushes and a commutator, or else slip rings can be used. Slip rings apply continuous power to your rotor, while fixed brushes and a commutator selectively switch in and out chosen windings that happen to be aligned with the brush axis at any particular point in time. Youll find quite a few different possible motor designs. But by far the two most economically important are that series DC motor using brushes, which is sometimes called a universal motor, and the AC induction motor. Lets look at each one in turn. Figure one shows you a universal motor and its torque versus speed curve. Youll recognize this one in your electric drill, any hot tub blower, vacuum cleaner, blender, sewing machine, or older car starter. There is a pair of wound stator coils that (when run on dc) will produce a constant stationary magnetic field. In series with the stator coil is a pair of brushes with a commutator, which
STATOR
ROTOR
TORQUE
STATOR SPEED
Fig. 1 A UNIVERSAL SERIES motor with brushes as is found on an electric drill, along with its speed-torque characteristics. The speed is easily and electronically controlled by using simple circuits.
38.1
Hardware Hacker
vice versa. Self regulating. Since you have lots of torque at zero speed, your motor self-starts. It might overheat badly if you stall it permanently. But, given half a chance, a universal motor easily gets itself up to speed. And stays there. Any motor is simultaneously a generator, and vice versa. A universal motor that is lightly loaded and running at a high speed produces an output voltage and resultant current in series with your input current. This voltage generated is known as a back emf, which opposes the input current. When lightly loaded, very little current is drawn from the supply, since the back emf and its resultant current nearly cancels out the input current. Which is real handy from a conservation of energy standpoint. If you are doing no useful work, and if your motor isnt getting particularly hot, then you should not need too much input energy just to spin things. As you add mechanical load, your series machine will become less of a generator and more of a motor. The back emf and its cancelling current drops, and the input current goes up. Your new energy needed is mostly transferred to your mechanical load. Again by conservation of energy, you are putting more energy in so you can get more energy out. Neglecting nonlinearities such as air resistance, the speed of a universal motor for a given load is determined only by the input current. You raise your current to go faster; reduce it to slow down. Which says you can just throw any old high power resistor in series with your universal motor to control its speed, just like used to get done on sewing machine motors. Better yet, you can use a far more efficient triac style light dimmer for control. Or, best of all, you can sense that back emf or otherwise measure your actual motor speed and use electronic feedback with a triac or SCR circuit which is only slightly more complex than a plain old dimmer. This gives you a tightly regulated speed control plus the ability to run real slow with lots of torque. The bottom line is that it is trivial these days to electronically regulate the speed of a universal motor. Weve seen several circuits in past issues and in those Hardware Hacker II reprints. They are also ten bucks at your hardware store or $3.95 surplus. As we have seen, LSI Systems is a good source for fancy universal motor controller chips, while lots of detailed ap notes appear in the usual triac data books from Motorola, SGS, Texas Instruments, and several others. So what is wrong with universal motors? If they are so universal, why arent they used everywhere? The biggest problem lies in those brushes. Brushes wear out. They are inefficient. They are both acoustically and electrically noisy. They spark and can start fires or explosions. And these problems really get out of hand when you need more horsepower. Look around, and youll see that virtually all universal motors in your home are only run on an intermittent basis. Mostly because your brushes cannot be run continuously without grinding themselves into oblivion. Or driving you up the wall with noise. Ideally, we would like to conjure up some scheme to get power onto the rotor without any brushes or other physical contact. And thats where the induction motor comes in. Figure two shows us a brushless beastie known as an AC synchronous motor. It has a wound stator and a permanent magnet rotor. Assume that we are plugged into the 60 Hertz ac line and that the rotor just happens to already be spinning at 3600 RPM. On a positive line peak, the north magnet pole will get attracted to the upper stator coil. By the time your magnet is pointing straight up, the current will be going through zero and there will be no mutual attraction or repulsion. Soon afterward, the line current swings negative, and that opposite polarity starts repelling the magnet, continuing it on its merry way. The exact opposite happens to the other pole, and the motor will continue to spin. The synchronous speed is set only by the number of poles and the line frequency, and is totally independent of your input current. The two pole synchronous motor runs at 3600 RPM off the 60 Hertz line; a four pole job spins at 1800 RPM, and so on. We thus see that a synchronous motor is real handy for maintaining a constant speed. Important use areas are for electric clocks, phonograph motors, timers, chart drives, and other lower power uses where an absolutely constant speed independent of the load or input current is essential. What happens whjen we apply too much of a load? If your magnet gets 90 mechanical degrees out of phase, there will be no attraction or repulsion, and zero power routed to the load. If it gets further out of phase, it actually tries to stall itself. Thus, the slower the speed, the less the torque. You really have a two speed device here 3600 RPM and 0 RPM. One con-
STATOR
ROTOR
S
TORQUE
STATOR SPEED
Fig. 2 A SYNCHRONOUS AC motor as is found on an clock, along with its speed-torque characteristics. The speed is determined only by the line frequency. The input current only determines how strong a load can be driven.
38.2
January, 1991
sequence here is that any true single phase synchronous motor will not start by itself. You will have to help it along with some switchable starting winding, a pole shading, an ferrous hysteresis cup, or a second or a third winding driven from a two phase or three phase source. A second consequence is that the speed of a synchronous motor with varying load or current could end up unstable. If it ever breaks out of sync, you almost surely will stall. This happens because a slowdown produces less torque, which in turn will produce less speed. Just the same as an auto in too high a gear for the grade will buck or stall. The load current does decide how large of a load could get driven at the synchronous speed. The more input current, the more power you could deliver to the load without breaking your synchronization. Figure three shows us a variation on a synchronous machine known as the AC induction motor. This one is by far the most common motor in use today. You will find these on your air conditioners, heater blowers, washers, most table saws, dishwashers, dryers, drill presses, water pumps, and many circulation fans. These are also the ones you would most like to be able to cheaply control the speed of. Especially to raise the efficiency of heating and cooling. The concept for the ac induction motor was positively brilliant. Tesla strikes again. Somehow you have to get currents onto the rotor, but we definitely want to avoid any brushes or other mechanical contacts. Question: What do you have when you have coils and iron driven from an AC current? Answer (A): a motor. Or (B): a transformer (!). Instead of brushes or slip rings, you transformer couple the rotor current in an induction motor. This is usually done by creating a special low impedance transformer secondary known as a squirrel cage. The purpose of the squirrel cage is to serve both as the transformer secondary and as rotor coils to simulate the rotor magnet of figure three. But not so fast. Literally. If you are running at synchronous speed, no lines of flux will be cut, and there will be no current induced into the squirrel cage. Which gives you synchronous speed, but zero power. Now, let your speed slip just a little bit. There is now a very low frequency ac current induced in the squirrel cage. This does create a changing magnetic field, and you now have more torque than you do at synchronous speed. Lower the speed slightly more, and you create even more torque. But slow down too far, and your torque starts dropping radically. Thus, an AC induction motor has optimum torque at a speed which is only modestly less than synchronous. Which is why your "quarter horse" motor is usually rated at 3450 RPM for a two pole motor and 1734 RPM for a four pole one. As you can see, we have a well behaved torque-speed curve between the synchronous speed and the optimum but only slightly slower peak torque speed. Below this peak speed, youll have the same instability and dropout problems that youd have with a pure synchronous machine. Like a synchronous motor, a single phase induction motor will not start itself. Usually there will be a second starting winding that gets kicked in only when getting the motor up to speed. Other starting variations can include a pole shading or multiple phase windings. In theory, you could reverse your induction motor by reversing the direction the starting method shoves the rotor. Details vary with the motor type. In some cases, the only way to reverse the motor is to remove the rotor and put it back in pointing the other way. Another problem with induction motors is that the optimum motor speed is usually far too fast for useful real-world work. You almost always have to use pulleys and belts or some other speed reduction scheme. Or go to lots and lots of poles, as is done in those fancy ceiling fans. Thus, any attempt to control the speed of an induction motor with a series resistance or a simple phase control is doomed to failure. This flat out will not work and probably will burn up your motor as well. Its like plowing with a pig. Now, you can use current control to alter the speed of your motor between its synchronous speed and its slightly slower optimum speed. But this range is usually far too narrow to be useful. Unless you are willing to vary the frequency applied to your motor. But since the impedance of a motor varies
STATOR
ROTOR
TORQUE
STATOR SPEED
Fig. 3 AN AC INDUCTION MOTOR as is found on a washing machine, along with its speed-torque characteristics. The current only controls the speed over a very narrow range between the optimum slip speed and the synchronous speed. To control the speed of this type of motor over a wider range, you have to use a complex cycloconverter circuit that can change both the input frequency and current, as well as accurately measuring the speed. You also have to be sure your starting windings (not shown here) do not cut back in at lower speeds.
38.3
Hardware Hacker
with frequency, youll also want to vary the voltage or the input current as you change frequency. Keeping the magnetic flux at a constant strength seems like a good starting point, so raising the voltage with the line frequency seems like a good idea. Youll also have to quite accurately sense your motor speed at all times, and feed back all of this information into your controller. In addition, you do have to guarantee that the starting windings or whatever do not kick in at slower speeds. Most starting windings are for very intermittent duty only and will rapidly burn up if run continuously. Circuits which sense motor speed and vary both the frequency and the current are often known as cycloconverters. Yes, you can get them. One major supplier is Asea Brown Bovari, a Swiss firm having a wide range of induction motor speed controls. But they are not cheap. And cycloconverter controlled induction motors tend to whine a lot, owing to nonlinear magnetic harmonics in the mid audio range. Tweeeeee. But a $19.95 wide range and quiet quarter horse induction motor speed control is not likely to show up in your hardware store in the next few months. Even though we do have bunches of better magnetics and new intelligent power integrated circuits available that should make the task far easier than it once was. What you are more likely to see are modifications to traditional induction motors which can make them more amenable to electronic speed control. Such things as dual stators, variable frequency rotors, and internal speed sensors. The bottom line? A lot of thought and time and effort has gone into induction motor speed controls for several decades now. So far, nothing really useful has yet seen the light of day. Despite an incredibly big bag of nickels waiting for the winner. I strongly urge you to try and hack this one. But dont expect any prompt or easy results. Needless to say, we here at Radio-Electronics editorial will pay very well for the first hackable, wide range, and sanely priced quarter horse induction motor speed control construction project.
+5V
14 13 12 11 10 9 8
VCC SQ W0 H0 L0 CQ DQ
DS1267
VBB H1 L1 W1 RST CLK GND
1 2 3 4 5 6 7
-5V
Fig. 4 A DUAL DIGITAL POTENTIOMETER using the Dallas Semiconductor DS1267. A 17-bit serial word determines which of the 256 steps for each pot gets selected. Unlike EEPOT chips, the settings are easily read but forgotten on power down. A CQ serial output lets you daisy chain your digital commands.
38.4
January, 1991
step potentiometer. This dude was a single channel device, had a permanent memory that remembered even when it was unpowered, and let you increment or decrement the pot setting with a simple interface. But it had no way to sense the present setting, nor any way to jump to any setting without moving through all of the intermediate ones. Sequentially. Dallas Semiconductor has just introduced a new DS1267 dual digital potentiometer chip that has strengths where the X9103 was weak, and vice versa. As figure four shows us, this one gives you two 256 step pots in a single package in your choice of 10K, 50K, or 100K total resistance. These are not memory devices. They return you to a 50% "mid-wiper" position on power up and must get rewritten each time. As is becoming common on many new chips these days, there is a three-wire serial control provided, intended to interface with three computer port lines. To set up your pots, you clock in seventeen data bits by using the CLK and DQ lines, while keeping the RST line high. Eight of these bits are for the first pot, eight for the second, with the final bit being used to optionally cas- cade the pot pair into a single pot with 512 steps of resolution. Drop the RST line to enter the new settings for the potentiometer pair. Theres also a serial OUT that lets you read out the present settings or else cascade chips for such things as multi-band equalizers. As long as RST remains high, your old setting gets saved. Thus, you can do 17 clocks to read out your existing settings, and 17 more to enter the new ones, without any noise or glitches. The digital end of the chip works off your usual 0 and +5 volts. The analog end can be anywhere from 0 to +5 volts with a grounded VB substrate pin, or can be used over a -5 to +5 range with -5 on your substrate pin. Thus, you can easily handle bipolar analog signals that go above and below ground. Your max clock frequency is 10 Megahertz, meaning that you can upgrade your pot settings as often as 580 Khz or so. Their series wiper resistance is typically 400 ohms, and you are allowed a max potentiometer or wiper current of one milliampere. Be sure to see the data sheet for additional specs and timing details. For our third contest this month, just tell me what you would do with one or more cascaded dual digitally controlled potentiometers. with electronic signals. Besides the intuitive or "real world" time domain, Fourier techniques let you create a separate frequency domain. Things not at all obvious in the time domain become quite clear in the frequency domain, and vice versa. For instance, a square wave in the time domain is just a signal that keeps bouncing up and down between two levels. In the frequency domain, that same square wave can be shown to consist of an infinite string of sinewaves. Specifically, you can build up a square wave from its fundamental sinewave, one third its third harmonic, one fifth the fifth harmonic, and so on up the line. A few of the zillions of places that the frequency domain can become important include spectrum analysis, holograms, video image compression, music synthesis, side looking radar, picture deblurring, biophysics, earthquake detection, neural computing, geophysical exploration, human aural and visual processing, HDTV channel squashing, pattern recognition, hard disk storage, nonlinear harmonics, speech generation, robotic sensing,
Making Wavelets
There is a brand new math revolution taking off that appears certain to profoundly change much of what electronics is and what it will be able to do. All the noise is over wavelet theory, an incredibly powerful new technique that promises to blow the 200 year old Fourier analysis, synthesis, and transforms completely out of the water. Why worry about some obtuse new math theory? Well, first because it is there. And second, because it is a sure fire winning topic for a school paper. And thirdly, theres a lot of electronic doors about to suddenly get slammed in your face if you do not quickly pick up on exactly what wavelet theory is and what it can do for you. Fourier analysis is (er better make that was) a method of giving you a second way of looking at and dealing
A New wave in applied mathematics B. Cipra, Science 24 Aug 1990, pp 858-859. New wave number crunching C. Brown E.E. Times, 5 Nov 1990, pp 31-34 Video compression using 3D wavelet transforms. A. Lewis, Electronic Letters, vol 26 no 6 pp 396-8. Non-orthogonal wavelet representations in relaxation networks. J. Daugman, New Developments in Neural Computing, pp 233-50. A theory for multiresolution signal decomposition. S. Mallat, IEEE Transactions on Machine Intelligence, v11-7 pp 674-93. Wavelet transformation in signal detection. F. Tuteur, 8th IFAC/IFORS Symposium, vol 2 pp 1061-5. Entropy reduction and decorrelation in visual coding. J. Daugman, IEEE Trans. Biomedical Engineering, vol 36 no 1 pp 107-14. Wavelet transformation in signal detection. F. Tuteur, ICASSP Speech Conference 88, vol 3 pp 1435-8. Complete discrete 2-D Gabor transforms. J. Daugman, IEEE Trans Acoustics & Speech, vol 36 no 7 pp 1169-79. Dispersive noise removal in t-x space. Beresford-Smith, Geophysics USA, vol 53 no 3 pp 346-58. Adaptive deconvolution by lattice filters. S. Persoglia, Bulletin of Geophysics Theory, vol 27 no 107 pp 169-83. A critique of seismic deconvolution methods. A. Jurkevics, Geophysics, vol 49 no 12 pp 2109-16. Statistical pulse compression. E. Robinson, IEEE Proceedings, vol 72 no 10 pp 1276-89.
Fig. 5 SOME KEY PAPERS on the new wavelet math theory which is now revolutionizing just about everything in electronics.
38.5
Hardware Hacker
in two years or so. New data books for this month do include that Special Purpose Linear Devices entry from National. Be sure to check out their LMC835 digitally controlled graphics equalizer youll find on page 1-227. Three other new data books are the Logic Databook from the Integrated Device Technology folks, that new Optoelectronics and Image Sensors from Texas Instruments, and that new Integrated Circuits Data Book #33B Supplement from Burr-Brown. Static memory RAM specs cleverly disguised as baseball trading cards are being offered by SGS in an unusual promotion. These are the same folks that previously gave you soup cans full of assorted free integrated circuits and soap boxes full of EPROMs. Three interesting new surplus catalogs showed up in todays mail. They do include Circuit Specialists, International Micro Electronics, and H&R Enterprises publications. Free samples this month include 1K Serial EEPROM s from International CMOS Technology, and SCOPE Octal drivers from Texas Instruments. TI also has a new three volume set on Digital Signal Processing Using the TMS320 available. Turning to my own products, for the fundamentals of digital integrated circuits, be sure to check into my classic TTL and CMOS Cookbooks. And, as you can tell from my nearby Synergetics ad, weve now got Hardware Hacker III and Ask the Guru III reprints available, as well as some new PostScript books by other authors. Also, a reminder that I do have this great new PostScript PSRT roundtable and library up on GEnie. Youll also find lots of Hardware Hacker and all of the Midnight Engineering preprints and other resources here. Finally, I do have a new and free mailer for you which includes dozens of insider hardware hacking secret resources. Write or call for info. Our usual reminder here that most of the items mentioned appear either in the Names and Numbers appendix or in the Motor Resources sidebar. As always, this is your column and you can get technical help and offthe-wall networking per that Need Help? box. The best calling times are weekdays 8-5 in Mountain Standard
ELECTRIC MOTOR RESOURCES Appliance 1110 Jorie Boulevard Oak Brook, IL 60522 (708) 990-3484 Appliance Manufacturer 29100 Aurora #200 Solon, Ohio 44139 (216) 349-3060 Asea Brown Bovari 1460 Livingston Avenue N. Brunswick, NJ 08902 (201) 932-6000 Bodine Electric 2500 West Bradley Place Chicago, IL 60618 (800) 7BODINE C&H Sales PO Box 5356 Pasadena, CA 91117 (800) 325-9465 Design News 44 Cook Street Denver, CO 80206 (303) 388-4511 Fair Radio Sales PO Box 1105 Lima, OH 45802 (419) 227-6573 Fasco Industries 500 Chesterfield Ctr #200 St Louis, MO 63017 (314) 532-3505 Grainger 2738 Fulton Street Chicago, IL 60612 (312) 638-0536 Herbach & Rademan 401 East Erie Avenue Philadelphia, PA 19134 (215) 426-1700 IEEE/Energy Conversion 445 Hoes Lane Piscataway, NJ 08855 (201) 981-0060 IEEE/Industrial Apps 445 Hoes Lane Piscataway, NJ 08855 (201) 981-0060 JerryCo 601 Linden Place Evanston, IL 60202 (708) 475-8440 Lindsay Publications PO Box 538 Manteno, IL 60950 (815) 468-3668 Machine Design 1100 Superior Avenue Cleveland, OH 44114 (216) 696-7000 Molon Motor & Coil 3737 Industrial Avenue Rolling Meadows, IL 60008 (800) 52-MOTOR Motion Box 6430 Orange, CA 92613 (714) 974-0200 Northern PO Box 1499 Burnsville, MN 55337 (800) 533-5545 PCIM 2472 Eastman Bldg 33-34 Ventura, CA 93003 (805) 650-7070 SAE 400 Commonwealth Drive Warrendale, PA 15096 (412) 776-4970
electrocardiograms, outline detection, seismic profiling, correllation, and data compression. In its spare time, Fourier analysis and synthesis leaps tall buildings in a single bound, milks cows, and even puts the cat out. The latest algorithms even sometimes may include a soft ice cream dispenser. But the big problem with Fourier analysis is that everything was connected to everything else. Make even the slightest change, and you had to go back to square one and recompute everything. And while Fourier series is a great way to handle "most" of a square wave, it sure has troubles with the suddenly changing leading and trailing edges. But many real world problems are like that. A mix of some stuff that is suddenly and dramatically changing, combined with other areas that are more or less stationary or otherwise uninteresting. What wavelet theory does is let you selectively mix and match wavelets
that can both deal locally with sudden changes and globally with averages and backgrounds. You can thus selectively apply all your math power precisely where it will do you the most good. And do so with speeds and efficiencies that were totally unheard of with traditional Fourier analysis. Figure five shows you a few of the newer key papers involving wavelet theory. Start out with the Science and E.E. Times overview summary stories before you get into the heavy stuff. Ive recently added a WAVELET.PAK program to my GEnie PSRT. Let me know if you want to get into this any deeper.
38.6
January, 1991
Time. Lets hear from you. 3
38.7
January, 1990
25.7
Don Lancasters
Hardware Hacker
April, 1991
Machine shop resources Infrared people detection Elegant simplicity contest Liquid crystal battery tester Exploiting paradigm shifts
total waste of time. Similarly, if useful computer touch screens would ever have done anything, these should have taken off years ago. Your bottom line is that people do not like to touch computer screens, and trying to convince them otherwise is probably fruitless. The same goes for Dvorak or other "improved" keyboards. The whole point of this Hardware Hacker series is to try and identify those paradigm shifts that you can work with and profit from. We have seen many recent examples. These include anything and everything involving PostScript, magnetic refrigeration, wavelet theory, desktop finishing, visible laser diodes, direct toner printed circuits, and book-ondemand publishing. To name just a few random samples. For some others, see my Eminently Hackable Emerging Technologies story found in that July-August 1990 Midnight Engineering. And downloadable as my GEnie PSRT file #116 EMERGOP.TXT.
his months topics include finding new hackable project ideas, working with infrared people detectors, checking out those "in-package" battery testers, another contest, and a fresh look at machine shop resources. Lets start off with some of the key secrets behind
hopelessly crippling all of traditional publishing. CD ROM . That insanely great PostScript language. Your key hacker opportunity: The people who are doing the shifting of the paradigm usually do not have the slightest idea what they are doing, since they will be always rearwardly focusing on the way things were. When a paradigm shifts, all sorts of new hardware hacking opportunities immediately open up. Especially if you are able to view reality from a different perspective. Or can reduce the cost of something by 50:1, totally changing the market to something totally foreign (and totally misunderstood) by the way things were. If there has not been any recent paradigm shift, the chances are it will be tricky to come up with a useful product. For instance, there has not been one iota of improvement in any of those Peltier thermoelectric cooling modules in the past two decades, owing to the inherently low efficiency of these devices. Any hacking work done the "old way" here is probably a
0.0003
0.0002
0.0001
0 0 10 20 30 WAVELENGTH IN MICROMETERS
Fig. 1 THE FAR INFRARED SIGNATURE of a 98 degree F person in a 70 degree room environment. Note that these are extremely weak power levels.
39.1
Hardware Hacker
A human will normally radiate at 98 degrees Farenheit, compared to the other objects in the room which will typically radiate at a 70 degree range. A human body is a very weak radiator which becomes even more so when compared against the ambient. Figure one is a typical curve of human body radiation in a normal room. As you can see, the radiated energy is centered on the eight micrometer range in the far infrared. One sensor which is capable of detecting the radiation from a person is known as a pyroelectric infrared detector. These are available at very low cost from the Amperex division of Phillips, among others. As well see shortly, Amperex has lots of good ap notes and data sheets available. The pyroelectric infrared detector consists of one or two detectors that in turn input to a field effect transistor source follower. The basic detector is a capacitor, across which several hundred microvolts of dc voltage will be generated in the presence of a warm and nonmoving human body. The detectors are often used in side-by-side differential pairs that are imaged slightly differently. They will tend to cancel out stationary sources. One very big gotcha here: These sensors are basically a capacitor so they cannot indefinitely produce a dc output voltage. Even the tiny bias current of a field effect transistors gate is enough to flatten any long term dc level. Thus, the pyroelectric infrared detector only is able to respond to changing levels of infrared energy. Stationary sources are ignored. So, the trick is to make the infrared signals appear to be rapidly changing. One obvious way is to have the person run through the beam. This can produce a usable transient. But something better is clearly needed. The traditional method was to chop the beam by putting a bladed fan in front of it. This would input an infrared square wave that represents the difference between the body and fan blade temperatures. By knowing how fast those blades whipped past, you could also do a synchronous demodulation that could also increase your detection sensitivity. But moving fan blades and synchronous detectors are expensive.
Fig. 2 SINCE PRYOELECTRIC INFRARED SENSORS are capacitors, they cannot hold a dc or stationary level. To emphasize changes in motion, special Fresnel lenses are often used that have "hot" and "cold" areas as shown here.
The opposite side of the paradigm shift is the sucker bet. Any thoroughly plowed ground that, for one reason or another, just did not and will not ever hack it. Ratholes into which countless corporate dollars have been foolishly dumped with zero visible results. Or results having the exact opposite of the intended effect. Obvious sucker bets include the UNIX language, the NeXT computer, TrueType, or most anything involving Teletext.
39K
10K
14
13
12
11
10
KRX11 ir sensor
LM324
OUT1 IN1- IN1+ VCC IN2+ IN2- OUT2
1 2 3 4 5 6 7
.01 100K
390K
100K
330K
4.7
Fig. 3 A "PEOPLE DETECTOR" intended for use as an office lighting control. The output goes high on any motion detection that would increase or decrease the input far infrared signature.
39.2
April, 1991
Somehow you have to gater your infrared energy from your area and concentrate it on the detector surface. Although mirrors are one solution, a plastic Fresnel lens is both a better and more economic one. Fresnel is pronounced "Fra-nell" Now for the tricky part. Instead of making the lens operate uniformly over the surveyed area, it is purposely striped so that there are "strong" and "weak" sensing areas. A typical lens pattern is shown in figure two. As the person walks through the beam or otherwise moves, they travel between the strong and weak lens areas, creating more of a varying signal than they would otherwise. Figure three shows you a schematic of a simple people detector using a dual element pyroelectric detector and a quad op amp. The Fresnel lens has strong and weak areas that alter the strength of the infrared signature of a moving person. This is sensed by the detector and routed to a x600 ac amplifier. The combined frequency response of the detector and the amplifier is an 0.3 to 5 Hertz range. Which is usually an optimum for most people movements. The output of your amplifier is routed to a window detector or dual comparator. This detector will output a signal on any sudden change in the infrared signature. Usually, the output of the window detector is routed to a counter of some sort to minimize false alarms. In security applications, an alarm output is created. For the occupancy detectors, the lights are quickly turned on, and then left on for a selected number of minutes. A fifteen minute delay is often an optimum for people who are usually sitting at a desk or bench. Each time they move, the on time gets extended. One commercial source of readyto-go occupancy sensors is Leviton. These fit in an ordinary power outlet. Occupancy sensors can dramatically reduce the power bills in most larger commercial buildings. Another alternative to pyroelectric detectors is the Kynar Piezeo Film from the Atochem folks. While less sensitive to infrared and much more tuned in to motion or vibration, this approach can let you integrate your lens and sensor into one single thin assembly. Penwalt also does have ap notes and data sheets available. While the electronics involved in people detection are both simple and straightforward, your mirror or lens design is not. Thus, you are better off using some already developed and debugged commercial lens/detector combination than trying to work one up from scratch. Figure four lists some of the more readable Amperex ap-notes and data sheets on people detecting. This is a very good starting point for picking up all the infrared sensing basics. UPDATE: Shortly after this story was run, Amperex discontinued all infrared detector work, and their fine ap-notes are no longer available. The reason was that the new Kynar piezo sensors by Pennwalt/Atochem were far simpler and cheaper to use. Because of the new California law mandating these sensors in all new office building constructions, infrared people detectors have become a low cost commodity, readily available at any electrical wholesale supply. Both Herbach and Rademan and Heath offer ready to go units for far lest than the cost of building them up yourself. Especially when you factor in the tricky Fresnel lens design. Typical surplus pricing is in the $12 range. A great buy.
Pyroelectric infrared detectors for movement sensing. Technical publication #134, 1984 Remote level sensing using pyroelectric infrared detectors. Technical publication #135, 1980 Low cost remote sensing radiometer using the RPY89 infrared detector. Technical publication #138, 1980 Low cost automatic light switching using passive infrared sensors. Technical publication #147 February 1985 Ceramic pyroelectric infrared sensors and their applications. Technical publication #163 Passive infrared (PIR) intruder alarms. Technical publication #213 April 1986 Movement sensing using a multi-element fresnel lens. Amperex ap note, November 1988. KRX10 dual element pyroelectric infrared sensor. Data sheet, September 1988. KRX11 dual element pyroelectric infrared sensor. Data sheet, September 1988 RPW100 dual element pyroelectric infrared sensor. Data sheet, September 1988 Fresnel lens data sheet and explanatory notes. Phillips data sheet, April 1986
39.3
Hardware Hacker
portion of the wedge at the bottom will offer you a higher resistance per unit length and thus will get hotter than the upper and wider part. Thus, this particular resistor will set up a temperature gradient that ends up hottest at the bottom and coolest at the very top. How hot? Well, that all depends on how much current your cell puts out under load. Power equals the current squared times the total resistance. A thin liquid crystal coating gets placed on the reverse side of the resistor. This is the same stuff used in clinical and desk thermometers. At a certain transition temperature, your liquid crystal will turn a bright green. Below that temperature it will remain black, and above it a dark grey. The more the available current under load, the higher the green spot on the display. Thus, a "good" cell will have a "high" green spot, and a "bad" cell might have a "low" green spot, or perhaps none at all. The results surely change with the ambient temperature and the size of the cell being tested, but they assume you are testing your cells at room temperature, and that the cell being tested is the type and size provided in the package. So, basically what you have here is a three cent 0 to 500 DC milliammeter. Done elegantly and simply. Which other uses can you come up with for this new concept? Or, better, yet, lets have us an
we just gotta have a stainless steel potato chip, since the real ones break on the third or fouth time into the dip. Yes, you could use a P-38, but Lets have your thoughts on elegant simplicity, electronic or otherwise. As usual, be certain to send all of your written entries directly to me here at Synergetics, rather than on over to Radio-Electronics editorial.
Green liquid crystal spot will move up and down following the critical temperature set by the quality of the test cell.
Fig. 5 THE LOAD RESISTOR in the Duracell battery tester is wedge shaped so that it gets hotter per unit length at the bottom than at the top. A liquid crystal coating changes color at a critical temperature, moving a green indicating spot up with a strong cell and down with a weak one.
39.4
April, 1991
and plastic for you in tiny quantities; and Lindsay Publications, who have a mind boggling array of lower priced machine shop books in stock. One old line "stocks everything" distributor is McMaster Carr. See if you cant cop one of their humongous 3000 page catalogs. Others include Enco, KBC Tools, J & L Industrial Supply, Rutland, and Wholesale Tool. And, of course, Sears. Two useful surplus sources include Bordens and C & H Sales. My favorite two mechanical trade mags are Design News and Machine Design. The zillions of others include American Machinist, the Used Equipment Directory, and Metlfax. There is also a huge collection of oversize "throwaway" shoppers that include New Equipment Digest (great free product samples!), Industrial Product Bulletin, Industrial Equipment News, and Metalworking Digest. On the hobby side of the fence, the Model Railroader obviously belongs on this list. Two others include Home Shop Machinist and Modeltec. Once again, weve got a rather long list here, but I do have this hollow feeling Ive missed something major. For our second contest this month, just tell me about any hacker useful machine shop resource that I dont already know about. As weve seen time and time again, the industry trade journals can be an incredibly cost effective information gateway which no hacker could do without. See Uhlrichts Periodicals Dictionary at your local library for a complete listing.
MACHINE SHOP RESOURCES American Machinist 826 Broadway, 4th Floor New York, NY 10003 (212) 477-6420 Bordens Surplus Center 1015 West O Street Lincoln, NE 68501 (800) 228-3407 C&H Sales PO Box 5356 Pasadena, CA 91117 (800) 325-9465 J&L Industrial Supply 31800 Industrial Drive Livonia, MI 48150 (800) 521-9520 Model Railroader 21027 Crossroad Circle Waukesha, WI 53187 (414) 272-2060
Modeltec KBC Tools Box 286 6300 18-Mile Road Sterling Heights, MI 48314 Cadillac, MI 49601 (616) 775-6296 (800) 521-1740 Kepro 630 Axminister Drive Fenton, MO 63026 (314) 343-1630 New Equipment Digest 1100 Superior Avenue Cleveland, OH 44114 (216) 696-7000 Roper-Whitney 2833 Huffman Blvd Rockford, IL 61103 (815) 962-3011 Rutland PO Box 587 LaPuente, CA 91747 (800) 289-4787 Small Parts PO Box 381966 Miami, FL 33238 (305) 751-0856 Used Equipment Dir. 601 Pavonia Avenue Jersey City, NJ 07306 (800) 526-6052 Wholesale Tool 12155 Stephens Warren, MI 48090 (800) 521-3420
Lindsay Publications Design News 44 Cook Street, Suite 210 PO Box 583 Manteno, IL 60950 Denver, CO 80206 (815) 468-3668 (303) 388-4511 Enco 5000 West Bloomingdale Chicago, IL 60639 (800) 621-4148 Home Shop Machinist 2779 Aero Park Drive Traverse City, MI 49684 (800) 447-7367 Industrial Equipt News Five Penn Plaza, 8th Floor New York, NY 10001 (212) 629-1500 Industrial Prod Bulletin PO Box 650 Morris Plains, NJ 07950 (201) 292-5100 Machine Design 1100 Superior Avenue Cleveland, OH 44114 (216) 696-7000 McMaster-Carr Box 54960 Los Angeles, CA 90054 (213) 692-5911 Metalworking Digest 310 Gibraltar Drive Morris Plains, NJ 07950 (201) 292-5100 Metlfax 29100 Aurora Road #200 Solon, OH 44139 (216) 248-1125
Service Robot Association. Burr Brown has just introduced an applications and product info BBS at (602) 741-3978. In our free samples department, Exel has borrowed a marketing ploy from the magazine clearing houses a series of stamps. Pick any three on the return card for free samples of their non-volatile memory products. And Annulus has free samples of their 8PDT switches which are great for switching serial ports or any EPROM chips into or out of circuit. For mechanical stuff, the definitive books and videos on glass etching are available through Professional Glass Consultants. Turning to my own products, for the essentials of making your hardware hacking profitable, do check out my Incredible Secret Money Machine. I
have also just released the Hardware Hacker III and Midnight Engineering book-on-demand published reprints. I do have this great new PostScript PSRT roundtable and library up on GEnie. Youll also find lots of Hardware Hacker and all of the Midnight Engineering preprints here. Finally, I do have a new and free mailer for you which includes dozens of insider hardware hacking secrets. Our usual reminder here that most of the items mentioned appear either in the Names and Numbers appendix or in that Machine Shop Resources sidebar shown above. As always, this is your column and you can get technical help and offthe-wall networking per that Need Help? box. The best calling times are weekdays 8-5 in Mountain Standard Time. Lets hear from you. 3
39.5
Hardware Hacker
39.6
Don Lancasters
Hardware Hacker
May, 1991
very now and then, it seems a good idea to go back over some of our older Hardware Hacker subjects and bring them up to date. Certainly one of the most popular topics ever found on our helpline involves
Reloading toner cartridges New electronic phone book More Santa Claus machines Secret laser printer manuals VHF & microwave resources
usually replaced whenever a cartridge is recharged. Note that just washing a wiper pad is a no-no. Our first rule: Toners vary from machine to machine. Typical Canon laser printers use what is known as a black write system, since laser diodes will last much longer this way. On the other hand, all but the newest and most expensive copiers use a white write system so that light ends up as white and dark as black. Thus: Copier and Laser Printing toners must NEVER be interchanged or substituted for each other! There are usually mechanical interlocks that prevent you from plugging a copier cartridge into a laser printer and vice versa. If you attempt to defeat these interlocks, you will end up using the wrong toner. At the least, this gives you useless copies, and at worst, this can cause serious damage. Similarly, toner chemistry changes from printer to printer. Most especially between manufacturers. Our second rule: The refill toner you use must be pretested in and rated for the exact cartridge you are refilling. So where do you get refill toner and all those wiper pads? My two favorite sources are Don Thompson and Lazer
and then electrostatically charged. It next gets selectively discharged by a laser beam, leaving a charge pattern on the drum. As the drum rotates, it passes very near the magnetic roller and the toner selectively jumps onto the drum, sticking by electrostatic forces only where you want an image. As your drum rotates further, it passes very close to a highly charged piece of paper, and the toner particles then jump over onto the paper. Any remaining toner that was left on the drum gets scraped off and routed to a spent toner holding tank. The photosensitive drum then continues on its way for another cycle. Meanwhile, you now have your image on the paper. But it is only held there by gravity and by rather weak electrostatic forces. This will easily smear if you touch it. The paper then goes on to a fusion roller assembly. Heat and pressure will melt the toner and force it into the paper, giving you a fairly durable final hard copy. One very important part of most fuser assemblies is the wiper pad. The wiper pad has a small quantity of silicon oil in it that both lubricates and cleans up any remaining toner on the pressure rollers. Your wiper pads are
HP PRINTER LaserJet I
LaserJet IIP
Personal LW NT
PS410
Fig. 1 HERES A LIST of all the most popular PostScript laser printers, their equivalent service manuals, and the style of toner cartridges they use.
40.1
Hardware Hacker
Non-PostScript laser printers, of course, are an utterly useless ripoff, so be absolutely certain that your used machine can speak genuine Adobe PostScript. Before we begin, note that the photosensitive drums must never be exposed to strong light, or to any light at all over any long period of time. Never get fingerprints on the drum. Cotton gloves are a good idea. Toner is an ultra fine powder that can end up all over everything. It is sometimes best to work outside, and possibly wearing a mask. Toner can, in theory, explode a vacuum cleaner, but this rarely will happen. But do be careful, just in case. The general steps in refilling any cartridge are fivefold:
(1) Remove and discard any waste toner from the spent toner holding tank. Do not reuse this toner; (2) Refill your fresh toner supply tank with a new bottle; (3) Lubricate the drum with a light dusting of Pixie Dust (see below); (4) Replace the oiled wiper felt on the fusion assembly elsewhere in the printer; and finally (5) Update accurate life and service records on a suitable label.
To refill an older CX cartridge with the punch-and-go method, you first snap off the cardboard label and then drill a toner filling hole
Drill a 5/8 inch hole using a #3 Vise Grip Unibit; carefully clear all chips.
A second CX hole is needed to let you empty out your spent toner holding tank. This area is found underneath the cartridge
Drill 3/8 inch hole using a #3 Vise Grip Unibit; carefully clear all chips.
Fig. 2 THE CX CARTRIDGE is easily recognized by its large drum and its obvious "lunchbox" handle. Here are my "punch and go" refilling secrets.
Products. There is yet another outfit known as Black Lightning that stocks specialty toners for T-shirt and fabric printing uses. The toner industry has its own trade journal. It is called Recharger, and is chock full of supplier ads and useful industry info. There are also at least a dozen recharging associations who do have lots of seminars and conventions. Details on all of these usually appear in Recharger magazine. The great news today in toner refilling involves new third party hard coated drums. For some reason or another, the factory stock drums are made needlessly soft. The third party drums instead are ultra hard and can easily be used for dozens of reloads. One leading importer of hard drums is CopyMate Products. Lets look at some specific refilling details. Certainly the most popular cartridges are those used in the Canon CX, SX, and LX engines. Figure one lists many popular laser printers and
the specific engine used in each one. This figure also reveals to you the outstanding Hewlett-Packard repair and service manuals involved. These manuals, and all major parts can be had overnight via VISA /800. For some reason which I simply cannot fathom, Apple Computer absolutely insists that you use the HP service manuals to keep your Apple LaserWriter printers alive. As near as I can tell, this is some sort of a top secret rebate policy. At any rate, you can recognize the older CX cartridges by their larger three inch drums, their red-yellowgreen end dial, and the obvious big lunchbox handle. While the original LaserWriter and all similar printers using these are rather dated, they do remain very useful, especially when printing lots of heavier stock. Many used bargains are now cropping up involving these older machines. Two sources are Don Thompson and The Printer Works.
Figure two shows you those CX refilling details. There is really never any reason to tear down a CX cartridge, except to substitute a hard drum. The original factory drums are big enough that you can often get four or five refills as is. For most people most of the time, a total teardown will cause many more problems than it will solve. Remember that your ultimate goal should be minimizing all of your per-page toner costs, not maximizing the number of recharges for each drum. An extra recharge is pointless if it costs the end user more per page to do so in the long run. If you absolutely have your heart set on taking any CX cartridge apart, youll need two special tools. One is a special tamperproof Torx bit. This is EVCO part number #945B700 and is available from Jensen Tools as well as most refilling supply houses. The second is a special pin-pulling tool called a CX Glompenstractor and is available once again by way of Don Thompson. Should you use my punch and go
40.2
May, 1991
method, you will have to drill two holes in the cartridge on your first reload. Which is best done using a rather unusual step-drill called a #3 Vise Grip Unibit. These are available from Jensen Tools or from any larger electrical contracting supply house. When used with a variable speed hand drill, the Unibit should cleanly cut a perfectly round hole in brittle plastic, while producing a single and easily grabbed chip. After drilling the holes, the spent toner is shaken or vacuumed out. You could reseal this hole with plain old Scotch Tape (be VERY careful to get a secure seal!), or else you use a nickel Caplug. Your fresh toner bin hole gets similarly used to accept a bottle of new toner and then resealed. Figure three shows you the SX cartridge recharging. The LaserWriter NTX is a typical machine that uses this cartridge. The SX cartridges have a one inch drum and end up rather flat looking, being much wider than thick. Should you decide to tear down the cartridge or upgrade to a hard drum, a different glompenstractor is needed having a narrower snout. The details of my punch-and-go refill method remain pretty much the same. First time around, you drill a suitable filling and emptying hole. Once again, the #3 Vise-Grip Unibit in a variable speed hand drill is ideal for this. To refill, drain and discard the spent toner and reseal. Then fill the fresh toner tank and reseal. Figure four shows you the newest LX cartridge recharging. The personal laser printers, such as the new QMS PS-410, do use this cartridge. The LX cartridge is recognized by its small size, an obvious spring and its "white trim" gears and bearings. No holes are required. To access the toner bin, pull the two pins by using ChannelLock #357 end pliers. The tank can then be refilled through the existing Caplug. To change the drum or drain the spent toner, remove the four Phillips screws on those nylon drum bearings and then pull the drum. Spent toner can be simply vacuumed or shaken out. After a recharge, it is a good idea to very lightly dust any drum with a suitable lubricant. Many of these are based on plain old zinc sterate. Only dont substitute baby powder since the
The SX cartridge punch-and-go refilling process is similar to the CX, except for the hole locations. The filler hole is shown here
Drill 5/8 inch hole using a #3 Vise Grip Unibit; carefully clear all chips.
One or more spent toner drain holes must also get added to the SX cartridge. The plastic is thin, so use a conical step drill
Drill 3/8 inch hole using a #3 Vise Grip Unibit; carefully clear all chips. Be certain that the new hole is centered between the die sink marks!
Fig. 3 THE SX CARTRIDGE is wider than it is high, has a small drum, and is the most popular cartridge for the larger 8 PPM PostScript laser printers.
perfume and oils will do you in. The usual name here is Pixie Dust. Pixie dust is available through most of the recharging supply houses at very low cost. You can make a "duster" from the toe of a childs athletic sock and a rubber band. Once again, the wiper pad on the fusion assembly should get replaced every time you change the cartridge. You normally keep your old wiper wand and drop a new peel-and-stick oiled nomex felt strip in place. While you can obtain toner tank resealing strips, travel of any kind is extremely rough on toner cartridges. I do not recommend ever moving a cartridge further than you can gently and personally hand carry it. Nor do I recommend ever swapping your own cartridges for unknown outsiders. I strictly limit my personal recycling service to a six mile radius. Yours also should be. Pizza style service. Reuse of toner removed from spent toner holding tanks is not in the least
recommended, nor is recycling your own wiper pads. There does remain plenty of "zoo" aspects to the toner recycling. Certain irresponsible manufacturers have now begun some high profile national "recycling" programs which in fact destroys the cartridges rather than recycling them. The hope here is to permanently get the cartridges out of circulation before they could be refilled and reused. Only an absolute idiot would participate in any program of this sort. If you do nothing else, you can sell your empty cartridges locally for $5 to $10 each, and then contribute as much of the proceeds as you care to to your favorite environmental group. Any salesman that tells you that normal use of any properly recycled cartridge can automatically void your printer warranty is telling you an outright lie for which they could be criminally prosecuted. Some others are literally gold plating stuff that does
40.3
Hardware Hacker
not in the least need gold plated. Yet others will substitute shoe polish for proper hard drum recoating. Those recharge/repair schools run the gamut from outstanding and high quality bargains on down to outright ripoffs. To tell one from another, ask the school for a list of all the previous students in your area. Then call one or two of them. You will have to pay very careful attention to details. But the new toner recharging industry is fast maturing and now offers all sorts of exciting and cost-saving new hardware hacking opportunities. But the number two place I always go to is the Electronic Industry Telephone Book from Harris Publishing. Although it lists for around $50 per year, sometimes you can get a free or promo copy from a sales rep. This national coverage gem works just like any other phone book, with alphabetical white pages and by-topic yellow classifieds. Their listings are very thorough, and I am continually amazed how often this one volume can solve so many problems. the above", lets look at a pair of new alternatives. Have you ever played around with your glue gun? While not too readily available, you can get polyethelene rods to use as glue sticks. This gives you a method for encapsulating components in high quality plastic or making your own custom connectors. Or of doing plastic casting at a tiny fraction of the usual mold costs. Lets carry this one step further per the crude system of figure five. Say you wanted to produce some large display letters in various styles and sizes. Just take a modified glue gun on a linear stepper and a no-stick base on a second linear stepper, and you should be able to put a plastic bead down that follows the shape of the letter. Repeat the process until the entire letter is created. The host computer traces out the proper path to build up the letter one bead at a time. Admittedly, this is a rather crude system which is limited to thin two dimensional objects. And we havent properly addressed the third dimension at all. But it is a good starting point that could lead to some exciting new developments. One suitable stepper would be the Hurst model SLS. Ive been meaning to work up some more details on this and on Hursts new EPC-015 controller. Maybe in a future column. There is a commercial variation of the "use a glue gun" desktop prototyping method. This is the brand new Stratasys 3-D Modeler. They refer to their process as a Fused Deposition Modeling, or FDM. The system starts with a large roll of .020 or .050 diameter plastic or wax filament. The filament is heated just hot enough to make its outside tacky. The filament is then laid down into an existing pattern in the same way you can do artsy-craftsy stuff with string soaked in glue. A three dimensional object is then built up, literally one string at a time. While they have an elaborate CAD software system based upon NURBS splines, the PostScript language and any old word processor should be able to do a vastly better job far faster and much cheaper. The FDM method seems especially well suited for modeling containers and other hollow packaging products.
The LX cartridge does not need any modifications before any refill, hard drum upgrade, or any spent toner draining
THIRD, Remove the screws to change the hard drum or drain out the spent toner.
SECOND, pull these pins to remove the fresh toner tank for easy refilling.
But be EXTREMELY careful not to touch the photosensitive drum or expose it to any strong light or ANY light of long duration.
Fig. 4 THE SMALLER LX CARTRIDGE is used in the "personal" 4 PPM printers and is easily spotted by its obvious spring and the "white trim". NEVER use copier toner in a laser printer!
40.4
May, 1991
But sharp edges appear tricky to do, especially gear teeth. Ultimate costs should be low, since no lasers, fumes, high temperatures, costly materials, or exotic chemicals are involved. Your typical speeds approach 1000 inches per minute. One big problem with the system: Some prototypes can end up looking like something that missed hitting the reject bin in the Special Education crafts class. Finer filaments can cure this, but build more slowly. Im wondering if a better prototyping solution might not involve two steps. Homes are usually built in a "rough" and the "finish" stage. And machinists often work with near net stock to try and minimize their total production time. And modelers will often build their model first and then superdetail it later. Possibly some system that gets the shape pretty near the way you want it quickly and crudely, and then the second step which measures and then modifies what you have to give you your final precision fit and finish. Lets have your thoughts on this. wave circuit could severely disrupt what you are trying to measure. Not to mention that high frequencies and weak signal levels are involved. Most suitable test equipment also tends to be very specialized, arcane, and quite expensive. And hard to justify. Third, VHF and microwave parts can be ridiculously costly, since most of them are aimed at gold-plated low volume military uses. Fourth, the math, the field theory, and all of the advanced technical skills needed to do anything useful at these frequencies goes way beyond electronic fundamentals. A lot of useful microwave information tends to require advanced degree skills, can get very hard to find, and often will even be classified. And finally, much of microwave design work ends up far more art than science. To this day, black magic can be involved, and you either have the right feel for what you are doing, or your circuits simply will not work. For this months resource sidebar, I have tried to gather together some
VHF and microwave stuff you might
find of hacker interest. Obviously, youll want to start with the old Radio Amateurs Handbook published by the ARRL. Many of the ham magazines and club activities will also involve the VHF and microwave frequencies. Per a recent rule change, certain new ham licenses no longer require any Morse code. Which was long overdue By far the leading surplus house carrying microwave radar goodies is Radio Research Instruments. Only these folks tend to concentrate upon whole systems and refuse to show prices in their flyers. Alternate places to check do include Fair Radio Sales, Surplus Traders, and in the Nuts and Volts shopper. While utterly ancient, McGraw Hills MIT Rad-Lab series should still be scrounging around on the dusty back shelves of the larger technical libraries. Volume One in particular, Ridenours Radar System Engineer-
PO
LY
ET
E EL
NE
RO
MO
DI
FI
ED
GL
UE
GU
N XI AX S LIN
EA
ST
EP
PE
Y-
AX
IS
LIN
EA
ST
EP
PE
R LO
ST
IC
TA
BL
R
40.5
Fig. 5 A "SANTA CLAUS" MACHINE for producing large display letters. The polyethelene bead from the glue gun is programmed to build up the entire letter one strand at a time. While slow, there are no mold charges and the size and style can be instantly changed. Logos and custom characters are a snap.
Hardware Hacker
LR64700 video compression chips that
VHF & MICROWAVE RESOURCES ARRL Handbook 225 Main Street Newington, CT 06111 (203) 666-1541 Defense Electronics 6300 S. Syracuse Way #650 Englewood, CO 80111 (303) 220-0600 Fair Radio Sales PO Box 1105 Lima, OH 45802 (419) 227-6573 Hewlett-Packard PO Box 10301 Palo Alto, CA 94303 (415) 857-1501 M/ACom Semiconductor 43 South Avenue Burlington, MA 01803 (617) 272-3000 Microwave Journal 685 Canton Street Norwood, MA 02062 (617) 769-9750 Microwave Product Digest 34 Evergreen Place Tenafly, NJ 07670 (201) 568-5835 Microwaves & RF 611 Route #46 West Hasbrouck Heights, NJ 07604 (201) 393-6286 Mini-Circuits Labs PO Box 350166 Brooklyn, NY 11235 (718) 934-4500 Motorola 5005 E McDowell Road Phoenix, AZ 85008 (800) 521-6274 Nuts & Volts Box 1111 Placentia, CA 92670 (714) 632-7721 Plessey Semiconductor 13900 Alton Pkwy, Ste 123 Irvine, CA 92718 (714) 455-2950 Radio Research Instruments 584 North Main Street Waterbury, CT 06704 (203) 792-6666 Rogers 100 S Roosevelt Avenue Chandler, AZ 85226 (602) 961-1382 SGS-Thomson Microelectronics 211 Commerce Drive Montgomeryville, PA 18936 (215) 362-8500 Surplus Trader Winters Lane Box 276 Alburg, VT 05440 (514) 739-9328
ing has been reprinted a number of times and remains highly useful to this day. A pair of more recent classics are Skolniks Introduction to Radar Systems and his Radar Handbook. As with any field, the free trade journals are your foremost day-to-day info source. While there are several dozen, four Ive found useful include Microwaves and RF, the Microwave Journal, Microwave Product Digest, and Defense Electronics. Be sure to check out Mini-Circuits Labs for all their low cost broadband amplifier chips, and Plessey for their unusual new Satellite Cable and TV Integrated Circuit Handbook. They also have lots of info on frequency synthesis and high speed dividers. Hewlett Packard, of course, makes all kinds of microwave components, as well as high performance micro-
wave and VHF test instruments. And Motorola has a wide variety of high frequency semiconductor chips and application notes available. Two other chip sources include SGS and M/A Com, while Rogers supplies printed circuit materials and oddball dielectrics useful for microwaves. Well, that should be enough to get you started. Please let me know what else you think should be added to our resource files.
should revolutionize both still and moving picture image storage and transmission. I do get lots of calls asking about hacker friendly sources for custom crystals. Two of my favorite sources are Statek for low frequencies and Crystek for higher frequencies. Both have catalogs, data sheets, and ap notes available. A newly detailed bibliography on magnetic refrigeration is available through Jerry Hagen at his own Cory Laboratories. Our unusual hacker magazine for the month is Homebuilt Rotorcraft from the Rotary Flight International folks, while that free Maxim Engineering Journal has all sorts of semiconductor goodies in it. Especially for audio/video switches and power supplies that work off a single AA cell. Free samples of Powerpole connectors are available from Anderson Power Products Inc. These are both snap-together modular and sexless. They should be ideal for such things as solar panels, but perhaps not quite rugged enough for use in cave rescue communications. An interesting sandblasting catalog is available from National Sandblast. This is one quick way to spruce up any metal on your prototypes. Turning to my own products, for the fundamentals of digital integrated circuits, do check into my TTL Cookbook and CMOS Cookbook. Or to pick up all the "oldies but goodies" at once, try out my Lancaster Classics Library. We also now have the Hardware Hacker III reprints available, which has the latest and best of all these columns in them. All edited, revised, corrected, and indexed. Finally, I do have a new and free mailer for you which includes dozens of insider hardware hacking secret resources. Write or call for info. Our usual reminder here that most of the items mentioned appear either in the Names and Numbers appendix or the Microwave Resources sidebar. As always, this is your column and you can get technical help and offthe-wall networking per that Need Help? box. The best calling times are weekdays 8-5 in Mountain Standard Time. Lets hear from you. 3
40.6
Don Lancasters
Hardware Hacker
June, 1991
e might start off with the reminder that we do have our RE technical helpline available for your use per the box below. Here is where you can go for tech help, referral to highly qualified consultants, for book and software purchases, and for general off-the-wall networking. Because of time and funding limits, this service is strictly restricted to US Hardware Hacker readers only. Your best calling times are on weekdays 8-5 Mountain Standard time. Before you call, please re-read the entire column and especially the Names and Numbers and Resource boxes. Hardware Hackers calling without a pencil or a pen handy will get chopped up and fed to the cows. You could also reach me by way of my personal BBS, otherwise known as GEnie PSRT. Call (800) 638-9636 for voice connect info. But, please do note that I am an independent developer and author sitting here on my sand dune in the middle of the Upper Sonoran desert. I have very little input to any RadioElectronics editorial policy and know next to nothing about RE projects by other authors. For more info on these, call Brian Fenton at RE editorial (516) 293-3000 , or else you might use the Radio Electronics BBS youll find at (516) 293-2283. Yeah, I sometimes do welcome any visitors that call in advance. But do note that Gurus are supposed to be hard to reach, because (A) it adds to the mystique, and (B) the Gurus and Swamis Union Local #415 rules demand it. At any rate, there is a seven hour drive involved in reaching the nearest airport from here. I wont even mention the deadly Gila Monsters or hostile Indians. Every once in a while Ill get a call that sounds reasonable at the time, but after thinking things over For instance, one Hardware Hacker did want to crystal control that low cost BA1404 FM stereo broadcaster weve looked at in past issues. He wanted to do this so that several actors in a play
Electronic tuning diodes Two unusual newsletters Parametric Amplification Association book resources Preventing modem dropouts
input frequency or voltage. Or you can use ceramic or SAW filters that are so precise and so repeatable that zero adjustment is needed. You also tend to use a few higher quality filters, rather than lots of individually tuned and cascaded LC circuits. But after all of this, there remain times and places where youll still need a few electronically variable capacitors. Selecting a station on a radio or TV are obvious examples. One very popular, well performing, and ultra low cost electronic tuning method is called a varactor diode or simply a varicap. Varactor diodes are diodes which have been optimized to look and behave as a high quality electrically variable capacitor. Most any diode conducts current in the forward direction and blocks it in the reverse direction. Specifically, when you reverse bias a diode, you create a depletion region containing neither electrons nor holes. As you increase your reverse bias, this depletion region gets thicker, and vice versa. Thus, any diode will behave as an electrically variable capacitance as you vary its reverse bias voltage. For most diodes, this unavoidable depletion region capacitance is a flaw that restricts your maximum speed of operation. But in a varactor diode, the depletion region capacitance is purposely made rather large, quite high quality, and very controllable. Figure one shows you a typical circuit. From the electrical control
could be on the same frequency. Uh, whoops. Yew caint get there frum here. Sure, you can crystal control your BA1404. And you certainly can put as many of them as you want on the same frequency. But one of the key properties of FM reception is that you will only receive your strongest station, with virtually zero pickup of any of those others. This normally desirable action is known as the FM Capture effect and is caused by the hard limiting present in virtually all FM receivers. As little as a fraction of a decibel can cause any one signal to utterly and totally dominate. Sorry about that. We return you to our column already in progress
Electronic Tuning
The traditional method of tuning a resonant circuit to your desired frequency is to use some adjustable or variable capacitor. While obvious and cheap, these capacitors are often large and sometimes expensive; could be sensitive to fields and vibration; and usually require human intervention for their use. These days, it is much better to go to electronic tuning methods. The simplest method of electronic tuning is to never do it. As much as possible, you purposely design out any need for a variable capacitor or a direct replacement. For instance, you use digitally synthesized frequencies instead of any local oscillator or BFO. You use switched capacitor filters or active filters that are tunable by an
Fig. 1 ELECTRONIC TUNING USING A VARACTOR DIODE. The doide depletion layer capacitance varies with the applied reverse bias. The large series capacitor serves as a dc block to prevent shorting out the tuning voltage.
41.1
Hardware Hacker
One way to increase the range of a varicap is to cheat and use a lower bias voltage. Your capacitance will usually increase dramatically for very low values of reverse bias. But at this point the diode will start to conduct and very much reduce the available Q or selectivity for your tuned circuit. The linearity will also be awful. For wider tuning ranges, special varicaps are obtainable which have different doping profiles. Varicaps with a medium tuning range have an abrupt doping profile, while those with very high tuning ranges use a hyperabrupt profile. The tradeoffs for a wider tuning range are more nonlinearity, somewhat higher cost, lower circuit Q, and harder tuning. Youll also become vastly more sensitive to noise and the precision of your tuning voltage. The AM broadcast frequencies of 550 to 1650 kilohertz have a 3:1 range. Thus, you should use a hyperabrupt varactor having at least a 9:1 and preferably a 10:1 tuning range here. The hyperabrupt Motorola MVAM108 is one good choice here, having an extreme 15:1 range. Those television frequencies are spread out over a very wide range. To prevent having to tune these all at once, three varactor tuners are separately used in several individually selected circuits. One for the lower VHF channels 2-6 ; a second for the high VHF channels 7-13; and a third for the remaining UHF channels. And sometimes a fourth for special cable channels. Note that there is a big frequency gap between channel 6 and channel 7 that holds the FM broadcast band and emergency services. Should you instead want to restrict the tuning range of a varicap, you can either put a fixed capacitor in parallel with it or else use a narrower voltage control range. The fixed parallel cap is often the better choice. Varactors work best with tuned signals in the microvolt range. Should your signals being tuned get above several millivolts, the signals themselves can add to or subtract from the tuning voltage. Thus, your positive signal excursions will increase the resonant frequency and vice versa. Which can introduce moderate to severe second harmonic distortion. You might resolve this possible
Fig. 2 A PAIR OF BACK-TO-BACK varactor diodes can be used to prevent larger signals from detuning themselves. A positive signal swing raises the capacity of one varactor and lowers that of the other. The changes largely cancel out.
side, you can simply reverse bias your diode by way of a large series resistance or some other method that has a very high RF impedance. When you change the value of this voltage, you electrically change the capacitance of your varactors depletion region, and thus tune your circuit. Typically, you change your tuning voltage over a 3 to 30 volt reverse bias range. On the resonant circuit side, you do have to provide a DC return path to ground for the tuning voltage bias. You also do have to provide a series blocking capacitor to keep any other DC path from shorting out your diode. Normally this series capacitor is very much larger than the varactors capacity, so it does not significantly alter your resonance calculations. Sadly, the varactors capacitance changes nonlinearly with the reverse voltage. Depending on the varactor, you might have 60 picofarads at 1 volt reverse bias, 45 pf at 2 volts, and 18 pf at 20 volts. Thus, your first couple of
volts of reverse bias will by far give you the most variation. The plot of capacitance versus reverse voltage is roughly linear when it gets plotted on semilog paper. Varicap capacities can go from a fraction of a picofarad with exotic microwave devices on up to several hundred or more picofarads for use in audio filters or AM tuning. You can sometimes use giant silicon power diodes for lower frequency varactor experiments. But the Q will often be low when you try this, and the tuning range will be limited. The capacitance range of a varicap is usually defined as the ratio between your 3 and 30 volt bias settings. An ordinary varactor will often have a capacitance range of 3:1. But note that a frequency change varies with the square root of your capacitance change in any resonant circuit. So, this type of 3:1 varactor can only shift a resonant frequency by 1.73 or so. Not all that much.
R1
1N914 R2
Fig. 3 A TEMPERATURE COMPENSATED varactor tuner. The forward drop temperature drift of the ordinary diode can be used to offset the capacitance drift of the reverse biased tuning varactors. Resistor R1 has to be large enough to not load the tank circuit. Resistor R2 adjusts the tempco.
41.2
June, 1991
distortion problem by using a pair of varactor diodes as shown in figure two. The two varactors are in parallel as far as the DC tuning voltage goes, but are in series with any signals being tuned. Thus, on a positive peak, the capacitance of the first varicap will increase as the other decreases, and largely cancel each other out. Dual varactors in a single package are becoming popular. The Motorola MV104 is one example. Note that two capacitors in series give you one half the total capacitance. Be sure to allow for this in your designs. It is extremely important to have a very stable reference for your tuning voltage, since any drift at all could detune your circuit. If possible, you will also want to use some sort of feedback to keep your tuning locked on channel. Various types of AFC (automatic frequency control) can sometimes do this for you. Sadly, varactors drift over temperature. Their values will increase with an increasing temperature. One typical value is in the 200 parts per million range. On the other hand, a regular diode that is forward biased should have a current-dependent drift with a negative tempco. Figure three shows how to add an ordinary diode in series with your tuning voltage to temperature compensate a varactor diode. The load resistor of the diode is adjusted to give a minimum overall drift. Sometimes a parallel capacitor can also be added having a chosen tempco. A precise temperature compensation over a wide range could get tricky. At the least, everything has to be tightly heatsunk together. More details appear in the Motorola Ap Note #AN551. As weve just seen, Motorola is one leading supplier of a wide variety of low cost and easy-to-get varactor diodes. Some cost under a dollar. See their RF Device Data II handbook for data sheets and ap notes. ultra low noise, high gain, and high frequency amplifier that used nothing but a diode sounded like a great idea. Today, low noise and high gain microwave transistors are a buck each from such outfits as Mini-Circuits Lab and Avantek. Paramps are largely limited to esoteric ultra-microwave lab circuits and for optical and infrared experiments. Although I do strongly suspect youll shortly see a stunning resurgence of paramps in a brand new application area. Figure four shows you how the parametric amplifier works. This is exactly the same idea as pumping a swing on the playground. The local oscillator called the pump frequency causes the capacitance of a varactor to change in a time varying manner. The "parameter" here we are varying is the diode capacitance. A low level input signal known as the input frequency is also routed to the same time varying capacitor. A filter extracts an output signal known as the idler frequency. The net result can be a very strong and low noise amplification and a possible frequency conversion for your input signals. Since a purely reactive capacitor is in use, there are theoretically none of those noise problems associated with resistance or the traditional tube or semiconductor amplifier circuits. I once used a plain old three cent 1N914 computer diode to produce 20 decibels of gain and a 2 decibel noise figure at 600 MHz. The diode was dc back biased to -3 volts or so and a suitable pump frequency was capacitively superimposed. Long ago and far away. Paramps can be designed as upconverters with gain, as downconverters with loss, or as negative resistance devices with potentially high gain but possible instabilities. The key math behind paramps is known as the Manley-Rowe relations. One horses mouth classic on this subject is Blackwell and Kotzebues Semiconductor Diode Parametric Amp- lifiers, published by Prentice Hall, way back in 1961. Included is a good summary of Manley-Rowe plus an extensive bibliography. For more mod- ern info on paramps, check out the Dialog Information Service.
input
VARACTOR DIODE
PUMP OSCILLATOR
Parametric Amplifiers
Surprisingly, varactor diodes were not initially designed for electronic tuning. Instead, they were created for a unique beastie once known as a parametric amplifier. Back in the days of tube style UHF tv tuners with 35 decibel noise figures, the idea of an
Fig. 4 A VARACTOR DIODE PARAMETRIC AMPLIFIER in which a diode provides low noise amplification. The pump frequency causes the varactor capacitance to vary in such a time-dependent "parametric" way that its interaction with the input frequency produces an amplified output at the idler frequency.
41.3
Hardware Hacker
There are zillions of professional societies, most of which do seem outrageously expensive and who simply do not deliver what they think they do. On the other hand, many of these groups have association bookstores that do have outstanding tutorials, reprints, and information real hard to find elsewhere. To get you started, our resource sidebar for this month lists a few of the professional organizations. Here is a brief rundown The American Association for the Advancement of Science publishes the "must have" Science magazine and provides outstanding book reviews. The Association of Energy Engineers carries dozens of books including a Small Power Production Manual and Efficient Electrical Systems Design. The old Institute of Electrical and Electronic Engineers is a rather poor performer that is way overblown and far overrated to the point of being a ripoff. But their reference books do include Visual Communication Systems and Navigation: Land, Sea, Air, and Space. The American Society of Materials has hundreds of books on just about anything mechanical. Their Metals Handbook series are classics. The Society of Motion Picture and Television Engineers is big on time codes, on video production, and other hard-to-find resources. The Journal of the Audio Engineering Society is your foremost resource for any music synthesizer theory, and all other audio. And weve already seen that the Society of Automotive Engineers has lots of great automotive electronics books. And finally, one real sleeper, that Society of Photo-Optical Instrumentation Engineers which publish a Milestone Series which includes such topics as Fiber Optic Gyroscopes, Digital Image Processing, and Ultrahigh Speed Photography. Naturally, there are bunches more where these come from. For the first of our two contests this month, simply tell me about any professional organization in any field that has book or publication resources of possible interest to hardware hackers. There will be all the usual Incredible Secret Money Machine book prizes, with an all-expense-paid (FOB Thatcher, AZ) tinaja quest for the very best. picks up any extension phone could blast you off the air. This can get especialy frustrating and infuriating just before the end of a long and costly upload or download. Radio Shack offers a product called their 43-107 Teleprotector. It costs eight bucks and its only tiny problem is that it may end up working too good for you. Any inactive telephone line has around 48 volts of direct current on it. When the first phone is picked up, the working voltage drops down into the 6-8 volt range. So, if you simply put a plain old voltmeter across the red and green wires on your extension phone, you can quickly tell if the phone line is in use by another extension. Figure five shows you the circuit for their teleprotector, along with some possible mods. An extension phone normally uses only the red and green wires. The teleprotector breaks the green wire and inserts a full wave diode bridge. The purpose of this full wave bridge is to make the current in the use-sensing circuit always go in the same direction, even during the ringing or if the phone polarity somehow ends up backwards. Initially, there is only a very high value 390K resistor inside the bridge. This is much too large a series value to let the phone work. When you pick up the phone, your full available line voltage initially appears across this resistor. Should you have at least 18.7 volts available (meaning that there is no other extension in use), your SCR (Silicon Controlled Rectifier) turns on, stays on, and will let the extension operate. If another phone is being used, there will not be enough voltage to turn on the gate of the SCR, and the extension will remain off. So far so good. Any SCR will stay on only so long as its main anode to cathode current drops to zero. This particular SCR can get turned off one of two ways. If you hang up your protected extension, the current in the green wire obviously drops to zero and resets you for the next call. But theres also that mysterious 8.7 volt zener in series with your SCR . Now, should some other phone get picked up, there wont be enough voltage left on the line to keep the zener con- ducting, so the SCR turns off, as does the extension phone.
Teleprotection Surprises
One of the ruder surprises youll discover when you first start using a modem extensively is that anyone who
MODULAR PLUG
black yellow red
MODULAR JACK
green
Fig. 5 CIRCUIT DIAGRAM for the Radio Shack 43-107 teleprotector, along with two possible mods. Unmodified, the extension phone is only allowed on the line when no other phone or modem is in use. The BYPASS switch will defeat the teleprotector entirely; the ME FIRST switch lets your extension remain active if another phone gets picked up during a call.
41.4
green
2N6565
June, 1991
Thus, your protected extension phone will never turn on if another phone is off hook. Should you be using your protected extension phone and should another one get picked up, you will get unceremoniously cut off the line! All of which tells us that their teleprotector works exactly as they advertise. There is no way to eavesdrop with a teleprotected extension phone. There is also no way to let the teleprotected extension blast your modem off the air. Or a FAX machine for that matter. Another trick is to hook up their teleprotector up to your answering machine. When you pick up your remote extension, your answering machine drops off the line, stopping the now unwanted message. Which also can be handy. One key point: Your teleprotector does not go on the modem phone. It goes on the interfering extension. Now for the big problems. Your teleprotected extension can not be used for any conferencing! You also cannot have a receptionist answer by picking up with their teleprotected extension and then listening in long enough to verify the correct person picked up the call. Two possible circuit mods are also shown in figure five that might be more suited to you should you need conferencing, but your main goal is to eliminate any modem blasting. A BYPASS switch completely defeats the teleprotecting. Whenever any two phone conversations are desired, you slide this switch into its BYPASS position, shorting the works out. You do, of course, have to remember to flip the switch back after any needed two phone conversations. Closing the ME FIRST switch lets the protected phone initiate or receive a call so long as it is the first phone picked up. It will stay active should additional phones be picked up, so two-phone conferencing can still be done. But the protected phone still can no longer break in on a modem or fax in progress. If you really want to get hairy, you could even let your protected phone purposely break in on any ongoing voice conversation when in the ME FIRST position. Just shout at the other person to pick up their phone. Once
ASSOCIATION BOOKSTORE RESOURCES American Association for the Advancement of Science 1333 H Street NW Washington, DC 20005 (202) 326-6400 American Society of Materials PO Box 473 Novelty, OH 44072 (800) 368-9800 Association of Energy Engineers PO Box 1026 Lilburn, GA 30226 (404) 925-9558 Audio Engineering Society 60 East 42nd Street, Rm. 2520 New York City, NY 10165 (212) 661-8528 Institute of Electrical and Electronic Engineers 445 Hoes Lane Piscataway, NJ 08855 (201) 981-0060 Society of Automotive Engineers 400 Commonwealth Drive Warrendale, PA 15096 (412) 776-4841 Society of Motion Picture and Television Engineers 595 West Harsdale Avenue White Plains, NY 10607 (914) 761-1100 Society of Photo-Optical Instrumentation Engineers PO Box 10 Bellingham, WA 98227 (206) 676-3290
you are sure their phone is picked up, hang your phone up for half a second, and pick it up again. Both of you should now be in conference. But my particular setup requires three phones, any-way conferencing, and no modem blasting. As a second contest for this month, lets have your thoughts on how to elegantly handle this problem. The Radio Shack teleprotector unit is cheap enough that you probably would not want to build one up from scratch. But should you decide to experiment completely on your own, that oddball 2N6565 sensitive gate SCR is a Teccor product available by way of Mouser Electronics.
consumer electronics items for use on an ongoing royalty basis is Mark Gottleibs DesignTech. Two rather impressive "labor of love" high quality newsletters are Don Parhams Homebuilt Rotorcraft for the Gyro Gearloose crowd; plus Ed Zim- mers Inventor-Entrepreneur Network Newsletter, thats intended mostly for Michigan hackers. A pair of interesting trade journals include Medical Electronic Products and Medical Equipment Designer. Yes, they have info on brain wave electrodes. But do make certain you know what you are doing before you hotwire your neighbors cat. Turning to my own products, for the fundamentals of microprocessors and microcomputers check into my Micro Cookbook I and Micro Cookbook II. Or, to pick up all those "oldies but goodies" at once, do check into my Lancaster Classics Library. We also now have the Hardware Hacker III reprints available, which have the latest and best of all these columns in them. All edited, revised, corrected, and indexed. Finally, I do have a new and free mailer for you which includes dozens of insider hardware hacking secret resources. Write or call for info. Our usual reminder here that most of the items mentioned appear either in the Names and Numbers appendix or in the above sidebar. 3
41.5
Hardware Hacker
41.6
Don Lancasters
Hardware Hacker
July, 1991
Power electronic resources Video crosshatch generator More wavelet breakthroughs Secrets of video compression Understanding transformation
certain new hardware hacking doors permanently closed to you if you do not learn to use and love transforms? Well, a transform is simply finding a different way of looking at some problem. Done properly, a transform should let something "magic" happen. What kind of magic? That depends. Transforms may just change things so they appear differently; may break something down into its fundamental parts; may extract non-obvious stuff in non-obvious ways; could help you separate signals from noise; may let you do more with less; or can perform any of a number of other unique (and often unexpected) tasks. For instance, you can either listen to music or else transform that music to notes-and-staff notation on a piece of paper. Beside it being a compact way of representing a song, the sheet music notation lets you know what is going to happen next. It also, for better or worse, standardizes the way the song is supposed to be played. These days, most transforms are done digitally. As figure one shows us, to do a transform, you take a pile of numbers in a grouping called an array and then apply some rule or set of
do seem to get several helpline calls each day from those of you who just "cant find" a technical paper at your local library. Its getting so bad that I am thinking of starting up a special "Couldnt find a pig in a dishpan" contest. Pretty near all libraries provide the Interlibrary Loan Services, through which you can get any copy of any paper from anywhere. All you have to do is ask. Theres also the UMI folks that can very quickly ship you a copy of anything from anywhere. Finally, at extra cost, you can instantly get a copy of any paper by way of the Dialog Information Service that is becoming available at most libraries.
More on Wavelets
For those of you that came in late, theres a math revolution taking place that we are intently following as an ongoing subject here in Hardware Hacker. This involves wavelet theory, a profound and an earth-shattering improvement on klunky old Fourier transformation which is certain to totally change just about any and all advanced electronic topics. Theres now a brand new book on wavelets that aims at becoming the "horses mouth" classic. It is known as Wavelets (a catchy title, eh what?). And authored by a long list of name brand wavelet biggies which start with G. Beylkin. It is published by Jones and Bartlett. It is not easy reading. Meanwhile, a spunky new outfit by the name of Aware Incorporated is -er- making a large splash in actual out-the-door wavelet products. They have some wavelet demo and nice simulation software that you can run on a PC, and they seem to be the first ones out the gate with new wavelet transform accellerator chips. Their initial chips are intended for video compression of 512 x 512 images to either eight or eleven bit accuracy. There seems to be several dozen new wavelet papers being released per month now, so it really is a zoo out there. This new wavelet tool is so fundamental and so important that
everybody from the oil exploration geologists to cardiologists appear to be showing intense interest. Most of these wavelet papers are, of course, totally unreadable. But to tune yourself into them, you use the Dialog Information Service from your local library, and zeroing in on their Inspec, Compendex, or Mathsci data bases. A monthly update of all the latest key abstracts should cost you under $20 . While you can get the actual papers through Dialog, they are much cheaper through UMI, and cheaper still through the Interlibrary Loan service. Wavelets operate by transforming arrays of numbers into other arrays of numbers. Unlike the classic Fourier transformations, you can simply and quickly zero in on specific portions of a transformation problem, and can do so both locally and globally. You also can eliminate "windowing" problems.
Understanding Transforms
Before we take a new look at video compression, which is one of hottest emerging new wavelet uses, lets ask the even more fundamental question: Just what is a transform? Why are they so crucially important? And why are
Fig. 1 A DIGITAL TRANSFORM takes an array of numbers that usually represents something from the real world. It then applies a math rule or rules to those numbers to create a new but related array of numbers in an imaginary transform space. Certain "magic" things will happen in the transformation process that gives you elegant new ways to solve electronic problems.
42.1
Hardware Hacker
unlike many other languages, PostScript automatically can do so continuously, invisibly, and on the fly, maintaining a separate user space and a machine- dependent device space. Figure two shows you a simple linear graphical transform that can convert plain lettering into isometric lettering. Lets look at some numbers. The linear graphical transform takes any point x,y in your original artwork and converts it into a new point x',y' by using these formulas x' = Ax + Cy + E y' = Bx + Dy + F
Fig. 2 THE LINEAR GRAPHICS TRANSFORM is both a simple and very important example of transform use. Here ordinary graphics and lettering have gotten changed into isometric lettering by using the math shown.
OM S
C RI T
rules to all these numbers. These rules will most often involve multiplying, adding, trig functions, shifting, and other math stuff. Your result is a second array that holds transformed "answers". You might call your first array the real world plane and the second array the transform plane. Think of them as two different worlds. If you can undo what you just attempted with your transform, the transform is said to be lossless, and your reverse process is called an inverse transform. For instance, you might take a picture and transform it. You might
then modify the new numbers in the transform plane to extract the edge information, to recognize patterns, count objects, look for differences, or deblur the image. A final inverse transformation could then give you a "new" picture which has changed in some way from the original. One fairly simple yet extremely important transform is known as the linear graphical transform. This is used to change your sizing, position, rotation, and shape of an image. The PostScript computer language is especially adept at making both linear and nonlinear transforms. And,
Values A through F are plain old constant numbers for any selected transformation. If we think of these as some array of [ A B C D E F ], your A value determines the horizontal scale. The B value decides how much climb you are going to get. And C decides how much lean. The D value sets the vertical scale. The E value provides for a horizontal shift, while the final F value sets the vertical shift. For instance, you might use A and D to separately change the width and height of a font character or message. A positive C value would make your message oblique or italic, while a negative C would do reverse italics. Similarly, a positive or negative E can give you a superscript or subscript. And a sneaky trick involving sines and cosines lets you rotate an image, again just by altering A through F in just the right way. Specifically, if is your angle of rotation, just use A = cos B = sin C = -sin D = cos E= 0 F= 0
POPULARITY
white
Fig. 3 A HISTOGRAM OF A TYPICAL VIDEO PICTURE can show you the "popularity" of each gray level. Since no use gets made of the untransformed edge, pattern, texture, and area info, only a very small amount of useful data compression is actually possible.
Which may look ugly but works like a champ. And in general, the linear graphics transform lets you translate, scale, or rotate most any 2-D object. The point here is that most other transforms work exactly the same way. You can start off with some pile of numbers, apply some rules to them, and end up with a second pile of numbers. And some very good things can happen along the way. The granddaddy transform of them all is known as the Fourier transform, while its computer implementation is called a Fast Fourier Transform or one of its improved offspring. Fourier
42.2
July, 1991
transforms move you from the real world frequency domain on over into the time domain. As an example, if you use Fourier transforms to analyze a square wave, you will break the wave down into a fundamental sinewave and weaker third, fifth, seventh, and other odd higher harmonics. Similarly, you can take a pile of sinewaves and add them together by using the inverse Fourier transform in order to get the actual squarewave back. Finding the frequencies is called analysis; Building the waveform is known as synthesis. Fourier transformation and all its newer offspring get used for nearly all the advanced electronic applications. These include detecting earthquakes and heart attacks, robotic vision, side looking radar, in video compression, picture deblurring, for high definition television, music synthesizers, sonar, geophysical exploration, holograms, and spectrum analysis. As weve seen, the real excitment behind wavelets is that it does all of the things that Fourier transformation already does. Only wavelets do it far faster, far better, and much cheaper. Wavelet transformations also seem adept at localizing certain features, such as the leading and trailing edges of a square wave. Which is something that gives Fourier fits. Probably the hottest example of wavelet mania today involves And what is even more exciting, it seems that researchers in human and animal vision are freshly discovering that either the wavelets or something very much akin to them is involved in natures way of resolving the same types of visual and aural problems. So, something really heavy seems to be coming down here. What makes a picture different from completely random pixels? The usual answers are edges, patterns, and large areas where the colors and shading change little if at all. If a picture is nearly random and if you must have everything in that picture, then there isnt much you can do to squash it down to a smaller size. On the other hand, if there are plenty of edges, patterns, and large areas, and if some amount of visual degradation can be allowed, then you can often squash your picture down considerably, possibly as much as 100:1. Figure three shows you a histogram of a typical image. In this case, there are 256 gray levels, and each level gets used by the height of its bar as shown. Despite your usual image having lots of edges, patterns, and big areas, the histogram looks more or less like random noise. But, apply a wavelet transform to your same picture, and you get the surprising histogram of figure four. Wavelets are chosen that zero in on extracting key edge, pattern, and area information. The amazing result is that most of the numbers in your new histogram are zero and thus can be safely ignored! An array with lots of zeros in it is called a sparse data set. By going to a fancy Huffman or any similar coding scheme, sparse data sets can be stored or transmitted much more compactly than an array in which every value is equally important. There is usually a two step process involved in video compaction. You first try to concentrate all the picture energy in as few array numbers as is possible, zeroing out as many of the others as you can. Then you code the final results as densely as you can. As a reminder, a wavelet transform works just like any other transform. You start off with the first pile of numbers, play several simple math games with them, and end up with the second pile of new numbers. In the wavelet case, your math games are simple additions, shifts, and rotations. Math which is far less hairy than the fancy trig needed with conventional Fourier or DCT Transforms. There are many popular wavelets, and new ones are getting discovered daily. A typical compression wavelet need act only on a very few adjacent values in your original array and can use only simple shifts and adds. The
Video Compression
We are smack in the midst of a multimedia revolution that will let you directly manipulate all studio quality pictures, artwork, and even superb animation on any personal computer. The big problem here is to squash the information content of a high quality color image down into a manageable space without degrading the picture too much. The HDTV high definition television people are grappling with their version of the same problem How can you get all of those pixels squashed down into an acceptable transmission bandwidth? Wavelets promise an exciting new solution to video compression. One that is fast, cheap, easily implemented in hardware, massively parallel, low on artifacts, and can give you compressions as high as 100:1.
POPULARITY
negative
zero
positive
Fig. 4 WAVELET TRANSFORMED HISTOGRAM OF THE SAME PICTURE creates a "sparse" data set having many zero values, by concentrating and redirecting the picture energy. Yet EVERY detail of the original picture is fully and exactly preserved! Extensive compression can now be done.
42.3
Hardware Hacker
wavelets are usually applied to rows in the array, and then reapplied to the resultant columns. So far, we have thrown nothing whatsoever away. With some sneaky coding, your exact picture can now be stored in one third or less the space as the original. Or modemed in one third the time or broadcast in one third the bandwidth. And you can reverse the process at any time to get your exact picture back. The eye can resolve some things quite well, and others poorly. You simply do not see some details, and others are too subtle to worry about. By zeroing out several more of the remaining numbers, and by being crude about how accurately you code the rest of them, you could trade off picture quality versus compression. And your changes in quality are progressive and subtle. Instead of picking up obvious "jaggies" as you reduce your pictures information content, you simply add broadband noise to it. Which is a much more pleasant and far less noticable form of image degradation. There are two papers on wavelet video compression that Ive found both readable and useful. The first is Application of Compactly Supported Wavelets to Image Compression from William Zettler and the rest of their crew at Aware. Free single copies are available from them on request. Secondly, while nearly anything on wavelets by John Daugman is a good choice, his Entropy Reduction and Decorrelation in Visual Coding by Oriented Neural Receptive Fields, in the IEEE Transactions on Biomedical Engineering, volume 36 , number 1 , January 1989, is outstanding. Dont let the long title scare you away on this one. It is both a good intro to wavelet video compression and a look about how nature may be already using the wavelets to solve thorny visual problems and similar edge effects. Should you want to explore the concept of histograms further, check out my HSTOGRAM.PS, available on GEnie PSRT as file #146. Discrete Cosine Transform. A crew known as the Joint Photographic Experts Group is in the final throes of making the DCT a standard. You can pick up a reprint copy of the JPEG Technical Standard, Rev. 8 through Adobe Systems if you want further details. This is a "must have" document if you are at all interested in video compression. At any rate, an awful lot of people have spent an awful lot of time and money to make the DCT a major new video compression standard. And the PostScript Level II now has built-in DCT filters, and dozens of firms are offering or about to offer DCT compression solutions. The only little problem here is that the wavelet transforms pass the DCT as if it was sitting up on blocks. It appears that anything that the DCT can transform, wavelets can do faster, cheaper, better, and with far fewer objectionable artifacts. In particular, there are background "tiling" effects using the DCT , but none using wavelets. Your wavelet transform computations are quite simple, and much easier to realize in standard or near-standard hardware. Wavelets are also good at building up the "big lumps" first, followed by finer detail. This can get important in video animation. So, I guess we are going to shortly see the answer to the question "What happens when a big bunch of people spend lots of time and effort creating a new standard which gets clearly and ludicrously obsolete before the ink on the final version is dry?
+5V .033 1Kx4 20K 100K bar width adjust +5V 3.3K gray adjust 5K gray bars horizontal lines crosshatch white dots vertical lines
16 15 14 13 12 11 10 9
ZNA234E
GND X1 SYNC VID GRAY TST +5V OSC2
1 2 3 4 5 6 7 8
video output 1K
video level
Fig. 5 THIS MONOCHROME TV PATTERN GENERATOR uses the Plessey ZNA234E to generate sync, gray bars, lines, dots, and crosshatching.
42.4
July, 1991
An old line distributor by the name of Galco does seem to be the best one-stop source for most power semis. Their Galco Gazette is now free, while their 1500 page full line catalog is a refundable $15. Several of the many insider trade journals addressing power electronics include PCIM, Motion, NASA Tech Briefs, and Motion Control. Hacker surplus power electronics seems catch-as-catch-can. But you can check Fair Radio Sales, C&H Sales, Surplus Traders, or else Herbach and Rademan. Several of the more significant manufacturers of those higher power semiconductors and support products include Asea Brown Bovari, Siemens, Powerex Inc., Power Semiconductors, SGS, Plessey, International Rectifier, Motorola, Teccor, and Semikron. Most of these have extensive data books and ap notes available.
POWER ELECTRONICS RESOURCES Asea Brown Bovari 1460 Livingston Avenue N. Brunswick, NJ 08902 (201) 932-6000 C & H Sales 2176 East Colorado Blvd Pasadena, CA 91107 (213) 681-4925 Fair Radio Sales PO Box 1105 Lima, OH 45802 (419) 227-6573 Motion Box 6430 Orange, CA 92613 (714) 974-0200 Motion Control 800 Roosevelt SE-408 Glen Ellyn, IL 60137 (708) 469-3373 Motorola 5005 East McDowell Phoenix, AZ 85008 (800) 441-2447 Power Semiconductor 600 West John Street Hicksville, NY 11802 (516) 933-3000 Semikron PO Box 66 Hudson, NH 03051 (603) 883-8102 SGS 1000 East Bell Road Phoenix, AZ 85022 (602) 867-6259 Siemens 2191 Laurelwood Road Santa Clara, CA 95054 (408) 980-4500
NASA Tech Briefs Galco 41 East 42nd St. #921 26010 Pinehurst Drive Madison Heights, MI 48071 New York, NY 10017 (212) 490-3999 (800) 521-1615 Herbach & Rademan 401 East Erie Avenue Philadelphia, PA 19134 (215) 426-1700 International Rectifier 233 Kansas Street El Segundo, CA 90245 (213) 772-2000
Surplus Traders PCIM 2472 Eastman Avenue #33 Winters Lane, Box 276 Alburg, VT 05440 Ventura, CA 93003 (514) 739-9328 (805) 658-0933 Powerex Hillis Street Youngwood, PA 15697 (412) 925-7272 Teccor Electronics 1801 Hurd Drive Irving, TX 75038 (214) 580-1515
One good source for Mac repair parts and good service information is Soft Solutions. And Rolf Taylor does offer Apple II repairs. Exchange parts for both families are available through Shreve Systems. A new line of ultra high quality and fully professional RF ground plane breadboards is available from RF Prototype Systems. For a rather unusual mechanical resource Kindt-Collins offers all sorts of industrial waxes, you can use for machining, CAD/CAM , moldmaking, investment casting, protection, and similar needs. Even water soluable waxes are offered. A pair of wildly different ongoing hacker magazines that are going great guns are Elector USA, chock full of hands-on construction projects, many with a European or continental bent; and Jeff Duntemanns really great PC Techniques, an intense software and hardware computing resource. Two newsletters of interest to any of you digital ham radio types include Digital Digest and RTTY Journal.
If you are at all interested in the PostScript language, do start out with my LaserWriter Secrets book & disk combo, Adobes "blue" PostScript Cookbook, or their "red" PostScript Reference Manual II. All three are available through my Synergetics, or ordered via GEnie PSRT email. We also now have the Hardware Hacker III and Midnight Engineering I reprints available, which have the latest and best of all these columns in them. All edited, revised, corrected, and indexed. Finally, I do have a new and free mailer for you which includes dozens of insider hardware hacking secret resources. Write or call for info. Our usual reminder here that most of the items mentioned appear either in the Names and Numbers appendix or in the Power Electronics sidebar. As always, this is your column and you can get technical help and offthe-wall networking per that Need Help? box. The best calling times are weekdays 8-5 in Mountain Standard Time, year round. Lets hear from you. 3
42.5
Hardware Hacker
42.6
Don Lancasters
Hardware Hacker
August, 1991
et us start off with several updates to the infrared people detectors we looked into a column or two ago. It seems Amperex/Phillips decided to jump ship and abandon their line of IR detectors. Their great ap notes seem to have been discontinued. Several remaining pyroelectric detector chip samples are supposed to still remain available through their (401) 762-3800 sales support line. On the other hand, people detectors have instantly become a low cost commodity. So much so that it is now usually cheaper to buy retail and modify, rather than building your own motion sensor up from scratch. Especially when you factor in the critical Fresnel lens design. One big cause for the sudden price drop is a new California law which mandates these on nearly every light switch in new office building construction. Herbach and Rademan offers a great and ready-to-install infrared people detector as #TM90SE1491 for $12.50 . And the Heathkit folks have several new occupancy and intrusion detector variations in their free home security electronics catalog. Or, you might also want to check your local price club or a warehouse builders supply outlet. A reminder that we have a great help line available for tech questions, off-the-wall networking, and useful consultant referrals. All per the box below. US callers only, please. And please remember that we are using California time during the summer. Those 4:30 am calls tend to get just a tad old. And please, please carefully read the entire column, especially the Names and Numbers sidebar. Oh yeah. Please note the French mathematicians name is "Fern-ell", not "fresh-nell." And, of course, those correct pronounciations found on our ongoing voice helpline are "hee-lah" and "ten-ah-hah". Our biggie for this month involves telephone caller identification chips and standards. But, first, lets find out how you go about
Electric power research info Focused X-ray breakthrough C-60 and the new Buckyballs Getting telephone information Caller number delivery secrets
who is calling you is infinitely more important than protecting your "right" to make undetected obscene phone calls. Very sadly, at least one state (Pennsylvania) has stupidly banned this wonderful new service. In other areas, the caller is given the option of blocking their caller id, for those one-in-a-thousand calls when your anomynity might legitimately be desired. Maybe for a drug overdose hotline. Blocking can get done by entering a three digit code before you make your call. So what is caller id, and how does it work? Figure one shows you some of the more interesting Bellcore documents which involve caller id. But the key horses-mouth paper youll need is Technical Reference TR-TSY-00030 and titled SPCS Customer Premises Equipment Data Interface. The cost is $25 via VISA. Figure two shows you the exact placement of the tone codes. Caller number delivery is normally provided as a code burst between the first and second full rings while your phone is still on its hook. Enough delay gets provided to allow for the short half second breaks sometimes involved in selective ringing. After your first full ring gets detected, a data path is established. A data path is simply any method of receiving some modem tones. The id tones will then get routed to suitable circuitry to recognize the data burst and strip out the needed information. The format and codes are related to
1. NYNEX Catalog of Technical Information, #NIP-7400, Free. 2. SPCS Customer Premises Equipment Data Interface, #TR-TSY-0030, $25. 3. CLASS Feature: Calling Number Delivery #FSD-02-1051, $30. 4. CLASS Feature: Calling Number Delivery Blocking #TR-TSY-000391, $33 The main Bellcore service number is (800) 521-CORE. Of these resources, (2) is by far the most important.
Fig. 1 HERES THE KEY INSIDER SECRET PAPERS involved with telephone system caller number delivery.
43.1
Hardware Hacker
squarewave at 600 Hertz. Your caller-id receiver circuitry is supposed to use this "wake up call" and string of marks to prepare itself for valid data reception. After setup, the first valid data byte is called a Message Type Word. If this burst is for the caller number delivery, your message type word will be a hex $04. Other codes could get used for other purposes. For instance, a hex $0A means "message waiting" for pager applications. The next byte is called the message length word and tells you how many digits are to be provided in the caller id numbers that are to follow. This message length does not include itself or the checksum in its count. The actual ASCII characters for the caller phone number follow, starting with the least significant digit. For instance, a digit "2" should get ASCII coded as hex $32. Your message ends with a check sum word used for an optional error detection. The checksum is the twos complement of the 8-bit sum of the message type word, your message length word, and the number of data words provided. To provide an error detection, you run your own 8-bit sum of all the bits of all the data words, and then add the check sum to it. If there are no transmission or reception errors, you should get a zero result. A non-zero result means an irrecoverable error. Ive just described the single data message format to you. Some nasty complications can arise if you have multiple data messages or several CLASS services active, such as call waiting. See the Bellcore papers. There are a number of caller id devices commercially available, with pricing in the $60 to $500 range. Typical are the Allied incoming call identifier, the Bellsouth calling line identifier, the Cidco Slimline series, and the TC-1021 , TC-1080 , and the TC-1082 . One discount source of all these is VSI Telecommunications. Do note that your caller number delivery service must be provided to you before you can use any of these devices. If the code is not being sent, there is no way you can receive it. There might be certain state laws prohibiting their use as well. This, of course, can be cured by staking your
6 seconds
Fig. 2 THE CALLER ID DATA BURST is sent between the first and second rings as shown. The transmitted level is -13 DBM and enough delay is provided to allow for distinctive ringing patterns. A typical number id will take half a second. The phone must remain on hook while the id is being received.
the digital codes used in pagers. The tones sent out are plain old serial binary modem tones at 1200 baud. A digital logic one (or a mark) is defined as one cycle of a 1200 Hertz sinewave. A digital logic zero (or a space) is defined as nearly two cycles of a 2200 Hertz sinewave. Each eight bit character is preceeded by a space start bit and a mark stop bit. The transmitted signal level is -13.5 DBM. While the code transmission is asynchronous, phase coherence is preserved throughout the entire caller id message. No more than 20 marks are permitted between characters. Figure three shows you the data format. Which might appear fairly complicated at first, but making sure
the number is valid is quite important. Almost certainly, you will elect to use a computer, or at the very least, some simple microcontroller in your caller id display circuitry. The first thing sent is called the channel seizure signal. This consists of 30 bytes of a Hex $55 code. After decoding, another way to look at this "hey wake up!" signal is one quarter second of a 600 hertz square wave. After the channel seizure signal, a carrier signal is sent. This is used to condition your receiver for valid data. This carrier consists of at least 150 milliseconds worth of marks. After the decoding, you have a one-sixth second solid string of logical ones following your quarter second burst
CHANNEL MESSAGE MESSAGE CARRIER SIEZURE TYPE LENGTH SIGNAL SIGNAL WORD WORD
DATA WORDS
The CHANNEL SIEZURE SIGNAL is made up of a string of 30 bytes of Hex $55, and is equivalent to 250 msec. of a 600 Hertz square wave.
The CHECKSUM WORD is the twos-complement 8-bit sum of the message length and data words. It can be used for error trapping. The DATA WORDS are the actual caller id digits in ASCII character form, least significant digit first. Each digit byte is preceeded by a zero "space" start bit and followed by at least one "mark" stop bit. The MESSAGE LENGTH WORD is the number of characters being provided as the calling number.
The CARRIER SIGNAL consists of 150 mlliseconds of marks, equal to digital logic ones.
The MESSAGE TYPE WORD is a Hex $04 when providing for the identification of a calling number.
43.2
August, 1991
state representative to the nearest anthill. Finally, anything you connect to the telephone line has to be FCC Type 68 approved. Meeting the rules for type approval is fairly reasonable and simple. But the approval process itself is a bear. We do have several experimental caller id projects in the works here at Radio-Electronics . Along with some simulator and test software. But there are ongoing problems with service availability, with meeting FCC part 68 specs, and in keeping the price low enough. Probably some absolute minimum general purpose caller id interface kit that demands use of a personal computer programmed in machine language to develop your own circuits makes the most sense at this time. ones and zeros. The third id stage is the number extractor. The number extractor can take your serial data stream, makes sure it is valid, and then extracts and formats your calling number. This number might also be displayed. The number extractor almost has to be a computer or microcontroller, since it probably would be unbearably complicated and expensive otherwise. The final and optional stage is any "Gee Whiz" stuff. Things like looking up the actual name of the caller. Or pulling their sales or service records. Or keeping a full record of the last hundred calls. Or interacting with a humongus CD-ROM data base. Once again, be sure to let us know what you want to see in the way of further caller id projects. and foods. It bonds readily with many other elements, creating by far the richest assortment of useful chemical compounds. Until recently, there were only two known forms for any pure carbon molecules. These were the supersoft graphite and superhard diamond. But several chemists (Kroto, et al. Nature v318 p162 (1985)) have discovered a uniquely bizarre third form of carbon molecule. By taking 20 hexagonal groupings of six carbon atoms and then fusing them properly with 12 pentagonal groupings of five carbon atoms, a hollow geodetic pure carbon molecule of 60 atoms gets created. Since this hollow all-carbon C-60 molecule looks like a geodetic dome or a socker ball, they were promptly given the name of Buckyballs, named after the late Buckminster Fuller, a leading early proponent of geodetic dome structures. Initially a few C-60 molecules were painfully hand collected, measured, and tested. And, sure enough, C-60 is for real. An experimental proof of the Buckyball shape has recently shown up (Hawkins, 2Science, v252, p312) , along with a fine bibliography.
A Caller ID Chip
As figure four shows us, there are usually four stages to the caller id receiver. The first stage is called a FCC Part 68 Interface. This one is needed to safely and legally connect to your phone system. Full details on Type 68 interface secrets appear in the Hardware Hacker II reprints. The second stage forms the analog front end. The analog front end can provide filtering and amplification, and then converts all your tones into actual TTL ones and zeros. Figure five shows you an analog front end circuit that uses the brand new Sierra Semiconductor SC11211N caller id chip. You input the low level telephone tones and get out a serial data stream ready for your computer or a microcontroller interpretation. Inside the chip is a fancy filter, an energy detector, a tone demodulator, and the clock. An ordinary tv color crystal and four capacitors is all you need to use this circuit. Cost for this 16-pin minidip is in the $4 range. Theres also a fancier SC11210 chip in a smaller package that deletes the internal crystal oscillator, all the level setting, and the energy detection features. Sierra also has ap notes on suitable FCC part 68 interfaces. While these chips are low in cost and work rather well, note that they are only analog front ends. "All" they do is reliably accept low level tones from the part 68 interface and then convert them into a string of digital
PHONE LINE
NUMBER EXTRACTOR
Converts the serial data stream into numbers that can be displayed.
43.3
Hardware Hacker
After the discovery, several other chemists (Kratschmer, et al. Nature v347 on p354 (1990)) were playing around with carbon rods in a plain old arc welder and figured out how to make lots of Buckyballs in a process almost as complicated as smoking a piece of glass with a candle. As a result, the price of Buckyballs has plummetted to an astonishingly low $1200 per gram. Quantity pricing is even less, and 100 milligrams of Buckyballs are available for $250 if you are on a limited budget. You can get all the Buckyballs you want off the shelf from Materials and Electrochemical Research. The latest games include putting Osmium "handles" on a Buckyball to make them easier to align, and adding precise doping impurities to create several new families of "Dopeyballs". While nobody has yet come up with a proven commercial use for Buckyballs, they are now by far the hottest topic in physical chemistry. Buckyballs will even superconduct, although only at rather low temperatures so far. The hollow molecule suggests all sorts of stunning new possibilities which include super lubricants, new batteries, ultra-strong fibers, improved semiconductors, and entire new classes of materials and compounds. Needless to say, Buckyballs are a sure fire winner for a Science Fair topic or school research paper. And we here at Radio-Electronics would certainly be most interested in any Buckyball project which was even remotely releated to electronics. Let us know what you can do here. Ongoing info on Buckyballs will appear in those Science and Nature magazines, and, of course, through the Dialog Information System. consists of scads of glass capalaries, and the X-rays skip along the surfaces at low angles, rather than shooting right on through them. From National Semiconductor, a new General Purpose Linear Devices data book. And from SGS-Thomson a new Power Transistors Application Manual. Good stuff. Satco is a great source for oddball tools and supplies. They usually sell to those school vocational education programs. They do have a great new thousand page catalog available. Unusual scientific images of all sorts of geologic, oceanographic and atmospheric stuff is available on a GRIPS-2 CD-ROM from Meridian Data. It is part of a new program to release obscure government research to a wider audience. Our unusual trade journal of the month is Power Quality, aimed at the higher end users of uninteruptable power supplies. Chip capacitors and resistors for surface mounting are now finally becoming widely available in small quantities at sane pricing. While Mouser Electronics is a good source, some ready-to-go $49.95 resistor and capacitor kits are available through Communications Specialists. Thats three cents each for the resistors and fourteen for the caps. Turning to my own products, for the fundamentals of digital integrated circuits, do check out my TTL Cookbook and CMOS Cookbook. Or you can pick up all of those "oldies but goodies" all at once in my Lancaster Classics Library. I also do have a BBS up as GEnie PSRT. Besides all the PostScript and desktop publishing stuff, youll find all sorts of ongoing Hardware Hacker and Midnight Engineering resources here. You can get your voice connect info by dialing (800) 638-9636. We also now have the Hardware Hacker III, Ask the Guru III, and the new Midnight Engineering I reprints available, which have the latest and best of all these columns in them. All edited, revised, and indexed. I do have a new and free mailer for you which includes dozens of insider hardware hacking secret resources. Our usual reminder here that most of the items mentioned appear in the Names and Numbers sidebar.3
CALL PROGRESS INPUT SERIAL DATA OUTPUT ENERGY DETECT OUTPUT +5V nc
14 13 12 11 10 9 8
SC11211
VIN VIP EL1 EL2 EDC VRF GND
1 2 3 4 5 6 7
+ 0.1F 0.1F 1F
Fig. 5 A CALLER ID ANALOG FRONT END that makes use of the Sierra Semiconductor SC11211 chip. Low level tones are accepted from the FCC Part 68 interface and get converted into serial ones and zeros for further digital processing.
43.4
Don Lancasters
Hardware Hacker
September, 1991
et us first pick up on several updates to some of our earlier Hardware Hacker topics. One good source for those BA1404 FM stereo broadcasting kits is DC Electronics. They also stock the super new Signetics NE602 mixer/converter chips and the TEC-200 film for direct toner printed circuits. Another NE602 source is Active Electronics. Telephone caller ID is certainly one hot topic these days. And yet another source of call identifier magic boxes is Hello Direct. Prices start at $60. You must, of course, have the ID service available before you can use these magic boxes. States that have at least some local availablility of caller ID should now include AL, CA, FL, GA,
Driving inductive loads More on phone caller ID Bakerizing and laminating Alternators as stepper motors Programmable logic resources
that the diode appears "backward" so that it does not normally conduct any supply current. If you suddenly try to turn off the inductor current, a small induced voltage will immediately be created that in turn forward biases and turns on the protection diode. The current you had before can then continue on through your protection diode and back into the relay coil. The current will now drop on down to zero fairly quickly, dissipating itself in the forward drop of the diode and in the internal resistance of your relay coil. At no time is any voltage spike generated that exceeds the 0.6 volts or so of your diode forward drop. This simple diode despiker works quite well. But there are some minor side effects that can sometimes cause problems. Note that your relay will stay pulled in for a brief time delay after you thought you turned it off. This happens because there is still diode-provided current going through your coil. In a larger relay, this time delay could extend a few tenths of a second, and could cause you timing problems. The physical dropout of your relay can also end up slower and sloppier. Which could cause contact arcing in
+
your magnetic field changes, youll generate an induced voltage. And the faster the change, the more voltage you create. Since any current through a coil can generate a magnetic field, any change in your coil current should produce a change in the magnetic field, which in turn will induce a voltage spike. The greater or the more sudden the change in the current, the greater the induced voltage. The basic math here says that e = L i / t or, in English, your induced voltage across any coil is proportional to the size of the inductor and the rate of change of current through the coil. Say you decide to control a relay. You turn your relay on by sending a current through your coil. And then you attempt to turn your relay off by suddenly disconnecting your coil current. What happens? Your magnetic field will suddenly collapse, generating a horrendous voltage spike. You tried to make t zero, and, since youre now trying to divide by zero, you get a theoretically infinite voltage spike. Thus, suddenly ceasing the current in any inductance is guaranteed to create a humongus voltage spike. Sometimes you might choose to purposely do this. For instance, the current through the coil in any car ignition is suddenly broken to step up the 12 volt battery into many tens of thousands of volts of ignition spark voltage. And a related technique gets used for television high voltage. But, should you suddenly cease a current through any coil in any solid state circuit, the voltage spike youll get is almost certain to blow up the transistor of whatever happens to be controlling your coil. The rule here is simple: Never let the current through an inductive load suddenly drop to zero in any solid state circuit! Figure one shows you how to add a plain old power diode to your relay coil to provide spike protection. Note
IL, IN, MD, ME, MI, NC, NE, NJ, NV, OH, OK, SC, TN, VA, VT, WV, and Washington DC. Other areas are still
in the planning stages. Most services are still for local calls only. One handy and rather non-obvious benefit of this new service: When you come back from lunch, you have a complete and a time stamped list of everyone who tried to call you when you were out. Most useful. I thought we might round up a big collection of odds and ends for this months column
INDUCTIVE LOAD
DESPIKING DIODE
Fig. 1 ANY SEMICONDUCTOR can be instantly destroyed if you use it to suddenly turn off the current in an inductive load. The despiking diode shown here allows the coil current to continue long enough to safely dump all the magnetic flux energy without creating a killer transient.
44.1
Hardware Hacker
combo as an "inefficient diode" with a
+
DESPIKING DIODE INDUCTIVE LOAD ENERGYDUMPING ZENER
Fig. 2 A SERIES ZENER diode shortens and sharpens the dropout time. This minimizes turn-off delay and contact arcing in power relays. Your control transistor must be able to block the supply voltage PLUS the zener voltage.
higher current applications. Your protection diode also should turn on fairly fast. If you use a slow diode or if there is not enough stray circuit capacitance around, a large and possibly destructive spike can build up during the time your diode actually starts conducting. Figure two shows you an improved spike protection circuit. Here we have added a 24 volt zener diode in series with your protection diode. This combination will conduct no current in one direction and will conduct in the other direction only when the voltage across it exceeds the zener breakdown of 24 volts. You could think of this series
+
INDUCTIVE LOAD
TRANSIENT PROTECTOR
Fig. 3 ELECTRONIC TRANSIENT suppressors are bidirectional devices which are also known as varistors, MOVs, TRANSILS, or several other trade names. These have to be very carefully matched to the allowable voltage rise and the magnetic flux energy to be dumped.
Whenever you suddenly disconnect your relay current, a large but acceptable 24 volt high voltage spike is created, which turns on both diodes in the series pair. The current continues through the diodes, but will fall to zero much faster as you now have a 24 volt drop burning up much of your remaining coil energy. Thus, this circuit will still give you protection, but will shorten the excess holding time by a factor of 40 or so. Your contact release will also be that much faster. What happens is that youve now made a tradeoff. You are allowing a reasonable sized spike in exchange for a large reduction of the release time. Note that your transistor will see a maximum voltage of your supply voltage plus the drop of that zener during break time. For instance, on a 12 volt supply, your transistor would have to block at least 36 volts if it is not to be damaged. There are special back-to-back Zener-like components intended for spike protection. These go by the names of varistors, MOVs , or transient protectors and do have various brand names. They work the same way as figure two in that they do not conduct until spike time. Then they do conduct heavily and thus internally dissipate the inductors flux energy. Figure three is a typical circuit. SGS is one of many suppliers of the TRANSIL spike protectors. Their BZW04P23 is typical. At 25 volts or under, it draws only five microamps. Above 30 volts it starts conducting heavily, and by 41.5 volts it draws at least ten amperes. Despite the tiny package, these devices can stand fifty amps for ten milliseconds. Higher power units are also available. You will have to carefully match your protection device to the size of the spike you are willing to allow and the amount of energy that you need to dump from your coil. If at all possible, you should also isolate your coil drivers from more sensitive parts of your circuit. Optoisolators and individual power supplies are great for this. More info on spike protection and computer interfacing appears in my Micro Cookbook, volume II.
44.2
September, 1991
hold onto it when you do. The kick is definitely there. At best, I would guess that you could not get as much useful force with a car alternator than you can by using a much smaller "real" stepper motor. And the alternator would end up ridiculously slower to boot. Id like you to try and prove me wrong. Either as this months contest or for a winning school lab project, experiment with a car alternator and find out exactly how useful a power stepper motor you could convert it into. Could you in fact create a $5 machine tool power feed with one? How fast can you go? How much output force can you get? How good are the steps? What is the best computer interface? Therell be the dozen or so of our usual Incredible Secret Money Machine book prizes, along with a big all-expense-paid (FOB Thatcher, AZ) tinaja quest for two going to the very best of all. OK. Figure four shows you some conversion hints. Most real stepper motors do use a permanent magnet rotor. With an alternator, you would use this field winding and slip rings for a giant electromagnet, running as much current through it as you can without overheating. This forms a group of seven shaft attached magnets that you can rotate to a desired position by activating the stator coils. And speaking of which, there are normally three sets of stator coils. These are usually hooked up in what is known as a three-phase wye circuit. For stepper use, you will want to find the floating splice where your wye connection is made and bring it out as a separate positive terminal. Which should then give you three distinct and independent winding sets. Lets call the windings A, B, and C. Power the field via the slip rings, and pulse winding A . Keep your current down around an amp or two at first to prevent any overheating or driver problems. The rotor will align itself with the nearest pole piece and should lock itself to some position. Now turn off winding A and activate winding B. The rotor should now jump one step clockwise. Turn off B and whap C . And you should jump yet another step clockwise. To step on clockwise, use an ABC sequence. To step counterclockwise, just use ACB instead. Youll probably want to keep at least one winding energized at all times so that you can hold a position when not stepping. Note that you could end up in position A, B, or C, depending on the power sequence. You will have to remember where you are with your controller or host computer. A typical alternator should give you 21 possible positions, and a resultant step angle of around seventeen degrees. You would disconnect all of the alternator diodes during your conversion. These could later get used as protection diodes serving the driver circuits if you do not have anything better available. One hint: Your slip ring brushes can "explode" whenever you take any alternator apart. And the two brush springs will fly off into the hinterlands. Look closely, and youll find a toothpick size hole in the insulated brush holders. To reassemble, you just put a toothpick or a stiff wire through the hole to hold the brush springs compressed. Done just right, you should be able to remove the toothpick from the alternator outside after you do your reassembly. The rotor winding of an alternator is a fairly high resistance, usually around 5 ohms or so. It will safely current limit itself. But your stator windings are an extremely low impedance, typically under a quarter ohm. Thus, you must externally limit your rotor currents to keep things from burning up. Plain old power resistors are a good way to handle this for your early experiments. There are some tricks you could pull to improve the number of steps per revolution. One would be to allow two windings to be active at once. This could double your resolution. A fancier technique would be to allow several different values of the current for each winding. This leads to various new microstepping opportunities. Microstepping is a proven concept with real stepper motors. Yet another possibility is to use a
STATOR A
STATOR B
STATOR C
ROTOR
5,10W
5,10W
5,10W 3,10W
Fig. 4 CAN A CAR ALTERNATOR get converted into a $5 power stepper for a machine tool drive? Only hackers know for sure. To experiment, use the rotor as a powered electromagnet. Find the floating stator wye connection and bring it out as a power terminal. Pulse one stator winding at a time in an ABC (clockwise) or an ACB (counterclockwise) sequence. Be sure to safely limit stator currents. Especially for all of your initial tests.
44.3
Hardware Hacker
bridge type drive which lets you run current either way through any given winding. You could now use both attraction and repulsion, and, once again, should be able to double the number of steps. And one final resolution enhancer would be to put a nutplate of some sort on the shaft end, creating a linear stepper. A threaded shaft through the nutplate will then move forward or backward as the alternator steps. For instance, with a 1/4-20 thread, each full revolution would advance you fifty mils. One single step at 21 steps per revolution would advance you a mere 2.38 mils, besides giving you a really major mechanical advantage to boot. Simple gearing could also be used to increase your step resolution. But you ll have to watch out for backlash if you attempt this. And do note that the finer the resolution, the slower your max allowable operating speed. Sorry about that. You might also like to look at some dual current scheme that gives you a brief high current pulse when stepping and some lower holding current between the actual steps. After you really have your stepper working reasonably well, youll want to increase all the operating currents. Overheating, saturation effects, and overshoots set your ultimate limit. It might also be very interesting to rewind all your stator coils. Use lots more turns of a much smaller diameter wire, and try bridging only a single stator pole rather than three. Ampere turns is the name of the game here. You could also try improving the rotor flux paths and air gaps. For further resources on power stepping in general, check out Airpax and Slo-Syn for iron, the PCIM and Motion trade journals for info, or SGS, Sprague, or else Motorola for useful drivers. One distributor that stocks a wide selection of power electronics stuff is Galco. And good old J.C. Whitney has bunches of alternators and parts available, including some rewound 100 amp stators that go for around $18. You will find lots more on power electronics in our brand new Hardware Hacker III reprints. Theres lots of possibilities here. Lets see what you can come up with. integrated circuits. Some (such as
EPROMs ) are based on exhaustive
BAKERIZING
With BAKERIZING, your toner image is temporarily placed in close contact with a smooth film. Heat and pressure is then applied. The toner becomes much blacker and takes on a beautiful medium to high gloss. Some films can be reused.
LAMINATING
With LAMINATING, a thin clear plastic overlay is permanently attached to your copy by using heat and pressure. This process seems ideal for menus and book covers, or where extreme scuff resistance to "lock in" your toner is required.
Fig. 5 THESE TWO "MAGIC" FILMS can dramatically improve both the appearance and durability of any toner image.
table lookups. While others (such as PLAs) do multilevel AND-OR digital logic using combinations of gates and registers for their logic. Some are one-shot programmed by blowing fuses. Others are easily reprogrammable. And yet others use a flash technology where the needed connections can get saved to a nonvolatile RAM memory during power down times. The prices of some programmable chips are now down in the $2 range. The important advantages of these new devices are that you can place the exact circuit you want into one or two packages; that they are easy to debug and modify yourself; that some can recalibrate or update themselves later on in-circuit; and that you are (at least temporarily) the sole source of your "secret" inside programming connections. All with instant delivery and no staggering setup charges. Ive tried to gather together some of the more popular programmable logic suppliers into this months resource sidebar. You might want to get data books and individual ap notes from many of these sources. One side note that is both alarming and sad. Many of these houses refuse to divulge the key programming info needed to use all their chips. While I consider this a monumentally stupid way to cut off your nose to spite your face, their argument is that they most definitely do want you to use some "approved" programmer that is more likely to keep their products reliable in any final use circuits. Heres a trick that may help you get programming info should reasonable and direct tries fail. The magic new buzzword in chip programmers these days is "DAC per pin". Just tell those application engineering people that youre now about to ship production quantities of your new pc-based DACper-pin programmer, and that you would like to be able to include programmability for their chips. Our helpline has been full of horror stories on programming info, but this ploy seems to work so far. At least till they catch on. Please do continue to send in your horror stories on this topic. More details on the use of EPROMs
44.4
September, 1991
appears in my CMOS Cookbook and TTL Cookbook, while the listings in our resource sidebar and all the usual trade journals should also have lots of application info.
PROGRAMMABLE LOGIC RESOURCES Actel 955 East Arques Avenue Sunnyvale, CA 94086 (800) 227-1817, ext. 60 Altera 2610 Orchard Parkway San Jose, CA 95134 (800) 800-EPLD Advanced Micro Devices PO Box 3453 Sunnyvale, CA 94088 (800) 222-9323 Catalyst 2231 Calle de Luna Santa Clara, CA 95054 (408) 748-7700 Cypress Semiconductor 3901 North First Street San Jose, CA 95134 (800) 952-6300 Dept. C4P Fujitsu 3545 North First Street San Jose, CA 95134 (800) 642-7616 Gould AMI 2300 Buckskin Road Pocatello, ID 83201 (208) 234-6668 Intl. CMOS Technology 2125 Lundy Avenue San Jose, CA 95134 (408) 434-0678 Intel 1000 Business Ctr Drive Mt. Prospect, IL 60056 (800) 548-4725 Lattice Semiconductor 5555 NE Moore Court Hillsboro, OR 97124 (800) FAST-GAL Mitsubishi 1050 East Arques Avenue Sunnyvale, CA 94086 (408) 730-5900, ext. 2106 National Semiconductor 2900 Semiconductor Drive Santa Clara, CA 95051 (800) 272-9959, ext. 40 NEC Corporation 401 Ellis Street Mtn. View, CA 94039 (415) 965-6000 GEC Plessey 13900 Alton Pkwy S123 Irvine, CA 92718 (800) 927-2772 Raytheon 350 Ellis Street Mountain View, CA 94039 (415) 962-7930 Texas Instruments PO Box 117692 Carrollton, TX 75011 (800) 336-5236, ext. 3707 Toshiba 1220 Midas Way Sunnyvale, CA 94086 (800) 321-1718 Waferscale Integration 47280 Kato Road Fremont, CA 94538 (415) 656-5400 Xicor 1511 Buckeye Drive Milpitas, CA 95035 (408) 432-8888 Xilinx 2100 Logic Drive San Jose, CA 95124 (408) 879-5199
Bakerizing Secrets
Despite my screaming it from the rooftops for quite a few years now, Bakerizing seems to remain a top secret procedure. Which amazes me because anyone that sees what this process does is absolutely stunned. Bakerizing is a sneaky trick you can apply to most any laser toner image, especially all of your PostScript laser printed output. Bakerizing instantly makes toner a high gloss jet black, while at the same time dramatically increasing the durability and crease resistance. While excellent for business cards, Bakerizing works nearly anywhere you want a blacker and denser image. And the process is nearly free! All you have to do to Bakerize is take a sheet of special high gloss, slip coated polyester film. You then temporarily place this film in contact with your toner image and then apply heat and pressure. This remelts your toner while contacting a mirror smooth surface. Which then can produce a calandering process similar to that ferrotype drum on older darkroom glossy print processors. The simplest way to Bakerize is to put the magic sheet on top of your hard copy and shove it back through your printer while imaging a blank page. The best way to Bakerize is to use a Kroy Color or Canon fuser unit or some sort of laminating machine. But even a plain old iron can sometimes be used effectively. With care, any single Bakerizing sheet can be reused dozens of times. Which can drop your per page costs to a few pennies per copy. I have gone ahead and custom ordered a zillion sheets of Bakerizing film and will be offering it as a stock product here at Synergetics. Ive also now got some more traditional yet toner compatible laminating materials suitable for nice looking menus and book covers. Figure five shows you how these two films differ. Give me a call if you want to play with either of these really exciting new materials.
Lets see. Ive just reprinted my Incredible Secret Money Machine II, along with its new intro and update section. And we are now Book-ondemand publishing our Hardware Hacker III, Ask the Guru III and my Midnight Engineering I reprints. A reminder about my new BBS up as GEnie PSRT. Besides all of that PostScript and desktop publishing stuff, youll find all sorts of ongoing Hardware Hacker and our Midnight Engineering resources here. You can get your voice connect info by dialing (800) 638-9636. Finally, I do have a pair of new and free mailers for you. One includes dozens of insider hardware hacking secret resources, while the other is on PostScript and Desktop Publishing. Write or call for info. As usual, most of the items mentioned here appear in the Names and Numbers appendix or in the Programmable Logic Resources sidebars.3
44.5
Hardware Hacker
44.6
Don Lancasters
Hardware Hacker
October, 1991
ust how could you go about meeting FCC specs? Contrary to common belief, the Federal Communications Commission regs are very easy to find and are not really all that expensive. Most federal regulations appear in the humongous group of paperback volumes known as the CFR and short for their Code of Federal Regulations. Each government bureacracy has their own title in the CFR series. For instance, their title 14 is for Aeronautics and Space and title 37 is the Patent and Trademark office. Of our interest here, title 47 is the FCC regulations that govern any use of radio spectrum space, intentional or otherwise. You can view most CFR volumes in most any large library that has a government documents section. The volumes are also now conveniently available from the US Government Bookstore. Yes, VISA and MC are accepted, with one week delivery. The FCC regulations are organized into a hundred separate parts. Figure one shows you the more popular parts for topics of any hardware hacking interest. For instance, computers and peripherals have to follow part 15 . Telephone interconnection has to obey part 68. Citizens band is part 95, and amateur radio is part 97 . And those new direct broadcasting satellites are now part 100. There are five volumes in the title 47 series, whose costs range from $10 to $20 each. These are arranged as parts 0-19 , parts 20-39 , parts 40-69 , parts 70-79, and parts 80-102. Typical charges for a registration, certification, or a type approval are usually in the $300 to $1500 range, depending on whether any testing is needed. More details on fees appear in parts zero and one. Meeting FCC specs on your own can be monumentally frustrating and be horrendously expensive. Besides taking forever. There are specialists who are willing to help you for a fee. The foremost trade journal for this is called Compliance Engineering.
Curve fitting fuzzy data Tesla Coils and resources Meeting FCC regulations Flashlight battery supplies Thoughts on E field machines
built up around Tesla. After years of careful reading, I have yet to find even the slightest credible shred of evidence that (A) Tesla was an extraterrestial alien, (B) Tesla invented a unique free energy perpetual motion machine, or that (C) All Teslas real inventions got suppressed by a paranoid government conspiracy. For our resource sidebar for this month, Ive tried to gather together some of the more useful sources of accurate materials and information on both Tesla and Tesla Coils. Try the Tesla Book Company for a well done collection on books and publications on and about Tesla that range the gamut from historical fact on down through utterly absurd fewchips-shy-of-a-full-board fantasy. Lindsay Publications also stocks some of the better and more factual titles, as does Industrial Micro. One leading supplier for museum quality Tesla systems is Resonance Research. Their Modern Resonance Transformer Design Theory book is a
Part 0 Part 1 Part 2 Part 5 Part 13 Part 15 Part 18 Part 21 Part 22 Part 23 Part 25 Part 68 Part 73 Part 74 Part 78 Part 80 Part 87 Part 90 Part 94 Part 95 Part 97 Part 99 Part 100
Structure and organization of the FCC. Rules, fees, certification, type approval. Frequency spectrum allocations. Experimental radio services. Commercial radio operations. Radio frequency devices (includes computers, peripherals). Industrial and medical equipment. Public service fixed communications. Public service mobile communications. Shortwave broadcasting. Satellite communications. Telephone terminal equipment. AM and FM broadcasting. Auxiliary AM and FM broadcasting links. CATV cable television. Maritime and shipping communications. Aviation communications. Private mobile services. Microwave relay. Personal and CB citizens bands. Amateur radio. Disaster and emergencies. DBS Direct broadcast satellites.
Fig. 1 FEDERAL COMMUNICATIONS COMMISSION regulations are presented in title 47 of the Code of Federal Regulations. Here are how the more useful and popular parts are arranged.
45.1
Hardware Hacker
creating the insulation materials, the surface treatments, and the vacuum needed to support ultra strong E fields remains a very thorny problem. On the other hand, for free energy, all you need is an insulated stationary vertical antenna several miles high to tap the Earths 100 volts-per-meter calm day E field. Then you hack up some sort of efficient DC-DC downconverter at the bottom. Uh Huh. Sure.
Fig. 2 THE SWITCHING MODE STEP-UP CONVERTER. Briefly but repetitively closing the switch causes the current in the inductor to ramp up to a high value. On release, the high current is transferred to the load resistor, appearing as a voltage that is always higher than the input. The switching duty cycle sets the amount of voltage step-up you will get.
classic, and they do offer interesting design software and videos. Good sources for Tesla Coil parts include Edmund Scientific, JerryCo (now renamed American Science and Surplus), and Surplus Traders. One newsletter that has lots of hands-on low end Tesla info in it is Tec-Spec. One interesting hacker club is that Tesla Coil Builders Association. At one time before the personal computer revolution, building a Tesla coil for a Science Fair was a required rite of passage for any Hardware Hacker. Although these induction coils still can be spectacular and lots of fun, they are really nothing but an oversize auto ignition coil or a tv flyback transformer. Outside of the science museums, induction coils just arent that big a deal any more. Not to mention the incredible radio and television in-
terference that a sloppily done Tesla coil generates. Or the X-ray danger. Is there any point in hacking ultra high voltages? Typical electrical (or "E") fields have a magnetic (or "H") field associated with them. In theory, you could build up your motors and generators using only the electric field, using only the magnetic field, or using any combination of the two. But look around you, and youll find just about every common motor or generator is a magnetic one that uses the H field nearly exclusively. The only obvious exceptions that prove the rule are such nanopower niche items as piezo fans. Most previous hackers who have explored the E field have now found electric field machinery to turn out extremely large, besides having unacceptably low power densities and even lower efficiencies. To this day,
4.7 mHy +
14 13 12 11 10 9 8
MAX655
LX2 V+ VCC VRF LBO LBI GND
1N5818
.001F 1 F 220 F
+5 VDC OUTPUT
Fig. 3 THIS MAXIM POWER SUPPLY KIT gives you a tightly regulated +5 volts DC at 170 mils or less from a pair of ordinary alkaline AA cells. The secret is a second internal switching regulator that gets used to raise the efficiency of the main one. Do not substitute any parts.
45.2
October, 1991
first hardware hacking project. One warning: The two inductors and the Schottky diode are critical. If you do not chose exactly the parts supplied by Maxim or their exact recommended substitutes, your circuit can end up either inefficient or not work at all. Random junkbox inductors tend to have too high a DC resistance and may saturate. Fortunately, all three of these are cheap parts. In quantity, the entire supply can cost well under $6. Maxim also offers a MAX654 that runs on a single AA cell and can start at an incredibly low 1.15 volts. By the way, AA flashlight cells provide considerably more energy at lower costs in less volume than by using a 9 volt battery. Except when you need ultra low currents for a very long time. Which can give the 9 volt approach a slight edge. More on this in the free Maxim Design News.
Fig. 4 SMOOTH CURVES CAN be quickly and drawn through fuzzy or noisy engineering data by using the PostScript language and the least squares fitting method. Here are three popular types of curves which can give good fits to many hardware hacking problems.
documented PostScript code all ready to run as my file #294 FUZZYFIT.PS on GEnie PSRT. Typical downloading costs should be twenty cents. The best curve for any noisy data
usually uses what is known as a least squares fit. All this means is that you try to spread the errors out as evenly as you can. Your main reason for squaring is to make all the errors
(A) The LINEAR or FIRST ORDER least squares fit uses an equation of y = a1x + a0 Here are the two linear equations in two unknowns that you need to solve to get the best fit for your data (sumx1)a1 + (#pts)a0 = (sumy)
(sumx2)a1 + (sumx1)a0 = (sumx1y) (B) The QUADRATIC or SECOND ORDER fit uses an equation of y = a2x2 + a1x + a0 Here are the three linear equations in three unknowns that you need to solve to get the best fit for your data (sumx2)a2 + (sumx1)a1 + (#pts)a0 = (sumy)
(sumx3)a2 + (sumx2)a1 + (sumx1)a0 = (sumx1y) (sumx4)a2 + (sumx3)a1 + (sumx2)a0 = (sumx2y) (C) The CUBIC or THIRD ORDER least squares fit uses an equation of y = a3x3 + a2x2 + a1x + a0 Here are the four linear equations in four unknowns that you need to solve to get the best fit for your data (sumx3)a3 + (sumx2)a2 + (sumx1)a1 + (#pts)a0 = (sumy)
(sumx4)a3 + (sumx3)a2 + (sumx2)a1 + (sumx1)a0 = (sumx1y) (sumx5)a3 + (sumx4)a2 + (sumx3)a1 + (sumx2)a0 = (sumx2y) (sumx6)a3 + (sumx5)a2 + (sumx4)a1 + (sumx3)a0 = (sumx3y) (D) For QUARTIC or HIGHER ORDER least square polynominal fits, just extend the math in the obvious direction for n linear equations in n unknowns. Solve for the coefficients. Little is usually gained by going beyond a third order polynomonial. Especially in two dimensions. (E) For least square fits to OTHER EQUATIONS, write out the least squares error equation. Then take partial derivatives with respect to each coefficient and set them to zero. This again leads to n linear equations in n unknowns.
Fig. 5 HERE IS THE "MAGIC" MATH behind least squares curve fitting. My FUZZYFIT.PS PostScript program on GEnie PSRT does it all for you quickly and painlessly. Just shove in the raw fuzzy data and out pops a smooth curve.
45.3
Hardware Hacker
The (sumxy) expression means to multiply the first x value by the first y value. They you multiply the second x value by your second y value and sum them. And so on. Fortunately, PostScript does all of your grunge work for you. Quickly, invisibly, and automatically. Around half a second to do the whole job. Say you have five noisy data points and want to place the smoothest and most accurate possible cubic curve through them. All you do is [2 4 4 11 6 14 12 16 16 12] findcubicfit And thats all there is to it! In this specific instance, your first data point is at x = 2 and y = 4. Your second data point is at x = 4 and y = 11, and so on. The findcubicfit operator then will automatically generate your needed magic a 0 , a 1 , a 2 , and a 3 best fit coefficients for you. Among the other handy PostScript routines in my new FUZZYFIT.PS are plotxyvalues, which automatically draws the actual data points on the graph, and ploteqn, which sketches the smoothest possible curve through your points. Theres no sane limit to the number of data points allowed. Much of your typical engineering data gives a good fit to a cubic polynominal. Especially stuff that is only slightly nonlinear. Note that, roughly speaking, your a 0 coefficient is strongest at your low valued end of all the data. Your a 1 coefficient sets the initial slope of your data. Your a 2 coefficient works best in the middle, while your a 3 coefficient has by far the strongest influence at the high value end. You can easily go beyond a simple cubic fit. If you know the underlying physical laws to your lab data, you can try and make your least squares fit the expected math. Other fuzzyfit curves of engineering interest include exponentials, sinewaves, statistical distributions, hyperbolas, circles, and ellipses. Your same "set the partial slopes of the error equation to zero" idea works well with these. For graphics and typography, you can step up to a fancier method of curve fitting that permits loops and cusps. This goes by the name of cubic splines, or Bezier Curves. PSRT has lots more on this exciting topic.3
TESLA COIL & HIGH ENERGY RESOURCES American Sci & Surp 601 Linden Place Evanston, IL 60202 (708) 475-8440 Edmund Scientific 101 East Gloucester Pike Barrington, NJ 08007 (609) 573-6250 Industrial Micro Box 596 Logan, UT 84321 (800) 359-0466 Lindsay Publications PO Box 12 Bradley, IL 60915 (815) 935-5353 Resonance Research E11870 Shadylane Road Baraboo, WI 53913 (608) 356-3647 Science Probe 500-B Bi-County Blvd Farmingdale, NY 11735 (516) 293-0467 Surplus Traders Winters Lane, Box 276 Alburg, VT 05440 (514) 739-9328 TecSpec 3 Mine Mountain Road Cornwall Bridge, CT 06754 (203) 672-0123 Tesla Book Co Box 121873 Chula Vista, CA 91912 (805) 646-3371 Tesla Coil Builders Assoc 3 Amy Lane Queensbury, NY 12804 (518) 792-1003
positive, even if the data point lies below the curve you are fitting. You have lots of choices of what type of curve to fit to your data. One class of curves that make good fits to typical engineering graphs and most typography are usually called polynomonials. These sometimes will also go by the name of a Power Series Approximation. Here is a typical polynomonial y = a 0 + a1x + a2x2 +a3x3 + You pick as many terms for your polynomonial as you think you need. For instance, a single a 0 term would give you a horizontal line with half the points above and half below. Dual a0 + a1x terms give you a slanted line, again splitting the points. Three terms gives you a quadratic or parabolic approximation. And four terms for a cubic. While you can go beyond a cubic, youll usually gain surprisingly little extra in doing so. One way to derive a least squares algorithm is to write out an equation for the total error for a given number of terms. Then you try to minimize this error by finding out how to separately minimize the error created by each individual term. To find a minimum in an equation or on a curve, you find the slope of your curve and set it to zero. Math freaks will immediately recognize this minimization process as taking
the partial derivatives with respect to each coefficient and setting them to zero. Which results in a humongous mess known as n linear equations in n unknowns. You then solve the mess as best you can. Out pops your coefficients needed to plot your best fitting curve. I personally use an old method involving determinants. The needed tools appear in FUZZYFIT.PS , while the tools and a detailed background tutorial is separately available as GEnie PSRT file #289 LINEAREQ.PS. Incredibly, only one single pass is needed to find the best possible curve, no matter how many terms in your polynomonial or the number of points involved! Figure five does show you the equations to use for a linear slope, a quadratic, and a cubic fit. These are the ones you would be most likely to use. Im sure Ive just horrified all the genuine math freaks in our audience with my really weird notation here. But all (sumx3) means is to take your first x data value and cube it. Then take the second data value and cube it and add it to the first one, and so on for all the data. By (sumx3), I simply mean the sum of the cubes of each individual noisy x data value. #pts is simply equal to your total number of noisy x,y data point pairs that you are trying to put your smooth curve through.
45.4
Don Lancasters
Hardware Hacker
November, 1991
ell start off with our usual reminder that this is your column and you can get tech help and off-the-wall networking per the box below. Your best calling times are often weekdays 8-5, mountain standard time. We have some really heavy stuff for this month, so lets have at it
SMPTE video time codes Electronic dog tag contest Case & enclosure resources Solar energy breakthrough? Photovoltaic panel bargains
with the waste heat. Or trying to reshuffle or "downconvert" your solar spectrum to the magic frequency. But these seem mostly laboratory pipe dreams that, in my opinion, are unlikely to see the light of day. In fact, the existing solar electric power plants are actually being torn down, owing to low efficiencies and really bad economics. Many hackers should be interested in the great bargains in the used (and somewhat degraded) solar panels now being offered wholesale by Carrizo Solar and retail by Surplus Traders. Instead, "what if" our sun was a radio source? You would simply get an antenna and a rectifier, and high direct solar-to-electric conversion efficiencies can now be yours. This is called a crystal set, and all of the technology does appear to be fairly well proven. Efficiencies near 100
appears to be a fundamental and unavoidable physical limit. Out in the real world, silicon solar panels are much less than ideal, and you are very lucky to get an overall long term system efficiency above ten percent. And often less. Which is dangerously close to the seven percent "breakeven" level, below which any solar panel will never pay for itself, owing to the materials and labor that go into the system, and the time value of the money used to finance the construction. At seven percent efficiency, a one meter square solar panel will generate around half a kilowatt hour per day, or around $5 of electricity per year. Yes, there are tricks you can pull to raise your efficiency. Such as using a pair of different semiconductors having different work functions. Or using concentrators. Or heating water
(A) LATERAL TIME CODE
This older and nearly obsolete code needs an auxiliary audio channel and outputs one 1200 baud code string per video frame. The tape MUST be in motion for reading. The code is synchronized to begin on the fifth horizontal scan line of the first field in the frame. The low state equals 0 IRE units and the high state equals 80 IRE units. The data stream is 80 bits long and follows the format of figure 2. It uses Manchester, or biphase encoding that is self-clocking and can be read in either direction. Various bits in the data are assembled into BCD words identifying the frame, the seconds, the minutes, and the hours. Other bits are used for synchronization, time adjustment, simple parity and special use flags. Reference: SMPTE 12M-1986 (B) VERTICAL INTERVAL TIME CODE This newer and widely used code is embedded each field on single horizontal lines during retrace. The code gets repeated two, and sometimes four times, beginning on line 10 of each fields vertical retrace interval for improved noise immunity. The code begins 10.5 microseconds into the line. The tape can be stationary or moving in either direction. There are 90 bits to the code, sent at a stationary rate of approximately 1.8 MHz. The format appears in figure 3. It also uses Manchester, or biphase encoding that is self-clocking and can be read in either direction. Various bits in the data are assembled into BCD words for the field, the seconds, the minutes, and the hours. Other bits are used for sync, time adjustment, full CRC error checking, and special flags. Reference: SMPTE RP-108
Fig. 1 THE SOCIETY FOR MOTION PICTURE ENGINEERS, or SMPTE offers these two standard time codes for video production or editing.
46.1
Hardware Hacker
percent would be theoretically possible. With any care at all, the real world long term efficiencies would not have to be that much worse. Your same square meter panel could approach $80 per year in electricity, a much more attractive value. The only little problem is that, until now, nobody was quite sure what an optical antenna or an optical rectifier was. Enter an individual researcher by the name of Alvin Marks who uses crossed pairs of a special Lumeloid film that uses the antenna/rectifier method for a direct solar conversion. Figures now including an 80 percent efficiency and a penny per kilowatt hour are being bandied about. This does look legit. No obvious physical laws are being broken and the Electric Power Research Institute has seen fit to throw some cash at the idea. And we are certainly getting much better at working with stuff the size of optical wavelengths. All of this is brand new and was rushed to meet this months Hardware Hacker deadline. All the info I have on it so far is the brief note in Business Week, August 12, page 49. For more on direct solar conversion, stay tuned or check into my PSRT bulletin board on GEnie. Other obvious news sources include Science News, Science, and the technology section of the Wall Street Journal. For ongoing technical info, do try EPRI or else the Dialog Information Service. The search keywords could include solar, Marks, Lumeloid, power, EPRI, and energy. Finally, for the ongoing grass roots shirtshleeves solar energy info, be sure to check out that really great Home Power magazine and the new Photovoltaic Network News.
VITC BIT # 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 SYNC BIT "1" SYNC BIT "0" 1 2 4 8 FRAME UNITS
FRAME UNITS
FIRST BINARY GROUP SYNC BIT "1" SYNC BIT "0" 10 20 FRAME TENS DROP FRAME FLAG COLOR FRAME FLAG SECOND BINARY GROUP SYNC BIT "1" SYNC BIT "0" 1 2 4 8 SECOND UNITS
FRAME TENS DROP FRAME FLAG COLOR FRAME FLAG SECOND BINARY GROUP
1 2 4 8
SECOND UNITS
THIRD BINARY GROUP SYNC BIT "1" SYNC BIT "0" 10 20 40 SECOND TENS FIELD MARK "1" FOURTH BINARY GROUP SYNC BIT "1" SYNC BIT "0" 1 2 4 8 MINUTE UNITS
1 2 4 8
MINUTE UNITS
FIFTH BINARY GROUP SYNC BIT "1" SYNC BIT "0" 10 20 40 MINUTE TENS UNASSIGNED ADDRESS BIT SIXTH BINARY GROUP SYNC BIT "1" SYNC BIT "0" 1 2 4 8 HOUR UNITS
1 2 4 8
HOUR UNITS
HOUR TENS UNASSIGNED ADDRESS BIT BINARY GROUP FLAG BIT EIGHTH BINARY GROUP SYNC SYNC SYNC SYNC SYNC SYNC SYNC SYNC SYNC SYNC SYNC SYNC SYNC SYNC SYNC SYNC WORD WORD WORD WORD WORD WORD WORD WORD WORD WORD WORD WORD WORD WORD WORD WORD BIT BIT BIT BIT BIT BIT BIT BIT BIT BIT BIT BIT BIT BIT BIT BIT "0" "0" "1" "1" "1" "1" "1" "1" "1" "1" "1" "1" "1" "1" "0" "1"
SEVENTH BINARY GROUP SYNC BIT "1" SYNC BIT "0" 10 20 HOUR TENS UNASSIGNED ADDRESS BIT UNASSIGNED ADDRESS BIT EIGHTH BINARY GROUP SYNC BIT "1" SYNC BIT "0"
Fig. 2 THE SMPTE lateral time code needs both a seperate audio channel and a full frame to output at its 1200 baud rate. The tape must be in motion for reading.
Fig. 3 THE SMPTE vertical interval time code is sent each field on a blank horizontal line at a baud rate of 180 Kilohertz. The tape could be stationary or move either way.
46.2
November, 1991
locate and identify each and every individual field in the entire video. This normally gets done by using a video time code. Figure 1 summarizes the two most popular time code systems in use today. Both of these are standards published by the Society for Motion Picture and Television Engineers. A useful Time Code Handbook is now available through Cipher Digital. It includes these SMPTE code standards along with the related MIDI electronic music timing code. My copy was free, but they may panic when they see the humongous stack of reader service responses they are certain to get from this column. The human persistence of vision is somewhere around 40 to 50 Hertz, so tricks have to be played to minimize the flicker of many motion displays. For instance, with any 35 millimeter movie, the light is turned off and a frame is suddenly moved into position by using an intermittent action known as a Geneva Stop Mechanism. The light is turned back on by a beam interrupter. The light is turned back off. The light gets turned on again, projecting the exact same image a second time. Finally, the light gets turned off and your next frame is moved into position. We say the sound movie uses a frame rate of 24 frames per second. The field rate is 48 fields per second. In this case, the two fields per frame are identical. With standard television, you have a frame rate around 30 frames per second. Each frame gets broken into two fields, one containing the odd scan lines, and another containing the even scan lines. This is known as interlace and gets used to reduce the display flicker. The field rate is near 60 fields per second. Color tv introduces a further complication in that successive frames must be paired as Frame A and Frame B. This is done to reduce color subcarrier visual artifacts. Thus, only the multiples of two frames or four fields should ever be cut or edited. A glitch results if B does not always follow A, and vice versa. The original or Lateral Time Code standard was first intended for movie film, and later adapted to videotape. This largely obsolete code gets summarized in figure 1 and is detailed in figure 2. The lateral time code needs a separate audio channel. An entire frame is used to transmit 80 bits of digital data at a normal rate of 1200 bits per second. The tape must be moving, either frontwards or backwards before this time code can be read. There is also no provision for error trapping. A Manchester or biphase coding is used. This just means that each bit changes at its beginning. A one bit also changes at its middle; a zero bit does not. A code of this type is inherently self-clocking and can be read in either direction. The first data bit is supposed to start off on the fifth tv horizontal line Those individual bits are largely grouped by fours into Binary Coded Decimal words. These become the units and tens of the frames, seconds, minutes, and hours. There are also some special bits used for sync, user definition, and specific flags. This lateral code is slow enough that it is easily read with machine language code by just about any computer or microcontroller. Their newer, or Vertical Interval Time Code is related but is far more sophisticated. It is summarized in figure 1 and detailed in figure 3. The entire 90 bit time code is transmitted within a single horizontal line. The chosen line should fit somewhere between blanked retrace horizontal lines 10 and 20. The code must be transmitted twice per field on non-adjacent horizontal lines, and often is transmitted four times for extra noise immunity. A similar Manchester or bi-phase encoding gets used. This one has a much higher bit rate, typically around 1.8 Megahertz during a pause. The time code burst begins 10.5 seconds into the selected horizontal line. The VITC can get read while the tape is stationary or moving in either direction at any reasonable speed. Error trapping and other special bits are also included in the code. Unfortunately, the bits do fly by a tad too fast to catch using machine language code on most mainstream microcomputers. So a special fast decoder is needed. Faster yet if you want to search and anticipate during a fast forward or rewind mode. But the logic is simple enough for most PLA or EPLD devices. A dual port RAM or an intelligent peripheral driver could make a useful computer interface. I do not yet know of a custom time code single chip, but it is reasonable to expect a cheap one shortly. And several simpler but nonstandard field codes are being introduced in new video products, especially by NEC. There have also been lots of helpline requests for a simple and automatic time code inserter for home video editing use. Let me know what youd like to see on this.
BADGE, PASS, TOLL CARD, ANIMAL COLLAR, PATIENT RECORD, WORK IN PROGRESS, PRICE TAG, ETC ETC ETC
Fig. 4 ELECTRONIC DOG TAGS using the cheap and durable Maxim DS1990 series of touch memory chips can get stuck on virtually anything. They do offer electronic readable serial numbers and several thousand of bits of read-write non-volatile memory. These are the shape of a coin cell and last for ten years.
46.3
Hardware Hacker
Unfortunately, many of these are staggeringly expensive and hard to get. But, by playing in the cracks, you may be able to come up with a useful workaround. JerryCo, whove recently renamed themselves as American Science and Surplus sometimes have great cases at unbeatable prices. Their #10827 and #21085 both cost around $4. The most obvious source for the "suitcase" style packaging is Jensen Tools, who do stock everything from leather through plastic to aluminum to stainless steel. And the ultimate in military quality primo cases come from Zero Manufacturing. At the ultimate in military primo prices. But the fancy electronic case I am the most impressed with is not an electronic case at all. This is the Guard Box that is offered by Pelican Manufacturing. A reinforced structural resin case measuring roughly 3 x 6 x 8 , it comes in four bright colors, and costs around $7 in your smaller production quantities. This box is airtight and watertight to thirty feet of depth. A handle, a hinged O-ring sealed cover, and positive closure snaps are included. It is ideal for, perhaps, a solar panel controller. Larger sizes and custom colors are also available. Several other manufacturers of quality (e.g. expensive) electronic cases are also included in our listing. Please do let me know if I missed anything obvious or if you have any favorites of your own. durable stick-on data acquisition and storage system the size and shape of a coin cell battery. You glomp this onto anything, such as a badge, a ticket, an animal collar, a security access card, a medical patient, or a product being manufactured or shipped. Inside is a 10 year lithium cell and some unique electronics. Depending on the version, the fancy electronics gives you either a plain old 48 bit serial number or else thousands of bits of read-write memory. Obvious advantages over a printed bar code label are that (A) much more info can be stored, (B) you can update it at any time, (C) no expensive laser reader systems are needed, and (D) the stainless case is far more durable and rugged than a paper label. What is really unique is that there is only a single contact plus the case ground. Which is all you need for reading and writing. Figure 5 shows you how one lead can be used to both read and write. Your host computer or whatever will first input your serial activation data string. Your touch memory will then respond, returning a serial number and/or your data. The single contact is exceptionally easy to access. Theres none of the alignment or fragility problems that you might have with more traditional connectors. Several different styles are newly available. The cheapest outputs only a 48 bit serial number. Fancier versions can read or write to 1K or 4K bits of internal non-volatile ROM. On yet others, a security code prevents just anyone from reading or altering the contents. This gets important on such things as a bus fare pass. The reading and rewriting process is destructive, so very bad things can happen with an erratic contact or a partial write. To beat this, your info is first written to an internal scratchpad area. Only the complete and correct write later gets transferred to main memory. An internal checksum is also provided to let you know if your data is still valid. Free samples and more technical details are available directly from Dallas Semiconductor. Cost is in the $1 to $3 range, depending upon the chip features and complexity. For our contest this month, just tell me something off-the-wall you would
In standby mode, the touch memory monitors its receive circuitry and disables its send electronics. The host computer first enables its tri-state driver and sends out
a serial interogation and data code. The host computer then disables its tri-state driver and awaits a reply.
Your touch memory receives the interrogation code and activates its send electronics. The serial number or the requested RAM data is then transmitted. The touch memory then goes into its standby mode to await further commands. The receiver grabs and interprets the returned data. To prevent possible contact or destructive write problems, the
host computer normally does a read, followed by a write, followed by a verify.
+5V
enable
10K
send
receive
send
HOST COMPUTER
TOUCH MEMORY
Fig. 5 ONLY A SINGLE CONTACT IS NEEDED to read from or write to your touch memory. Here is how it gets done.
46.4
November, 1991
do with an electronic dog tag. There should be all of the usual Incredible Secret Money Machine book prizes, with a great all expense paid (FOB Thatcher, AZ) tinaja quest for two going to the very best of all. As usual, send your written entries to me here at Synergetics, and not to Radio-Electronics editorial.
FANCY CASE & ENCLOSURE RESOURCES American Sci & Surp 601 Linden Place Evanston, IL 60202 (708) 475-8440 Bud Industries, Inc 4605 East 355th Street Willoughby, OH 44094 (216) 946-3200 Carlon 25701 Science Park Drive Cleveland, OH 44122 (800) 321-1970 EAI 37 Sherwood Terrace S124 Lake Bluff, IL 60044 (708) 295-6664 Hoffman Engineering 900 Ehlen Drive Anoka, MN 55303 (612) 422-2177 Jensen Tools 7815 South 46th Street Phoenix, AZ 85044 (602) 968-6231 Lansing Instrument PO Box 730 Ithaca, NY 14851 (800) 847-3535 Pelican Products 2255 Jefferson Street Torrance, CA 90501 (213) 328-9910 Rose Enclosures 7330 Executive Way Frederick, MD 21701 (301) 696-9800 Zero Corporation 777 Front Street Burbank, CA 91503 (818) 846-4191
46.5
Hardware Hacker
46.6
Don Lancasters
Hardware Hacker
December, 1991
et us start things off with a bunch of updates to some of our previous topics. Along with some new answers and random comments to a few of our more asked helpline calls FM Stereo Broadcasters The Rohm BA1404 FM stereo transmitter chip really has been popular, and a dozen sources now exist for kits via ads in Radio-Electronics , in Nuts & Volts, and elsewhere. This stunning micropower and low cost chip was originally intended to ease adding a CD player to your car radio. You just unplug your regular antenna, insert a module containing the BA1404 into the radio, and your old antenna into the new module. Presto. Instant and hassle-free high quality stereo access to your car audio system. The fact that the chip made a really great wireless broadcaster was an afterthought. But there are problems with some of the kits. Some do not use a crystal for the 38 kHz stereo pilot subcarrier, that can lead to drift and adjustment hassles. A suitable 38.4 kHz crystal is cheaply available from Statek. A second problem is that many premium FM receivers are synthesizer controlled these days, and demand a frequency that is exactly on channel. Older and cheaper FM receivers had an AFC analog frequency lock that could grab any old frequency and follow it as it drifted around. Can you directly crystal control a broadcast quality FM transmitter? Not really. Because you are asking for a something that is both ultra "stable" and instantly able to accurately "drift" over a wide frequency range as it is being modulated. Instead, you normally would use an indirect method to stabilize your FM transmitter. You should add a small varactor diode (say 10 pf) across your tank so that you can tune your tank circuit by using a DC control voltage. Details on this in Rohms ap-notes. Then, you divide down your output frequency to average out your modulation and compare this against some
Fractals & chaos update Standards resource info More on FM stereo chips Home energy monitoring Current transformer ideas
animation stands and computer controlled wood routers. One videotape I received was quite impressive. This is an interesting example of synchronicity. If the time is right for something, it should emerge from dozens or hundreds of apparently independent sources. By the same token, attempting to patent your "new idea" is almost certainly futile, again because of synchronicity. Word has it that rewinding your alternator gives you by far the best results. Try something around twenty
reference. A DC control voltage is then derived to hold you onto your frequency. Phase lock loop style. Im checking out several suitable circuits, so stay tuned. Or else send me your favorite solution to this ongoing circuit hassle. Whats really needed here is a low cost, low power single chip synthesizer. Alternators as Power Steppers Yes, they definitely do work, and an amazing number of hackers have already picked up on this on their own. The important uses now include
0.01 (A) The standard way of measuring a DC current is to place a very small valued shunt resistor in series with your power load. The voltage drop across the shunt and Ohms Law will tell you the current. In this example, an 0.01 Ohm shunt gives you a tenth of a volt out for a measured current of ten amperes. R (internal meter resistance)
R / 999 (B) If an ordinary meter is used, its internal resistance must be included in the calculation of the shunt value. As shown here, to convert an 0-1 DC milliameter with an internal resistance of R into an 0-1 DC ammeter, you have to add an external shunt of R/ 999 Ohms. Note that the term "shunt" applies to the meter, and not to the load being measured.
SHUNT
(C) When ultra-low shunt values or high currents are involved, a four terminal Kelvin mesurement scheme should be used. This prevents unavoidable drops across the power connections from trashing your readings. Four terminal techniques are also used for remote load regulators and in superconductor research.
47.1
Hardware Hacker
technical details or else a fully open substitute to evolve shortly. Consult the usual suspects. To me, it makes infinitely more sense to derive a simple code based on channel, time, date, and duration. One that anyone can apply and use. Most computer manufacturers have learned long ago that open is the only way to go. Witness Adobe Systems fortune skyrocketing when they released the black book showing how their super secret type I PostScript fonts were encrypted. In reality, all their highly touted encryptions were an outright joke. Why? Because any computerliterate seventh grader could trivally crack their code with a little patience and zero insider knowledge. But thats another story. See my Ask the Guru II reprints for details. And I do stock some black books here. Fractals and Chaos One of the shortcomings of James Gleicks popular Chaos: The Making of a New Science was that it was a tad shallow in spots. It would lead you up to lots of great stuff, and then leave you with nowhere to go. This has now been cured in spades through Chaos: The Software from Autodesk. This $59.95 software package uses a color VGA on any IBM PC or clone. It includes stunning and spectacular interactive demos of Julia sets, Mandlebrot sets, chaotic pendulums, strange attractors, chaos games, fractal forgeries, toy universes, and bunches more. One major emerging use for the fractals is in video compression. The Iterated Systems people have just introduced a unique series of new products. These include their POEM Fractal Factory and POEM Fractal Slide Projector. Amazingly, these use software only to decompress a full animation in real time. And at compression rates approaching 100:1 . Sadly, the compression process is very slow and requires expensive and custom hardware and software. Their new compression scheme is ideal for such uses as CD ROM. Much more info on all of these topics does appear in my Hardware Hacker II & III reprints.
OVERLOAD OUTPUT
Fig. 2 THIS SIMPLE CURRENT LIMITER is adjustable over a 50 ma to 1 ampere range. Use the collector current to pull down a reference or otherwise reduce your input current or source voltage.
feet of #26 wire per pole and straddle only one single pole (instead of the usual three) with each winding. The driver circuits that I showed you purposely held back on the power to keep you from burning things up on your first cut. Naturally, youll go to higher power and properly current limited two-level drivers as you improve your designs. Video Time Codes I forgot to tell you that the SMPTE vertical interval time code is usually sent on lines 13 and 15 of your field retrace. In retrospect, this code is horribly klutzy and incredibly inefficient. Owing mostly to its having to evolve from an earlier movie film standard and from primitive extraction electronics. A bunch of newer time codes are
now emerging. All incompatible, of course. Both NEC and Sony are now pushing their own time standards on new consumer video products. More on these when I get a chance. Our second most popular helpline topic (second only to caller id) is on the new VCR time code. The one that lets you look up a magic number in a directory or a video guide and easily program your VCR for an unattended recording. It turns out that the two developers of this recording code are Daniel Quo and Henry Yen. More info on the use and distribution of the code is available through Gem Star Development Corp. At present, their code is "secret". Naturally, that wont last for long, and you can expect either complete
AC LOAD
AC SOURCE
Fig. 3 A CURRENT TRANSFORMER is just a transformer that has a very low impedance primary (often ONE turn or less) and a high step up ratio. The voltage across the secondary is proportional to the primary current.
47.2
December, 1991
Especially when AC is involved or when breaking the primary circuit is a no-no. Let us do a quick review of several of the current sensing fundamentals and see just what hacking opportunities it leads us to. In general, there are two main methods of sensing current. With the direct method, you will insert an appropriately low resistance or an impedance in series with your load and then measure the voltage drop across that impedance. With the indirect method, you try to accurately measure the strength of the magnetic field created by a conductor carrying your load current. Figure one shows us three stock direct current ammeter schemes. In 1-A , a low value resistor known as a shunt is placed in series with your load. You then measure the voltage drop across it with a "perfect" high impedance voltmeter. For instance, an 0.01 ohm shunt will give you a tenth of a volt out for ten amps in. Shunts can be anything from a piece of paper clip (thanks to Steve Ciarcia) to precision and temperature compensated devices. Sources of shunts include Simpson, Triplett, and various advertisers found in those Measurement & Control and Sensors trade journals. It gets slightly trickier when you use a shunt to increase the range of an ordinary milliammeter, rather than a "perfect" high impedance voltmeter. Details per figure 1-B. Say you have an 0-1 DC milliammeter of internal resistance R. Your shunt to convert it into an 0-1 DC ammeter should have a value of exactly R/999 . Such that the combined parallel resistance of your shunt and the meter resistance ends up precisely R/1000. The amount of voltage drop across your shunt will depend on the input current, the shunting resistance and Ohms law. The lower the R value of your meter, the less voltage drop you will get across your shunt when you measure high currents. Since all high current shunts are extremely low impedances, you have to be very careful that voltage drops across your meter and all your load connections do not inadvertently get included in your measurement. Which leads us to the four terminal or Kelvin connection shown in figure
1-C . You only measure the voltage drop across one precisely known portion of your shunt. This guarantees any external high current connections are outside of the voltage drop being accurately measured. A variation on this four terminal technique can let you use a voltage regulator remote from its load. By feeding back your voltage sensing leads, you could include your main current drops inside your regulator and get more precision results. The trick, of course, is to make absolutely certain that none of the load current travels through the sensing leads. A simple current limiter is shown in figure two. This one is adjustable over a 50 milliampere to a 1 ampere range. It is based on a silicon NPN transistor requiring a 0.6 volt drop before it conducts. Whenever your transistor starts conducting, its collector current gets used to pull down a regulator or whatever, thus safely limiting your current. As figure three shows us, current sensors for AC circuits often use a current transformer. This is a special transformer with a very low impedance primary and having a high step-up ratio. The ac voltage that is delivered to the load resistor is proportional to the current through the primary winding. There are some very important
rules for current transformer use. The load must never be disconnected from a current transformer! Dangerous and possibly lethal voltages could result. Also obviously, you must not ever disconnect the primary in any live circuit. Only the specified load resistor can be used or calibration will suffer. Often, the primary can simply be a single wire through the core of your transformer. If you do use one single wire, some care is needed to get an accurate and repeatable reading. The wire should go straight through the exact center of your current transformer. It should extend straight and normal for a few inches in either direction. It must not be allowed to move or vibrate. On the other hand, sensitivity and accuracy can be improved by using several primary turns. But there may not be room in the core. If the current transformer core can be opened in some manner, you can connect or disconnect your current transformer without physically disconnecting the current-carrying load wire. This gets very important when making any service and test measurements. Or when you want to avoid an electrician. A snap around ammeter is one example of an openable core. When you glomp around your wire to be
Fig. 4 A TOROIDAL CURRENT TRANSFORMER is economical and accurate. It is also self-shielding. But you physically have to disconnect and thread the primary wire through the core center.
47.3
Hardware Hacker
sensed, you are closing the core and completing a current transformer magnetic circuit. Two important points. You must snap around a single wire. If you snap around a wire and its return path at the same time, the two currents will largely cancel out, leading to a useless measurement. And the magnetic path must be identical for each snap. Thus, your core must close together exactly and precisely. The same way each and every time. Ive found three good sources for low priced and eminetly hackable current transformers. These include Amecon, CR Magnetics and Toroid Corporation. The latter two offer toroidal current transformers, similar to figure four. These are priced in the ten to twenty dollar range. Some models include built-in LED safety monitors. But by far the most hackable current transformer I have ever seen is the Amecon LD-50 load detector. This beauty costs under three dollars in quantity. Samples are available. Now for the unusual part. Their LD-50 comes apart so you can install it snap-around style, without breaking your primary load wire. Wires of #8 size or smaller are acceptable. Using a 50 ohm load, the LD-50 produces 50 millivolts per ampere and is useful over a 5 to 50 ampere range. Which is ideal for most home power monitoring uses. Speaking of which separate control receiver/transmitter is located a safe few feet away. This control box uses a lower frequency radio waveform to interrogate your isopod. If properly addressed, the isopod responds by returning a high frequency code burst equal to the current presently being measured. Finally, the control box intercepts the data burst and converts the sensed current into standard serial data that a computer can understand. As shown, the isopod would only measure and monitor current. This should be good enough for all typical home energy monitoring uses. But, by carefully synchronizing the start of your data burst to the a positive-going current zero crossing, your control box could measure its own voltage to determine the lag or lead and your actual real and reactive power. Several channels could be made available by using different isopod addresses. This solves the dilemma of simultaneously monitoring both legs of the normal 220 volt, center ground home power system. Or of monitoring individual breaker lines. Probably the simplest addressing scheme would be simple time delays in a time division multiplex setup. For instance a channel "3" isopod would wait 0.3 seconds (plus the next zero crossing) before any response. A channel "6" isopod would wait 0.6 seconds, and so on. This way, one single control box and one single interrogation command could read dozens of isopods. You can probably get by snapping the pod directly onto your incoming power leads ahead of the meter. Up on the roof where they are easily open and accessible. The tiny extra power "stolen" from your utility would add half a cent to a $200 monthly power bill. And the potential utility benefits in load shedding and load levelling would be much higher. I can think of lots of other uses for a cheap and fully isolated, snap-on current sensor. Safely measuring high voltage industrial systems can be one obvious use. Getting info onto or off a rotating shaft or moving machine is a second. Theres lots more. And last summer I had a slightly but infuriatingly intermittent frost sensor in my heat pump. It sometimes insisted upon switching over to high
Fig. 5 THIS "ISOPOD" CONCEPT can greatly simplify home energy management. The Isopod glomps on one of your incoming power cables. Inside is a current transformer and some micropower electronics. On any addressable VLF radio inquiry from a nearby control module, the isopod returns your present current consumption as a UHF radio signal. Response is synchronized to current zero crossings so that both real and reactive power can be remotely measured.
47.4
December, 1991
emergency heat during the middle of an Arizona desert summer afternoon. Those several wasted service calls and the perfectly good parts that were unnecessarily replaced could have been eliminated by several snap-on isopods that monitored and recorded what was going down. But why dont you tell me instead? For this months contest, just tell me what you would do with one or more isopods. Or else contribute somehow to our home energy dialog. There will be all of those usual newly revised Incredible Secret Money Machine II books going to the dozen best entries, with a great all expense paid (FOB Thatcher, AZ) tinaja quest for two going to the best of all.
STANDARDS ASSOCIATIONS & RESOURCES Am. National Stds. Ins. 11 West 42nd Street F13 New York City, NY 10036 (212) 642-4900 Am. Soc. Test & Mat 1916 Race Street Philadelphia, PA 19103 (215) 299-5400 Am. Radio Relay League 225 Main Street Newington, CT 06111 (203) 666-1541 As. Computing Mach. 11 West 42nd Street F3 New York City, NY 10036 (212) 869-7440 Compliance Engineering 629 Massachusetts Ave. Boxborough, MA 01719 (508) 264-4208 Intl. Standards Org. 1 Rue Varembe C-56 CH1211 GENEVE 20 Sw. 41-22-749-0111 MIL Spec Order Desk 700 Robbins Av B4-D Philadelphia, PA 19111 (215) 697-2179 Nat. As. Broadcasters 1771 N Street NW Washington, DC 20036 (202) 429-5300 Nat. Elec. Manuf. As. 2101 L Street NW Washington, DC 20037 (202) 457-8400 Nat. Fire Prot. As. Box 9101 Quincy, MA 02269 (617) 770-3000 Soc. Automotive Engs 400 Commonwealth Drive Warrendale, PA 15096 (412) 776-4841 Soc. Motpic & TV Eng 595 West Hartsdale Ave White Plains, NY 10607 (914) 761-1100 Standards Eng. Soc. P.O. Box 2307 Dayton, OH 45401 (513) 223-2410 Underwriters Labs 333 Pfingsten Road Northbrook, IL 60062 (708) 272-8800 US Govt Bookstore 720 North Main Street Pueblo, CO 81003 (719) 544-3142 -------------------------------Fed. Information Cen. (800) 359-3997 --------------------------------
places to start include the Encyclopedia of Associations on the reference shelf of your local library. Or, for government standards, try the main Federal Information Center phone number of (800) 359-3997. To make things easier for you, Ive gathered up most of those standards associations and resources of hacker interest together for you into our resource sidebar for this month. Two quick notes. ANSI is just the American arm of the international ISO. They are pretty much one in the same. And the US National Bureau of Standards has long ago changed their name over to the National Institute of Science and Technology, or NIST. Actually, RS232 has been largely replaced by the MacDonald interface. Jerome R. MacDonald is the senior member of a design team in the dairy science division of the US Department of Agriculture who has come up with a fast, convenient, error-free, and fully networkable serial comm setup. Its use rapidly spread to other government agencies, to hackers, and finally to industry. The great MacDonald interface is presently up for Electronic Industries Evaluationary (EIE) status. Thus, the old MacDonald farm interface is now an EIE I/O.
Electronic Ind. As. NIST (US Bureau Stds) 2001 Pennsylvania Ave NW ADMIN E-128 Washington, DC 20006 Gaithersburg, MD 20899 (202) 457-4900 (301) 975-3058 IEEE Electronic Eng. 345 East 47th Street New York City, NY 10017 (212) 705-7900 Nat. Standards Assoc. 1200 Quince Orchard Blvd. Gaithersburg, MD 20878 (800) 638-8094
video camera, digital audio, and new broadcast satellite chips. And from Silcon Systems, a pair of newer data books on Communications Products and Storage Products chips. Free user samples of the TPIC2801 intelligent power drivers and a nice Application Report are available from Texas Instruments. This one is particularly useful for automotive and solenoid driver applications. And through Aremco, a Materials Catalog M12 which includes a wide variety of high temperature ceramic materials, including machinable and pourable versions. Trade Winners now has a US phone and address. This Hong Kong answer to Computer Shopper gives you direct wholesale access to far eastern computer and electronic sources. The monthly EPRI Journal is full of power ideas, ranging from solar to energy conversion to cold fusion, to cogeneration. It is free to utilities or anyone associated with the media, but $29 per year otherwise. EPRI also has lots of videos and reports. The Alternate Energy Sourcebook is a combination catalog and resource
file offered at a refundable $14 from Real Goods. On their catalog side, you will find such goodies as solar panels, pumps, DC lights, and such. On the resource side are tech articles by Amory Lovins and directories to electric vehicle clubs. How to Automate your Home is a new self-published manual by David Gaddis and offered through Home Automation USA. Well done. I have newly revised and reprinted the second edition to my Incredible Secret Money Machine, and now have autographed copies in stock for you. And for those of you interested in Display PostScript, Ive now posted a freeware GhostScript clone to my GEnie PSRT board. Included are special Ghostscript fonts and the full source code. Youll also find Hardware Hacker and Blatant Opportunist preprints here, along with bunches of goodies not available elsewhere. Our usual reminder here that most of the products and services mentioned appear either in the Names and Numbers appendiz or the Association Resources sidebar. Lets hear from you.3
47.5
Hardware Hacker
47.6
Don Lancasters
Hardware Hacker
January, 1992
t seems I did miss an obvious source in our recent Tesla and High Energy Resources sidebar. This is the Tesla Society up in Colorado, who do run a great museum and now offer annual Tesla and nontraditional energy conferences. Their High Energy Enterprises division has lots of books and videotapes. These range the gamut from hardto-find and genuinely useful research materials on down (wayy on down) through bunches of "Boy-a-wholeflock-of-them-flew-over-that-time!" pseudoscience titles. On the other hand, if you are into antigravity, the 1000 miles per gallon carburetors, Russian weather control conspiracies, Reed motors, all those pangalactic happy faces on Mars, or zero point scalar energy, these books and videos are definitely for you. This is most fascinating reading and watching, either way. Actually, Ill freely admit that I firmly now believe in one trilateral conspiracy. Thats the one involving the first, second, and third laws of thermodynamics. But thats just me. And if you want to prove me wrong, just show to me any simple and an independently verifiable experiment that anyone can reliably duplicate. Some more info on wavelets: The Wavelets book offered by Jones and Bartlett is at long last in print. And a major new tutorial just came out in the October 1991 IEEE-SP Signal Processing magazine on pages 14-38.
Effective shielding ideas Computer monitor basics VGA Adaptor for Mac LC Recurrent flyback sweeps GPS Navigation Resources
going to suppress. Any old conductor should usually block an E field. One obvious hacker choice that works well are boxes built up out of double sided printed circuit board. See figure one. But there are gotchas. At lower frequencies, there is the lack of a well developed skin effect, which causes an E field to penetrate deeper into a conductor. Thus, thicker shields are needed to supress lower frequencies. Say 20 mils for an AM broadcast band use or 80 mils for the high ultrasonic frequencies. A shield can also act as a shorted turn to any nearby inductor, possibly lowering the Q and detuning. A slot or other continuity break can sometimes help this problem. While some small shield holes are often tolerable, any poor contacts are definitely not. A continuous solder bead or else lots and lots of individual screws may be needed for really high shielding effectiveness. Its very important that unintended currents dont ever run through any shields. Unwanted signal drops can often end up in series with your input signals, making things much worse rather than better. Which is also why you should keep all your digital and analog grounds separate. The H fields can be blocked by use
youre going to have interfering noise or signals present, one very powerful technique is to change over to fully balanced circuits. These sense only the differences between their inputs, rather than any absolute values with respect to ground. Thus, any common mode interfering signals that bounce both inputs up and down together will automatically be rejected. Or at least suppressed. Several fancy audio studios have even gone to totally balanced power lines to dramatically reduce their hum and noise problems. There are usually two paths that interference can travel. One is with radiated energy, which can propigate directly through space to emit from or interact with your circuit. The other is conducted energy, which enters along your power connectors and input or output cables. Since separate tricks are needed to deal with any radiated or conducted energy, step one is finding out which path is the one to deal with first. The radiated energy interference will usually have two components. These are that "E" or electric field, which largely concerns itself with the induced voltages; plus that "H" or magnetic field, which creates induced currents. Once again, you have to know which field component you are
Fig. 1 SIMPLE E-FIELD SHIELDS for radio frequency hacker projects are easily built up from pieces of double-sided printed circuit board stock.
48.1
Hardware Hacker
of any strongly magnetic materials. These materials are said to posess a very high permeability. The simplest magnetic shield is any piece of scrap sheet steel from the air conditioning shop or trailer hitch works. But things get messy in a hurry if you need lots of magnetic attenuation. Those special and fancy shielding materials are optimum only in certain thicknesses and over a specific frequency range. Worse yet, some of these cannot be cut or drilled without a complex reannealing process. And too strong of a magnetic field could saturate a shield, making it useless. To be effective, a magnetic shield should completely surround the volume it is protecting. A good free booklet on all of the fundamentals of magnetic shielding is available from Amuneal, while other sources of custom shield materials now include Advance Magnetics and Magnetic Shield Corp. Conducted interference can be best eliminated by some blocking filter. These filters should freely pass the signals of interest, but present a high series impedance and a low shunt impedance to all others. Murata-Erie is one good source. The ferrite bead is a remarkably cheap and super effective interference suppressor. As figure two shows us, you simply hang these onto a wire or else run a turn or two through them. Ferrite beads act as a broadband high frequency lossy transformer. There are various materials and sizes. Sources of ferrite beads include Fair-Rite, FerriShield, Ferroxcube, Intermark, and Siemens. One very little known ferrite bead gotcha: Do not ever cast a ferrite bead in epoxy or otherwise constrain it so
Fig. 2 THOSE FERRITE BEAD interference suppressors are slipped onto any conductor and behave as lossy and broadband transformers.
it cannot move. The beads must be able to physically change their size slightly, or else the performance will sharply degrade. Two trade journals that involve themselves with shields and shielding are Electronics Test and Compliance Engineering.
+V
DAMPER DIODE
H YOKE CURRENT
FLYBACK VOLTAGE
HORIZONTAL SWITCH
off
on
off
on
DAMPER DIODE
on
off
on
off
+V SUPPLY FUNCTION
sink
source
sink
source
Fig. 3 THE RECURRENT FLYBACK SWEEP on television sets and computer monitor displays is extremely energy efficient, but will only work over a VERY limited range of horizontal scan rates.
We sure do get a lot of computer monitor and tv compatibility calls. Lets start off with the obvious: The performance and bandwidth of an ordinary tv or VCR using composite NTSC video is severely limited. That is why all the computer folks went to special RGB monitors in the first place. And that is why nearly all the computers are totally incompatible with ordinary television gear. Yes, there now are all sorts of ways you can use your computer for real video editing or to record computer screens on a VCR. But note that there is no way I know of that you can record plain old 80 column text as composite video on your VCR . Or display it on any unmodified tv set. Especially in color. If you are going to put something out as broadcast video or record it on a VCR, it has to look like all the other stuff that you see on your cable channels. Which means a few big characters, rather than zillions of small ones. And far fewer lines of horizontal resolution. How does a monitor work? Inside is a cathode ray "picture tube" with one or more guns that squirts lots of electrons at a phosphor screen. At any instant, only one single dot appears on the screen. That dot gets moved around by the scanning process, and will get brightened and dimmed by
48.2
January, 1992
rapidly setting its intensity at a video modulation rate. To build up the illusion of a full picture, that scanned dot is moved rapidly and horizontally from left to right and more slowly vertically from the top to bottom. The decay characteristics of the phosphors selected and your human persistence of vision combine to create the illusion of a total picture. Television uses what is known as an interlaced scan. To build up a tv frame, the dot starts at the upper left and rapidly scans to the right and slowly on downward, painting every second scan line. When it gets to the bottom of this field, it goes back to the top and picks up what it missed, painting a second field. The normal field rate is usually 60 Hertz for black and white or 59.94 Hertz for color. The NTSC (or the Never The Same Color) standard uses 512 lines per frame, or 262.5 lines per field. Which leads to standard horizontal scan rates of 15750 Hertz for black and white or 15735 Hertz for color use. These two numbers end up slightly different to get all the rest of those color magic numbers to properly drop in place. Interlaced scan is used by NTSC to minimize flicker. But interlaced scan is totally unsuitable for data displays, especially small text! The reason is that interlace only works if successive lines are more or less the same. This is clearly not the case between dot lines of finer text. Thus, most computer screens demand a noninterlaced display where each field is complete and identical. Most computer monitors also will use a 59.94 or a 60 Hertz vertical scan rate. Get under this and youll end up with too much flicker. If you try to lengthen the phosphor persistence you may end up with "comets" for traveling balls. The horizontal scan rate is decided by how many horizontal lines you are using per field. As you increase the number of lines in a display, the horizontal scan rate goes up. Scan rates from 22 to 45 kilohertz are typical, with some exotic displays going much higher. Now for the kicker: Most monitors will only work over a VERY limited range of horizontal scan frequencies! Unless you go to some very fancy multi-sync techniques. Before you can use any monitor in any system, you must make sure the horizontal scan frequencies are compatible with your intended use! that e = L i/t
i/t = e/ L
Lets rearrange things a tad Now i is the change of current and t is the change of time, so i/t will be a linear current ramp whose rate of change should equal your supply voltage divided by the inductance of your horizontal yoke. Say that the switch in figure three is now open and has been that way for a long time. There is no current in the yoke, and no deflection. The spot will still be in the middle of the line. Now, close the switch, but just for around one-half of the live scan time. What happens? We now apply a positive voltage to an inductor, and start building a linear current ramp. That current ramp in the deflection yoke creates one linearly increasing magnetic field, and the spot moves to the right. When you get to the right side of the screen, youll have bunches of energy stored in the horizontal yokes magnetic field. What are you going to do with it? If you just burn it up as heat, youll end up with bad reliability and high
WIDTH ADJUST
DAMPER DIODE
FLYBACK CAPACITOR(S)
VIDEO
H RATE DETECTOR
Fig. 4 MULTI-SYNCING MONITORS change their sweep values to try and match an input horizontal scanning rate. But once selected, they still operate only over a very limited frequency range.
48.3
Hardware Hacker
power consumption. Besides needing far tougher electronics. Instead, lets work smarter instead of harder. Open your switch. Your equivalent circuit now consists of a yoke coil in series with the flyback capacitor. A standard and high Q series resonant circuit. But one that started out with zero volts on your capacitor and a strong current through the coil. At this time, all of the energy is in the coil and none is in the capacitor. Let the circuit resonate for a while, exactly like any other series resonant circuit. In one quarter of a cycle, the capacitor will be charged up to a large positive voltage, often in the 800-1200 volt range. At the peak, your current will be zero, and our inductor will be "empty" and field free. You have now transferred all of the magnetic energy that was in the yoke into electrical energy stored in the flyback capacitor! And done so with very little loss. Let the circuit continue to resonate. The current will reverse, and energy will start transferring back into the inductor. Note that the current is now going in the opposite direction. Let the circuit resonance continue until the capacitor gets to zero volts. You will now have placed all of the original magnetic energy right back into the deflection yoke, with one very important difference: Since the current is going in the opposite direction, youre now at the maximum left spot position, compared to the maximum right that you were half a resonance cycle ago. And, again, you have done this with very little loss of energy. This process is known as a flyback, and the retrace interval is called the flyback time. The beam is turned off during the flyback time so the retrace is invisible. Let the circuit resonance try and continue. The flyback capacitor will try and go negative, but at that time, a damper diode will turn itself on. Your circuit now consists once again of a coil connected between the positive voltage and ground. It will once again start generating a positive current ramp. Only this time, it starts from a negative initial current. Thus, your sweep starts linearly moving from the left to the center. As you move on towards the center of your scan line, the yoke energy gets sent back into the positive supply. When you get to the center of the screen, you once again close your switch to repeat everything for the next cycle. The damper diode shuts down just as soon as you cross center screen. Automatically. Once again: Close your switch to move from center to right. Transfer the coil energy to a flyback capacitor. Resonate half a cycle and transfer the flyback capacitor energy back to the coil inverting its sign and putting you far left. Turn on the damping diode to move from left to center. And repeat the process once each scan line. Several details Ive omitted: That flyback pulse also gets sensed and routed to a current stepup transformer called a flyback transformer. The flyback transformer steals a minor part of the energy and uses it to create the high voltage dc supply that is needed by the display tube. Other windings can be used for blanking, horizontal phase comparison, and boosted supply voltages elsewhere in the tv or monitor. And games have to be placed to keep any uncentering dc bias out of the horizontal yoke. But regardless of these details, the basic concept of recycling your deflection energy remains. At any preselected horizontal scan frequency, the flyback system works and works well. But if you try and change the scan rate by more than a few percent, the voltages, currents, and resonances quickly get out of hand. Thus, this very elegant, highly tested, and ultra conservative flyback scheme inherently works best only at one horizontal scan frequency. And that is why you cant normally get a monitor that can accept any old horizontal rate. Yes, there are multi-sync monitors. These usually work by measuring the intended input scan frequency and then switching in one or more flyback capacitors and adjusting the supply voltage accordingly. Our figure four shows one multisync scheme. Once switched, a multisync monitor is a narrowband system just like any other flyback-driven circuit. The bottom line: A monitor only works at its intended horizontal scan rate. And at no other.
RED RED GROUND BLUE BLUE GROUND GREEN GREEN GROUND HOR SYNC HSYNC GROUND VERT SYNC VSYNC GROUND VGA ENABLE GROUND
2 1 9 13 5 6 15 14 12 11 10 7 3 NC NC NC NC NC
1 6 3 8 2 7 13 10 14 4 5 9
RED RED GROUND BLUE BLUE GROUND GREEN GREEN GROUND HOR SYNC HSYNC GROUND VERT SYNC VSYNC GROUND
NC 11 NC 12 NC 15
Mac LC
Video DB15 male
4 8
VGA Monitor
Video VGA male
Fig. 5 THE MACINTOSH LC definitely can NOT be used with a NTSC montor. But this simple cable and jumpering lets you use it with most VGA color monitors. The pin7 to pin 10 jumper selects the VGA mode.
48.4
January, 1992
The Macintosh LC computer was designed for use with the Mac color monitors to the Mac color standards. Since these can be expensive, lots of hackers are often on the lookout for lower cost substitutes. But note that you definitely can not use a NTSC color monitor with these because of the higher scan rates on the LC. And also do note that you should never buy a substitute monitor without making certain it works and is good enough for your uses. Happily, there is a hidden "secret" provision on the Mac LC that lets you change your LC output so it is VGA compatible. Quality VGA monitors are often available much cheaper in far wider selections than are "real" Mac monitors. The secret jumper that changes the LC scan rates for VGA compatibility is shown you in figure five.
GPS NAVIGATION RESOURCES Ashtech 390 Potrero Avenue Sunnyvale, CA 94086 (800) 229-2400 Bancomm 6541 Via del Oro San Jose, CA 95119 (408) 578-4161 GPS World P.O. Box 10460 Eugene, OR 97440 (503) 343-1200 Magellan Systems Corp. 960 Overland Court San Dimas, CA 91773 (818) 358-2363 Magnavox 2829 Maricopa Street Torrance, CA 90503 (800) 421-5864 NASA Tech Briefs 41 East 42nd Street Ste. 921 New York City, NY 10017 (212) 490-3999 Rockwell Commercial GPS P.O. Box 568842 Dallas, TX 75356 (214) 996-5863 Speleonics P.O. Box 5283 Bloomington, IN 47407 (812) 339-7305 TI/GPS Products PO Box 869305, M/S 8449 Plano, TX 75086 (214) 575-4057 Trimble Navigation 585 North Mary Avenue Sunnyvale, CA 94086 (800) TRI-MBLE
Amazingly, the antennas needed are no big deal. They can easily fit inside a small shoebox. Unlike video satellites, the GPS antennas need a wide beamwidth, a narrow frequency bandwidth, plus the pseudorandom modulation which works acceptably well with lower signal-to-noise ratios. The antennas do require line-of-sight operation, so they have to be used outdoors, on a rooftop, or on the top of a vehicle or boat. Since the reception electronics is incredibly complex, it is unlikely you would want to try and build up your own GPS receivers by using discrete components. Thankfully, sophisticated chip sets are now becoming readily available from Rockwell, Magellan, and others. Their initial pricing is in the $500 range. But it is reasonable to expect Japanese GPS chips within a year or two for under $35. So now is the time to start on your GPS hacks. Even today, the pricing is certainly within the range of, say, trucking firms or land surveyors. The leading trade journal in this new field is called, of all things, GPS World. It is free to qualified subscribers. An interesting hacker newsletter that deals with navigation and comm of interest to cavers is Frank Reids great Speleonics. Some other GPS info appears from time to time in NASA Tech Briefs. Several other sources of GPS gear include Ashtech,
Bancomm, Texas Instruments, and Trimble Navigation. Trimble has a nice GPS intro book available titled GPS A guide to the next utility. Let me know if you want any more specific info on the GPS frequencies, codings, and user protocols. Ill be happy to get into this further if there is enough interest.
48.5
Hardware Hacker
48.6
AM radio interference 48.1 amateur television 37.5, 40.5 Amecon 47.4 American Colloid Company 48.5 American Design Components 26.4 American Micro Devices 32.3 American Safety Razor 29.3 American Science & Surplus 45.2,
46.4
A/D convert 34.4 Ademco 26.3 Adobe Systems 31.2, 32.2, 33.5, 42.4,
47.2
Adobe Type I Font Format 33.5 Advance Magnetics 48.2 aerogel 30.5 AEE Energy Books 31.5 air conditioner 25.2, 38.1, 38.3 air gaps 44.4
air handler 25.3
Alarm Installer & Dealer 26.3 aliasing 34.4 alien 25.4 alkaline cell 39.3 alkaline "D" cell 25.3 All Electronics 34.5 Allen, Jim 28.4 Alltronics 36.4 Alnico magnet 29.4 alternate energy 28.3, 43.4 Alternate Energy Sourcebook 47.5 alternates to patents 37.1 alternators 44.4, 47.1 altimeters 40.5 Alumilite 33.5
Alvin Marks 46.2
American Society of Materials 41.4 American Voice I/O Society 46.5 ammonia 28.3 ammonium persulfate 28.1 ampere turns 44.4 Amperex 39.2, 39.3, 43.1 amplitude 37.2 amplitude response 34.3 Amuneal 48.2 Analog Devices 26.3, 29.1, 34.3, 34.4 analog front end 43.3 Analog Integrated Circuits 34.5 Analog Products Highlights 34.5 analysis 42.3 Anderson Power Products Inc 40.6 angular velocity 35.3 animation 32.3 animation stands 44.2, 47.1 annodize 30.3 Annulus 39.5 ANSI 47.5 anthill 33.1 Antelope Springs 28.3 antique radio 31.5 antique radio tubes 30.5 Apple Computer 35.1, 40.2 Apple Assembly Cookbook 25.6 Apple IIe 32.3, 35.1, 36.1, 39.1 Apple IIgs 32.3 AppleWriter 36.1 Appliance 30.3, 38.4 Appliance Manufacturer 38.4 arc welder 43.4 arcane applications 25.1 Arco Solar 28.4 areas 42.3 Aremco 47.5 Arius 26.3 Arlin Shepard 30.1 armature 38.1 array 25.6, 42.1 art 40.5
artificial intelligence 25.4 Asea Brown Bovari 38.4 ASCII text strings 32.1 ASCII textfile 31.1 Ashtech 48.5 Ask The Guru 30.5 Ask the Guru III 38.6, 43.4, 44.5 ASM International 31.5 association bookstores 41.4, 41.5 Assn for Computer Machinery 32.2 Assn for the Advancement of Science
41.4
AST Servo 27.3 Astrophysical Journal 32.5 Atochem 39.3 atomic processes 31.1 attaboys 30.1 attenuate 37.4 attenuation 37.2 attraction 38.1, 38.2 ATV 40.5 Audio Amateur 31.1 Audio Engineering Society 41.4 Audio Handbook 46.5 Audio and Video 36.3 Audio & Video Data Book 47.5 audio level meter 33.5, 37.2, 37.3 audio tone 28.2 audio voltmeter 33.4, 33.5 audio/video switches 40.6 audiovisual control 36.2 audiovisual switcher 36.1 auto chips 34.5 autocorrellator 27.1 autodesk 47.2 Automotive Electronics 33.5 automotive ignition 29.5 avalanche effect 33.4 Avantek 41.3 avuncular sleezoids 45.3 AVIOS 46.5 Aware Incorporated 42.1, 42.4
Hardware Hacker III balanced circuits 48.1 bale 37.1 Bancomm 48.5 bandpass filter 34.1 bandwidth 36.3 bare bones startup 33.3 barrel shifting 34.4 baseball trading cards 38.6 BASIC 25.5 bass control 34.1 Batelle 36.2 batteries 43.4 battery 45.3 battery backup 29.5 battery tester 39.3 BBS 26.3, 30.5 Bear, Steve 28.4 Beauty of Fractals 25.5 Belco Electronics 35.1 Bell, FW 29.4 Bell System Record 43.1 Bell System Technical Journal 43.1 Bellcore 43.1 Bellcore Catalog 43.1 Beneath Apple ProDOS 32.4 Bend-Flex 33.5 Benson, Arizona 31.5 Bergey Windpower 28.4 Bessel filter 34.3 best books 31.4 Beta Phase 29.4 big lumps 42.4 "Big Mutha" semiconductors 42.4 biological research 34,4 Binary Coded Decimal 46.3 bingo cards 33.3, 35.4 bismuth telluride 25.1 bit 30.2 BitBlt operator 32.3 bitmaps 31.2, 32.3 black body radiation 39.2 black book 32.2, 33.5, 47.2 Black Lightning 40.2 black magic 26.3, 40.5 Black Spleenwort Fern 25.6 black write 40.1 bladed fan 39.2 blades 29.3 Blatant Opportunist 47.5 blocking filter 48.2 blue light emitting diodes 25.6 Bob Knight 44.2 Bobs Electronic Service 32.5 Bodine 38.4 book resources 31.4 x Book-on-demand 25.6, 26.4, 30.5,
33.5, 34.5, 36.1
Book-on-demand publishing 39.1 Book-on-demand Resource Kit 46.5 Bomarc Services 30.1 Bordens 39.5 Bourns 32.5, 37.5 Bowseretta 28.3, 33.4 breath alcohol detector 28.3 Brian Fenton 41.1 brickwall filters 34.4 bridge 44.4 bridge circuit 30.4 broadcast satellite chips 47.5 bromine 30.4 brushes 31.3, 38.1, 38.2 brushless motor controller 27.5 Buckminster Fuller 43.3 Buckyballs 43.3 Bud 30.3 burglar alarms 26.3 burn 28.1 Burr-Brown 29.1, 34.3, 38.5, 39.5,
41.5
business cards 44.5 Business News Publishing Co 46.5 Business Week 46.2 Butterworth filter 34.1, 34.3 Byte 25.5, 31.1
C 25.5
C&H Sales 30.4, 38.4, 39.5, 42.5 cabbage duster 29.5 cable 33.1, 40.5 cable systems 34.4 cables and fittings 33.5 CAD/CAM 26.3, 27.5, 29.3, 30.5,
31.2, 36.2
Caplug 40.3 car alternator steppers 44.2 Car Audio 26.4 car battery 25.2 carbon dioxide 25.3 Carrizo Solar 46.1 cascade 34.3 Case Against Patents 37.2 cases & enclosures 30.1, 30.2 cases & packaging resources 30.5 cash 46.2 castable resin 36.4 castable urethane 33.5 casting compounds 37.5 Castolite 36.4 cat 38.6 CATV Systems 29.5 Cauer filter 34.3 cave hunting 37.5 C-Cube 32.4, 36.3 CD-ROM 32.4, 39.1 CD-ROM data base 43.3 ceiling fans 38.3 cell polarization 39.3 cellular phone 41.5 Celtec 31.5 center frequency 34.1 Centigrade 34.5 CFR 45.1 chance 33.3 changes 32.3 channel seizure signal 43.2 ChannelLock 40.3 Chaos: Making a New Science 25.5,
47.2
CAD/CIM Roundtable 36.3 Cain Security Systems 26.3 calendaring 44.5 caller id blocking 43.1 caller id chip 43.3, 43.4 caller id data burst 43.2 caller id data format 43.2 caller number delivery 43.1 calculator 34.5 Calogic 27.5 Cambion 25.1, 25.3 Canadian Atomic Energy Commission 25.1 Canon 40.1 Canon fusing units 33.5, 44.5 capacitance inclinometer 30.4 capacitors 34.1
chaos resources 25.4 chaos science 25.1, 25.5 Chaos: The Software 47.2 characteristic impedance 32.4 character smearing 32.4 character string 32.2 Charles, Ron & Associates 31.5 Chebycheff filter 34.3 check sum word 43.2 chemical compound 43.3 Chemtrec 28.1 chief obstructionist 33.1 chillers 25.3 chime 31.4 chips 33.1 Chips & Technologies 26.4 Chomerics 27.5 chroma key 36.3 Cipher Digital 46.3 circuit analysis programs 30.5
Circuit Cellar Ink 31.1, 36.3 Circuit Design 36.4 Circuit Specialists 38.6 Circuit Works 36.4 Circuits Assembly 39.5 clamping paper cutter 27.5 clamshell 30.3 CLASS services 43.1 climb 42.2 clock 38.2 clock frequency 38.5 clockwise 44.3 CMA 33.5 CMOS 27.1 CMOS Cookbook 29.5, 34.5, 35.3,
36.4, 38.6, 40.6, 44.5, 46.5
CMOS Memory 46.5 CMOS Schmidt trigger 26.4 Coburn 30.3 cocobolo 30.3 Code of Federal Regulations 45.1
cogeneration 43.4
coil 44.1 cold 33.5 cold fusion 25.1, 31.1, 33.4, 37.4 cold fusion cell 26.2 cold fusion kits 26.1 cold fusion papers 26.1 cold fusion update 26.2 cold water 28.1 collodial absorbent 48.5 color palette 41.5 color shifts 32.4 command word 30.2 commodity 39.3 common mode 48.1 Communications of the ACM 32.2 Communications Products 47.5 Communications Specialists 43.4 community college 39.4 commutator 38.1 Compendex 42.1 complex signal 27.2 Compliance Engineering 31.5, 45.1,
48.2
composite objects 36.2 compound curves 36.2 compound miter cut 30.4 compress 36.3 compression 32.1 compression ratio 32.1, 32.4 compression stop 30.4 compressor 38.1 Computer Literacy Bookshop 31.5 Computer Shopper 25.6, 44.5
condense 26.3 condenser 28.3 conducted energy 48.1 conducted interference 48.2 conductive adhesive 29.3 conductive pen 36.4 connectors 29.4 Conpad 33.5 conservation 43.4 conservation of energy 38.2 Constantine 30.3 constant percentage 37.3 constant phase 34.4 consumer video 47.2 contact release 44.2 content dependent 32.1 contest 33.4, 35.3, 39.4 continuity break 48.1 continuous research 33.3 Control Automation 36.3 control terminal 31.3 controlled circulation 35.4 controller 25.3 cooler 35.3 cooling 33.4 cooling ideas 25.1 cooling mode 25.2 co-processor 32.4 copper board 26.3 copy machine chassis 33.5 CopyMate Products 40.2 corner frequency 34.1, 34.3 corner rounder 27.5 Cory Laboratories 40.6 COSMAC 34.5 cosmic cupcake syndrome 33.2 cost 34.4 count length 35.1 counterclockwise 44.3 counterexample 33.2 couplings 29.5 CPU 31.1 CR Magnetics 47.4 crap shoot 26.2 crease resistance 44.5 critically damped 34.1 cross axis sensitivity 30.4 cross correllation 32.5 cross-genre classic 32.1 crystal 25.1 crystal set 46.1 Crystek 40.6 Cryptosystems Journal 32.5 C60 43.3 cubic 45.4
cubic polynominal 45.4 CUPL demo disk 29.5 Curie Point 33.4 current 47.2 current limiter 47.2, 47.3 current monitor 47.4 current noise 29.2 current sensor 47.4 current sensing 47.2 current steering 26.4 current transformer 47.2, 47.4 cutoff frequency 34.1 cyan 29.1 Cycolor 29.1 Cycle-Flex 33.5 cycloconverters 38.4 cyliths 29.1 Cypress Semiconductor 46.5 cyrogenics 33.5 CX cartridge refilling 40.2 CX engine 30.1 CX Glompenstractor 40.2 CX refilling details 40.2
damper diode 48.4 damping 34.1, 34.3 data acquisition 46.4 data bit 46.3 Data Communications Handbook
30.5
data compaction 25.6 data compression 32.1, 36.3 Data Conversion Products Databook
26.4
data message format 43.2 data modulation 32.1 data path 43.1 data points 45.4 David Gaddis 47.5 DBM 37.4 DC current measurement 47.1 DC-DC converter 29.5 dc output 27.2 DCT 36.2, 36.3 DCT filters 42.4 DCT transforms 42.3 deblur 42.2 decibels 37.2, 37.2 Deco Shapes 29.3 decoder 46.3 xi
Hardware Hacker III decoupling 29.2 deep space probes 27.2 Defense Electronics 40.6 deflection power 48.3 deflection yoke 48.3 delta modulation 32.3 dentistry 29.4 depletion region 41.1, 41.2 descrambling circuits 34.4 Design News 38.4, 39.5 Designing With Plastics 30.5 DesignTech 41.5 desktop 36.2 desktop finishing 39.1 Desktop Prototyping 36.1 despiker 44.1 despiking diode 44.1 detail 25.6 detector 39.3 deuterium 25.1, 26.1, 26.2 device space 42.2 dew point circuits 26.1 dew point sensors 25.3 dew sensor 26.3 dew sensor test circuit 26.4 diac 31.4 Dialog Information Service 31.4,
33.1, 33.3, 42.1, 46.2 diamond 43.3 dictionaries 32.3 die-cutting 27.5 die punch 27.5 dielectrics 40.6 Diers, Don 30.5 difference 27.1, 27.3, 48.1 differential 39.4 differential scheme 48.5 differentiations 34.2 difficult terrain 37.5 diffraction gratings 30.3 digital 29.4 digital audio 47.5 Digital Digest 42.5 digital encoder 46.5 digital potentiometer 38.4, 38.5 digital signal processing 34.5 Digital Signal Processing 38.6 digital signal processor 34.4 digital sinewave generator 35.3, 36.3 digital synthesis 27.3 digital thermometer 34.4 digital transform 42.1 digital voice recording 26.4 direct mail books 33.4 direct method 47.3
direct order line 30.1 direct toner printed circuit 26.1, 26.3,
39.1
dirty environment 29.5 disallowed sequence 35.2 disallowed states 35.2 disco stars 28.2 discontinued semiconductors 44.5 discrete cosine transform 32.4, 36.2,
36.3, 42.4
editor 31.1 EDLCO 30.3 Edmund Scientific 25.3, 45.2 EDN 33.3, 35.4 Educational Products 31.5 EE Times 31.1, 33.3, 35.4 EEPOT 38.4 EEPROM 38.6 EER efficiency ratings 25.3 Efficient Electrical Systems Design
41.4
Discrete Semiconductor Catalog 30.5 disk drive 44.2 disk formats 32.1 display PostScript 31.2 displays 29.3 Distillery 31.2 distinctive ringing 43.2 distortion 37.3 Document Structuring 31.2 Don Diers 30.5 Don Thompson 30.1 DonJer 29.5, 30.3 Dopeyballs 43.4 doppler 27.2 dot and bar generator 42.5 Doubleday Science Series 33.4 doubly balanced modulator 27.1 downconverters 41.3 downconverting 27.2 dreams 28.3 drift 41.3 dross 28.2 drug deliveries 37.5 dry ice machines 25.3 dry ice temperatures 25.3 DTMF 39.5 dual comparator 39.3 dual stators 38.4 dual video driver 32.5 DuPont 26.3, 36.2 durability 44.5 Duracell 39.3 DV-I 32.4 Dvorak keyboard 39.1 DWI breath alchol tester 28.4 DynaArt 26.3 dynamic RAM technology 31.1 dynamic range 37.3
EIA standard RS-232 47.5 elastomeric battery clip 33.5 Elector USA 42.5 electric drill 38.1 electric field 48.1 electric motor resources 38.4, 38.6 Electric Power Research Institute
43.4, 46.2
electrical wholesale supply 39.3 Electricity by Serendipity 31.1 electrolytic level 30.4 electrolytic sensor 30.4 electromagnetic 44.3 Electronic Buyers News 37.5 electronic compass 29.4 Electronic Component News 30.3 Electronic Components & Applications 33.5 Electronic Design 33.3, 35.4 electronic dice 35.3 electronic die 35.2 electronic dog tags 46.3, 46.4, 46.5 electronic funds transfer 32.1 Electronic Goldmine 33.5 Electronic Industry Telephone Book
40.4
E field 45.2
earth sheltered houses 28.3 eaves-drop 41.5 EconoRAM 30.1, 30.2 edge information 42.2 edges 42.3
electronic levels 30.4 electronic multiplier 27.1 electronic music 27.2, 34.4 electronic musician products 46.5 Electronic News 37.5 Electronic Packaging 30.3 Electronic Products 33.3 electronic protractors 30.4 electronic trade journals 35.4 electronic tubing 46.5 electronic tuning 41.1 Electronics 33.3 Electronics Test 48.2 elegant simplicity 35.1, 39.3, 39.4 elevator 36.2 Elliptical ripple filter 34.3 ELTEC 37.5 Emergency Response Guidebook
xii
Emerging Technologies 39.1 EMERGOP.TXT 39.1 empty cartridges 40.3 Emulation Technology 34.5 Enco 39.5 Encapsulated PostScript 31.1 enclousure 46.3 Encyclopedia of Associations 33.3,
33.5, 41.3, 47.5 enemy 30.1 energy 33.2
energy conservation 48.3 Energy Depot 28.4 energy management 46.5 entropic heat death 33.1 entropy 32.1 environment 33.1 environmental quality 43.4 EPA 28.1 epoxy 48.2 EPRI 43.4, 46.2 EPRI Journal 25.1, 47.5 EPROM 29.5, 37.5, 44.4 EPROM samples 25.6 equal change 37.3 Equal Component Value filter 34.3 equal steps 37.3 equally tempered music 45.3 error intolerance 32.1 errors 32.1 ET watching 32.5 etchant 28.1 eutetic liquid 28.2 evaporator 28.3 EVCO 40.2 even scan lines 46.3 Exair 25.3, 35.3 Exar 34.4 Excalibur 36.4 excess heat product 26.1 Exel 39.5 exit temperature 35.4 expert systems 25.4 explode 28.1 exponential decay 31.4 extended range 29.5 extension phone 41.4 Extraordinary Science 34.5
Farenheit 34.5 Fasco 38.4 FASTFERN.PS 25.6 Fast Fourier Transform 42.2 FAX 31.2, 32.1, 37.5 FCC 31.5, 45.1 FCC Part 68 interface 43.4 FCC regulations 45.1 FCC specs 45.1 FCC Type 68 approval 43.3 FDM 40.4 Federal Information Center 47.5 feedback 27.1, 27.3, 41.3 felt 29.5 femtowatt 37.3 Fenton, Brian 41.1 FerriShield 48.2 ferrite bead 48.2 ferrotype drum 44.5 Ferroxcube 48.2 fiber optics 25.6 field 48.3 Field Programmable Gate Arrays
48.5
field rate 46.3 Figaro Engineering 28.3 filter 32.4, 34.1 filthy rich 25.4, 37.1 finish 40.5 fire attack strategy 28.1 fire department 28.1 Fisher-Klosterman 46.5 5-band graphic equalizer 28.2 555 timers 30.4 flame detector 28.2 flangers 27.2 flash technology 44.4 Flemings Rule 38.1 flicker 31.4, 46.3 flip flops 35.2 flock 30.3 floor sweepings 37.5 flux concentrating pole pieces 29.5 fluxgate magnetometer 27.4, 29.4 flyback 48.4 flyback capacitor 48.4 flyback deflection 48.3 flyback time 48.4 flyback transformer 48.4 foils 30.3 folder 27.5 Fomeboards 30.3 font cache 32.3 force sensitive inks 27.5 format 43.3
Fotofoil 30.3 four layer diodes 31.4 4-stroke engine 32.3 four terminal connection 47.3 four terminal techniques 47.1 Fourier analysis 38.5 Fourier synthesis 45.3 Fourier transformation 32.5, 42.1 FM Capture effect 41.1 FM reception 41.1 FM stereo broadcasting 44.1 FM stereo transmitter chip 47.1 Fractal FactoryV47.2 fractal fern 25.4, 25.5, 45.3 fractal geometry 25.5 Fractal Geometry of Nature 25.5 Fractal Resources 25.4 Fractal Slide Projector 47.2 fractals 47.2 frame rate 46.3 free samples 46.4 Freedom of the Press 31.2 Freeman Supply 29.3, 36.2 Freon replacements 33.5 frequency 29.2, 34.3 frequency conversion 41.3 frequency domain 38.5, 42.3 frequency selective network 34.1 Fresnel lens 39.2, 39.3, 43.1 Frontier Space Books 31.5 frost sensor 47.4 frostbite 25.4 fume and vapor detectors 28.3 Fused Deposition Modeling 40.4 fuser assembly 30.1 fuser roller 30.1 fusion roller 40.1 fuzzy data 45.3 fuzzy logic 31.1 FUZZYFIT.PS 45.3, 45.4
Hardware Hacker III Geneva Stop Mechanism 46.3 GEnie 30.5, 34.5, 36.1, 46.2 GEnie PSRT 35.1, 38.6, 39.1, 40.4,
41.1, 42.4, 42.5, 43.4, 44.5, 45.3, 46.5, 47.5 genlock 46.2 geodetic dome 43.3 geophysics 34.4 GhostScript 47.5 gigawatt 37.3 Gila Monsters 33.2 glass etching 27.5 Gleick, James 25.5, 47.2 globally 38.6 Global Positioning Satellites 48.5 glompenstractor 40.3 glue gun 40.4 glut 37.4 Godzilla vs the Night Nurses 32.1 gold plating 40.3 Golly gee, Mr. Science 26.2 gonzo justification routines 31.2 good deed 35.1 Goodyear blimp 37.5 GoScript 31.2 GPS 48.5 GPS: A guide to the next utility 48.5 GPS antennas 48.5 GPS World 48.5 graphics equalizer 38.6 WW Grainger 38.4 Grantham 36.2 granular powder 36.2 graphite 43.3 grinder 39.4 GRIPS-2 CD-ROM 43.4 GripWorks 29.3 ground currents 29.2 grounded sync tips 32.5 group delay 34.4 Group III compresion 32.3 Guard Box 46.4 Gurus & Swamis Union 41.1 gyroscope 26.4, 33.1, 36.1, 41.4
Hall voltage 29.4 Haltek Electronics 38.6 Hamlin 30.4 Hammond 30.3 hand-coded program 25.4 hand punch 39.4 handle 29.3 Hansen, Steve 48.5 hard coated drums 40.2 hard disk formats 32.3 hard drum recoating 40.4 hard limiting 41.1 Hardware Hacker 36.1, 40.4, 43.4,
46.5, 47.5
hardware hacking 37.1 harmonics 34.1 Harris Publishing 40.4 Haygeman Machine 25.6 hazardous materials 28.1 Haz-Mat 28.1 HDTV 32.1, 33.1, 42.3 heat flux 25.2 heat pump 25.1 heat sealing iron 37.5 heated roller 26.3 heater 31.3, 34.5 Heath 33.4, 39.3 Heathkit 43.1
heating mode 25.2 heatsink 25.2, 31.3
heatsink efficiency 25.1 heavy duty stuff 25.6 heavy water 25.1 Hello Direct 44.1 help line 33.4 Henry Yen 47.2 Herbach & Rademan 38.4, 39.3, 42.5,
43.1
H field 45.2
H&A Industries 31.5, 33.1 H&R Enterprises 38.6 hacker magazine 26.1 Hackers Holy Grail 35.4 Hagen, Jerry 40.6
Hall effect 25.6, 29.3 Hall effect devices 29.4, 29.5 Hall Effect Sensors 29.5 Hall Effect Transducers 29.5
Hewlett-Packard 30.1, 34.5, 40.2, 40.6 High Current Transistor 33.5 high definition television 42.3 High Energy Enterprises 48.1 high energy resources 45.1 high gloss 44.5 high pass filter 34.1 High Speed CMOS Logic 32.5 hippie commune 28.3 HIRES 32.4 histogram 42.3 Hitachi 36.3, 47.5 hitch works 30.3
hobby modelmaking 31.5 Hoechst Celanese 30.5, 34.5 hogwash 45.1 holding tank 40.1 Holy Grail 26.2 Home Automation USA 47.5 home energy monitor 47.4 Home Power 28.3, 46.2 Home Shop Machinist 39.5 homebrew cases 30.3 Homebuild Rotorcraft 40.6, 41.5 homodyne detector 27.1 homodyne detection 27.3 Hong Kong 47.5 horizontal line 46.3 horizontal resolution 48.2 horizontal scale 42.2 horizontal shift 42.2 horizontal yoke 48.3 horror stories 44.4 horse trailer 30.3 Hosfelt Electronics 31.5 host controller 30.2 Hostess Twinkies 39.4 hot foil die-stamping 27.5 hot-knife sign cutters 27.5 hotline 43.1 hot mercury 37.1 house organ 33.5 How to Automate your Home 47.5 howled 27.3 HPGL 26.3 HSTOGRAM.PS 42.4 hue 27.2 Huffman 36.3 Huffman coding 32.2, 42.3 Huffman compression 32.1 Hughes 35.3 human brain 25.4, 31.1 human vision 42.3 humidity control 46.5 humidity sensors 26.3 Hurst 40.4 HVAC News 28.4 hydrogen 31.1 hygrometer 26.3, 29.5 hyperabrupt profile 41.2 HyperCard 25.4 hysteresis 38.3
xiv
Hardware Hacker III IF amplifier passband 34.4 IGVT 34.5 IGUS 26.4 illustration packages 31.2 image compession 25.5, 32.1 Image Sensors 32.5 incandescent lamps 31.2 inclinometer 27.3, 30.4 Incredible Secret Money Machine
25.4, 26.2, 26.4, 27.5, 28.2, 29.4, 30.5, 31.5, 33.5, 34.5, 35.3, 35.4, 36.3, 36.4, 37.2, 37.5, 38.4, 39.4, 39.5, 41.4, 44.5, 46.5
Integrated Circuits Data Book 38.6 Integrated Device Technology 38.6 integrated programs 25.4 integrations 34.1 Intel 31.5 intelligent power drivers 47.5 interaction 37.3 interface 31.3 interfering signal 27.2 interlace 33.1, 46.3 interlace scan 48.3 Interlibrary Loan Service 33.3, 36.3,
42.1
independent research 33.2 indicating spot 39.4 indirect method 47.3 indium 28.2 induced currents 48.1 induced voltages 48.1 induction motor speed control 38.1 inductive loads 44.1 inductors 34.1 Industrial Education 36.2 Industrial Micro 45.1 industrial waxes 42.5 industry trade journals 31.4 inefficient 25.1, 25.2 inept motions 33.3 inertia 38.1 Infochip Systems 32.4 Infocom 32.1 information theory 32.1 infrared detectors 25.3 infrared film 37.5 infrared laser 36.2 infrared people detectors 39.2, 43.1 infrared signature 39.1, 39.3 initialize 30.2 initgraphics 31.2 inorganic solids 30.5 in-phase 27.4 in-phase term 27.2 Insiders Guidt to Rapid Prototyping
48.5
Intermark 48.2 intermittent 38.2 internal resistance 39.3 International CMOS Technology
38.6
International Micro Electronics 38.6 International Rectifier 34.5 International Security Conference
26.3
International Standard Periodicals Dictionary 26.3, 33.3 interpreted language 25.4 Intro to PostScript video 26.4 Introduction to Radar Systems 40.6 inventor 37.1 Inventor-Entrepreneur Network 41.5 inverse transform 42.2 I-Q demodulator 27.1 irresponsible manufacturers 40.3 ISO 47.5 isometric lettering 42.2 isopod 47.4, 47.5 Isotech 25.1 Iterated Systems 47.2
Lancaster, Jeff 26.3 large display letters 40.5 laser toner image 44.5 LaserWriter 35.1 LaserWriter Corner 25.6 LaserWriter NT 30.1 LaserWriter NTX 30.1, 40.3 LaserWriter Reference 30.1 LaserWriter Secrets 30.5, 42.5 LaserWriter technical information
30.1
Inspec 42.1 instability 27.3 instant cold 25.3 Instant Modeling 36.1 instant suede 29.5 Institute Electrical & Electronic Engineers 41.4 insulation 25.2 integrated circuits 29.5
Jones, Marlin 36.4 Jones & Bartlett 42.1, 48.1 Journal of Electroanalytical Chemistry 25.1, 37.4 JPEG 32.4 JPEG compression 33.1
latch 31.4, 36.4 latent image 29.1 lateral time code 46.1, 46.2, 46.3 lathe 39.4 laws of thermodynamics 33.2 Lazer Products 30.1, 33.5, 40.1 L/C ratio 34.1 LD-50 load detector 47.4 lean 42.2 least squares 45.3 least squares curve fitting 45.3 LED 37.4, 37.5 LED output drivers 31.4 Lehman Scientific 26.4 lens design 39.3 xv
Hardware Hacker III lethal shock hazard 31.3 letterhead 35.4 leuco dye 29.1 Leviton 39.3 library research 33.1 Lichtman, Jeff 32.5 Life Sceince Library 33.4 light dimmer 38.2, 48.1 Light Emitting Diode 27.5 Light Machines 36.2 Lightwave 25.6 Lindsay Publications 31.5, 33.1, 38.4,
39.5, 45.1 low cost memory 30.2
line frequency 38.2 line-of-sight 48.5 linear 29.4, 37.3, 45.4 linear actuators 25.1 linear current ramp 48.3 Linear IC Data Book 28.3 Linear Integrated Circuits 29.5 Linear Data Book, 1990 27.5 Linear Data Manual II 27.4 Linear Devices 38.6 linear graphics transform 42.2 linear predictive coding 32.3 linear sinewave 27.3 linear stepper 44.4 Linear Technology 27.5, 29.1, 30.5,
34.3, 34.4
low noise amplifiers 29.1, 41.3 low noise preamp 29.2 low pass active filters 34.2 low pass filter 34.1 34.3 low resistance 47.3 low temperature alloys 28.1 LSI Logic 40.6 LSI Systems 31.4, 47.4 lubricants 43.4 Lucas Sensing Systems 30.4 Lumeloid 46.2 lunchbox handle 40.2 LVDT 27.3, 36.4 LX cartridge recharging 40.3, 40.4
Materials Catalog 47.5 Materials & Electrochemical Research 43.4 Materials Safety Data Sheet 28.1 math 27.2 math freak 25.4 math rule 42.1 Mathsci 42.1 matter 33.2 Maxim 29.5, 32.4, 34.3, 34.4, 34.5,
45.2, 47.4
McMaster Carr 39.5 Mead Imaging 29.1 Measurement & Control 27.3, 47.3 mechanical goodies 29.3 Medical Electronic Products 41.5 Medical Equipment Designer 41.5 Melcor 25.1, 25.3 melody 31.4 melody chip 31.4, 35.2 memory 30.1 Memory Card Systems & Design
44.5
linear voltage differential transformer 27.3 linearity 37.2 liquid crystal 39.3, 39.4 liquid cyan 29.1 liquid gases 33.5 Liquisorb 48.5 lithium hydroxide 26.2 load detector 47.4 load levelling 47.4 load resistor 39.4 load shedding 47.4 locally 38.6 Loch Ness 34.5 lock-in amplifier 27.1, 27.2 locomotive 42.4 log 37.3 log amps 33.5 logarithmic 37.3 Logic Databook 38.6 Logical Devices 29.5 Lorenz attractor 32.5 loss 37.2 lossless 32.1 lossless transform 42.2 lost Guru columns 35.1 xvi
magnetic refrigerator 35.3 Magnetic Shield Corp 48.2 magnetized roller 40.1 magnetocaloric effect 33.2, 33.4 mailer 42.5 mailing list 32.1 Management Roundtable 48.5 Manchester coding 46.3 Mandlebrot 25.5 Mandlebrot Set 25.5 Manley-Rowe relations 41.3 Markel 46.5 Markoff scheme 32.4 Marks, Alvin 46.2 Marlin Jones 36.4 Master Product Catalog 46.5 masterCAM 36.2
Memory Data Book 30.5 mercury 30.4 mercury activator 28.1 Meridian Data 43.4 Merigraph 36.2 mesmerizingly awful reps 30.1 message length word 43.2 message type word 43.2 metal locator 27.4 Metalphoto 30.3 Metals Handbook 41.4 metric gears 25.6 Micro Cookbook I 36.4, 41.5 Micro Cookbook II 36.4, 44.2 Micro Linear 36.3 Micro Linear Data Book 33.5 MicroCAD News 30.5 microcapsule technology 36.4 microcomputer triac interface 31.2 microcontroller 43.2 Microcontroller Handbook 34.5 microcontroller 25.4, 30.4, 36.4, 43.3 Micron 33.5 Microperipheral Products Data Book
29.5
Hardware Hacker III microwave components 40.6 microwave frequencies 40.5 MicroWave Journal 40.6 Microwave Product Digest 40.6 Microwave Radio Astronomy 32.5 Microwave Transistors 32.5 Microwaves & RF 40.6 MIDI 46.3 Midnight Engineering 26.2, 27.5,
29.5, 31.1, 35.1, 35.4, 36.1, 37.2, 38.6, 39.1, 39.5, 42.5, 43.4, 44.5, 46.5 military infrared viewers 30.5 mine tailings dumps 30.5 mineral 27.4 Ming Engineering 26.4 Mini-Circuits Labs 40.6, 41.3, 48.5 misdirection 28.1 misnamed chip 27.3 MIT 35.3 MIT Rad-Lab 40.5 Mitsubishi 30.4 mixers 41.5 mixing 27.2 Model Railroader 37.5, 39.5 Modeltec 39.5 modem 41.4 modem tones 43.2
multiple feedback filter 34.2 multiple phase windings 38.3 multiplex 25.1 multiplication 37.2 multi-sync monitors 48.3, 48.4 Murata 26.3, 48.2 music 42.1 musical 37.3 MUTE 36.2 MWK Industries 29.5 My First PAL Design 29.5 MYFAVOR.TXT 40.4
N doping 25.1
NAND 35.2 nanopower memory 30.2 nanovolts 29.2 NASA Tech Briefs 32.5, 42.5, 48.5 National 27.1, 30.5, 33.5, 34.3, 38.6,
43.4
nozzle 35.3 NPN transistor 26.4 NRAO observatory 32.5 NSRA News 39.5 NTSC 27.2, 46.5, 48.3 NTSC color monitor 48.5 nudist colony 37.5 nuron computing 33,1 null 27.3 null seeking 27.3 number extractor 43.3 numdots 25.6 NURBS splines 40.4 Nuts & Volts 31.1, 34.4, 40.5 nylon bearings 26.4 NYNEX 43.1
Modern Resonance Transformer Design Theory 45.1 modified Huffman codes 32.3 modulo 35.2 moisture sensing paint 26.3 Mokrytzki, B 46.5 moldmaking 42.5 Molon 38.4 momentum 35.4, 35.3 monitor 48.4 monitor fundamentals 48.2 MOS Data Book 33.5, 46.5 Mother Earth News 28.4 motion 35.3 Motion 38.4, 42.5, 44.4 Motion Control 39.5, 42.5 motion detectors 30.5 motor 31.3, 44.1 motor commutation 29.5 motor speed 38.4 Motorola 27.5, 30.5, 31.3, 34.4, 38.2,
40.6, 41.2, 42.5, 44.4
Motortechnics 46.5 Mouser Electronics 30.3, 43.4 MOV 44.2 movie film 32.1, 46.3 multilayer prototyping systems 34.5 multiple feedback 34.3
negative resistance devices 41.3 neighbor 41.5 NE-2 neon lamp 28.2 net power generation 45.1 Network Analysis and Synthesis 34.2 neuron computing 31.1 Never The Same Color 48.3 New Age nineties 33.1 NewAge Industires 36.4 Newark Electronics 39.5 New Equipment Digest 39.5 NeXT computer 39.2 Nickel 36.4 nickel wire 26.2 nickels 37.1 911 number 28.1 NIST 47.5 noise 29.2 noiseless 32.1 noisy data 45.3, 45.4 noisy source 29.2 noninterlaced display 48.3 nonlinear 41.2 nonlinear harmonics 38.5 Northern Inc 38.4
OrCAD 27.5 order 34.2 organic materials 28.1 osmium 43.4 OTP EPROM 29.5 outright zoo 30.1 overdamped 34.1 Owls Mask 32.5 oxidizer 28.1 oxygen 31.1
P doping 25.1
P-38 can oener 39.4 P-51 Mustang 37.5 xvii
PacTec 30.3 pad printer 27.5 padding press 27.5 PAL 46.5 palladium 26.2, 31.1 Palmer, Wyatt 26.3 Panasonic 26.3 pangalactic happy faces paper clip 47.3 paper drill 27.5 paper jogger 27.5 parabolic 45.4 paradigm 39.1 paradigm shifts 39.1 parametric amplifier 41.3 paranoid secrecy 37.1 passive filters 34.1 part 15 45.1 part 68 45.1 part 68 specs 43.3 partial derivatives 45.4 patent examiner 37.2 patent office 33.1 patent process 37.1
patent victim 37.1
patents 37.1, 42.2 pattern recognition 38.5 patterns 42.3 Patti, Toni 32.5 payload 37.5 PCIM 38.4, 42.5, 44.4 PC Techniques 42.5 peakedness 34.1 Pelican Manufacturing 46.4 Pelier cooling 39.1 Peltier effect 25.1 pen plotters 44.2 people detector 39.2 perfect binder 27.5 perforating 27.5 permeability 48.2 perpetual motion 33.1, 36.1 persistence of vision 48.3 personal computer 39.1 Personal Engineering 46.5 personal experiments 33.4 personal network 33.4 personal resource file 33.3 persistant downloads 32.3 PH handbook 29.5 phase 27.1 phase angle 27.2 phase control 31.3 phase controlled dimmer 31.2 phase detector 37.1, 37.3 xviii
phase detection 27.2 phase locked loops 25.6, 27.2, 33.5 phase-locker 27.1 phase plane plots 45.3 phase-sensitive detector 27.1 Philips 27.4, 30.5, 33.5 Phillips 25.6, 32.5, 39.2 Phillips/Signetics 39.5 phosphor 48.3 photo film 29.1 photochemical machining 36.2, 37.5 Photocomm Inc 28.4, 44.5 photopolymer 36.2 photosensitive drum 40.1 phototriac isolator 31.3 phototriac optoisolator 31.1 Photovoltaic Network News 46.2 Photovoltaic Systems 28.4 physical chemistry 43.4 picoamperes 29.2 picture insertion 36.3 pixie dust 33.5, 40.2, 40.3 pizza style service 40.3 PLA 44.4 plastic gloves 28.1 plastic sheet 29.3 plastic tubing 36.4 Plessey Semiconductors 40.6, 42.5 Plessey Signal Technologies 34.4 plotters 44.2 plumb bob 30.4 plug-in cartridge 40.1 PM Research 36.4 PMI 30.5, 34.3, 34.5, 46.5 POEM Fractal Factory 47.2 POEM Fractal Slide Projector 47.2 poison 30.4 Pohns and Fleishman 33.4 pole shading 38.3 poles 38.2 Pollution Equipment News 28.1, 28.3 Polycase 30.3 polyester carrier 29.1 polynomonials 45.4 Popular Electronics 31.1 Popular Science 28.4, 37.5 popularity 42.2 positioning sensors 36.4
position to voltage transducer 27.3
positive cash flow 37.1 PostScript 25.4, 25.5, 26.2, 31.2, 31.5,
34.5, 35.4, 36.3, 39.1, 40.2, 40.4, 42.2, 44.5, 45.3, 46.5, 47.5 PostScript Cookbook 42.5 PostScript fractal fern 25.1
PostScript laser printers 40.1 PostScript level II 42.4 PostScript PSRT 39.5 PostScript Reference Manual 42.5 PostScript Roundtable 35.1 PostScript Secrets 25.6 PostScript Show & Tell 31.3 PostScript fonts 47.2 potentiometer 38.5 Power and Bulk Solids 29.3 Power Bipolar Transistors 33.5 power control 27.5 Power Control IC Handbook 29.5 power density 45.2 power diode 44.1 power electronics 42.4 power feeds 44.2 power line hum 48.1 Power Modules 46.5 Power Quality 43.4 power resistor 39.3 Power Series Approximation 45.4 power stapler 27.5 power steppers 47.1 power terminals 31.3 PowerMOS 27.5, 32.5 Powerpole connectors 40.6 PowerTech 33.5 Powertechnics 46.5 Precision Monolithics 29.1 pre-etch 26.3 pressure roller 29.1 Pressure Sensor 46.5 primary circuit 47.3 printed circuit 26.2 Printed Circuit Design 46.5 printed circuit hacker standard 31.1 printed circuit layout demo 27.5 printed circuit materials 40.6 printer paper feeds 44.2 printer warranty 40.3 Printer Works 40.2 printing 27.5 printing resources 27.4 printshops 27.5 prismatics 30.3 probabilities 25.6 probability table 25.6 Probe 31.1 Proceedings of the IRE 32.2 processing times 32.1 product development 37.1 Product Selection Guide 31.5 Professional Glass Consultants 39.5 professional societies 41.4
Hardware Hacker III professional video 46.2 programmable gain amplifier 27.1 programmable inverter 27.1 programmable logic 44.4 Programmable Logic Devices 39.5 programming info 44.4 propylene 30.4 Protection Devices 36.4 protection techniques 44.1 prototype 31.5, 36.1 prototype models 29.3 pseudo-compiling 31.2 pseudorandom sequence generator
39.1
pseudorandom sequences 45.3 PSRT 46.2 psychedelic lighting 39.1 public domain 37.1 Pulaski 37.2 pulleys 38.3 pulse width modulator 29.5 pump frequency 41.3 punch and go method 40.2 pure math 45.3 PVC cement 29.3 pyroelectric chip 43.1 Pyroelectric Detectors 37.5 pyroelectric infrared detector 39.2
Q 34.1, 48.1
QMS PS-410 40.3 quadratic 45.4 quadrature 27.4, 35.2 quadrature art 45.3 quadrature sinewave 35.2 quadrature term 27.2 Quadrax 36.2 quality factor 34.1 Quality Software 32.4 quantization noise 34.4 Quo, Daniel 47.2 Quick Shooter 25.6
radio astronomy 32.5, 40.5 Radio Astronomy Handbook 32.5 Radio Astronomy Transactions 32.5 Radio Control Modeler 37.5 Radio Electronics 33.5, 39.4 Radio Electronics BBS 41.1
Radio Observer 32.5 Radio Research Instruments 40.5 Radio Shack 30.3, 31.3, 34.4 radius of rotation 35.3 random clicks 28.2 random number generator 25.6 random process 25.6 rare earth 33.4 Rare Earth Resources 33.5 rate of change 44.1 ratio 34.1, 35.4 rats-nest 32.5 raw PostScript 36.1 Raytheon 29.5 RE editorial 41.1 read 30.2 read-write memory 46.4 real editing 46.2 Real Goods 28.4, 31.5, 47.5 real time 32.1 rear axle ratio 33.2 receiver "S" meter 33.4 recharged cartridges 40.1 Recharger 30.2, 40.2 recharging supplies 33.5 recoated SX drums 30.1 Rectifier Diodes 33.5 recurrent flyback deflection 48.3 recurrent flyback sweep 48.2 recycle 29.3 Red Spot 30.3 redundant 32.1 Reed motors 48.1 refill 30.1 reflect 32.4 regulated speed control 38.2 relative strenght 37.3 relaxation oscillator 28.2 relay 44.1 relay dropout 44.1 religion 33.1 REM statements 32.3 remarks 31.2 remote control toy 28.3 Remote Controls 42.5 remote load regulators 47.1 repair tools 30.1 repelling magnets 33.1 replication 36.1 repulsion 38.1, 38.2 resealing strips 40.3 research 33.2, 33.4 reset 35.2 resonance 34.1, 41.2 Resonance Research 45.1
resonant circuit 41.1 resonant energy storage 45.1 response curves 28.3 retired machinist 39.4 retrace 48.4 reverse 38.1, 38.3 RF breadboards 42.5 RF Communications data handbook
41.5
RF Components 29.5 RF/IF Signal Processing Guide 48.5 RF impedance 41.2 RF Power Module 27.5 RF Power Transistor 25.6, 27.5 RF Prototype Systems 42.5 RF remote controls 30.2 Ridenour 40.5 right hand generator rule 38.1 rip off 37.1, 40.4 roast beef 36.1 robotics 44.2 robotics applications 33.5 Robotics Now 39.5 Rochester Electronics 33.5, 44.5 Rockwell 48.5 Rogers Corp 26.3, 33.5, 40.6 Rohm 40.6 Roland Digital 36.2 Rolf Taylor 42.5 Ron Charles & Associates 31.5 room ambient 25.2 rotary coolers 35.3 Rotary Flight International 40.6 rotary power 38.1 rotate 42.2 rotating shaft 25.1 rotor 38.1, 44.3 rotor flux path 44.4 rotor winding 44.3 RoundTables 35.1 rough 40.5 router 39.4 roving satellites 48.5 RS-232 driver 29.5 RTL digital integrated circuits 39.1 RTTY Journal 42.5 rubber stamps 27.5, 36.2 rude surprise 37.1 rule of thumb 37.3 run length encoding 32.2, 32.3 Rutland 39.5
Hardware Hacker III safety isolation 31.3 Sallen-Key 34.2 Sallen and Key 33.4, 37.2 sample 29.3, 29.4, 34.5, 35.1 sample copies 26.2 Samsung 26.4, 28.3, 32.5, 37.2 sand 30.3 sand dune 41.1 sandblasting 40.6 Sandia National Laboratories 28.4,
35.3
Santa Claus update 36.1 Sanyo 25.6 Satco 43.4 Satellite Receivers 29.5 satellite 25.3, 33.1 Satellite Cable & TV 40.6 satellite downlinks 40.5 satin finish 30.3 saturation 27.2 "Save my baby!" 28.1 SAW device 34.4 SAW filters 34.4, 41.1 scams 37.1 Schaevitz 36.4 school report 26.2 school research paper 43.4 School Shop 36.2 Schottky diode 45.3 Science 26.1, 26.2, 29.5, 30.5, 31.1,
32.5, 41.4, 43.4, 46.2
Science Citations Index 33.4 science fair 25.3, 26.2, 43.4, 45.2 Science Frontiers 34.5 Science News 46.2 Science of Fractal Images 25.5 Scientific American Amateur Scientist 37.5 SCOPE Octal drivers 38.6 Scope Evaluation Kit 29.5 scoring 27.5 Scotch tape 40.3 SCR 4.14, 42.4 scroll compressor 25.3 SCSI 34.5 search 46.3 Sears 39.5 second harmonic distortion 41.2 second order filter 34.1 Security Dealer 26.3 Security Distributing & Marketing
26.3
Security Show, The 26.3 security systems 26.3 seismometry 27.3 Selectrim 32.5 self-clocking code 46.3 self-healing plastic 34.5 self-lubricating bearings 26.4 selfsimilar copy 25.6 self-starts 38.2 Self-Stick Chips 46.4 Semiconductor Databook 31.5, 34.5 Semicondutor Diode Parametric Amplifiers 41.3 Semiconductor Product Guide 26.4 sensing magnetic fields 29.4 sense 38.4 sensor 28.3, 39.2 Sensors 27.3, 47.3 sequence 35.1 Serco 30.3 serial clock 36.4 serial data stream 43.3 serial input data 36.4 serial interface 36.4 serial number 46.4 series DC motor 38.1 series motor 38.1 series resistance 41.2 serious researchers 26.2 Serpenski triangle 25.3 service 40.1 service bureau 36.1 service manual 40.1, 40.2 service records 40.2 SETI activities 32.5 settransfer 31.2 sewing machines 35.3 SGS-Thomson 29.5, 31.3, 33.5, 36.3,
38.2, 42.5, 44.4, 44.5 shaft encoder 27.2, 29.5 shake-the-box kit 45.2 shape alloy wire 29.4 Sharp 29.5, 30.5, 34.5 shear 39.4 sheet steel 48.2 shell 29.1 Shepard, Arlin 30.1 shield 48.1 shielded cables 32.4 shielding 29.2 shields & sheilding 48.1 shift registers 39.1 shock hazard 31.3 shoddy quality 30.2 shoe polish 40.4
shop courses 39.4 Short Circuits 33.5 shorted turn 48.1 shotgun technique 37.1 show-and-tell 26.2 Shreve Systems 42.5 shrink wrapper 27.5 shunt 47.3 Sickels, Bob 32.5 sideband filtering 34.4 side-looking 27.2 Siemens 34.5, 48.2 Sierra Semiconductor 43.3, 43.4 SIGGRAPH 25.5 sign cutters 44.2 signal conditioner 27.3 signal path 27.1 SignCraft 31.5 Signetics 25.6, 33.5, 41.5, 46.5 silicon controlled rectifiers 39.1, 41.4 silicon power diodes 41.2 silicon solar cell 46.1 Silicon Systems 29.5, 47.5 Silicon Valley Surplus 29.5 silicone oils 30.4 silver-plated particles 29.4 SIMM 25.4, 31.1 Simpson 47.3 sinewave generator 36.3 Singing Wind 31.5 sinkhole 37.5 sintering 36.2 skin effect 48.1 sledgehammer cure 35.2 sleeping bag zipper 39.4 slip 38.3 slip ring brushes 44.3 slip rings 38.1, 44.3 Slo-Syn 44.4 slop 25.6 slot 48.1 slow blow buse 37.1 SLS 40.4 Small Motor handbook 38.4 Small Parts 28.1, 39.4 Small Power Production Manual 41.4 smoothing fuzzy data 45.3 SMPTE time codes 46.1, 46.2 SMT grabbers 32.5 snap action 29.4 snap around ammeter 47.3 snap-together 40.6 Snow Belt Solar 28.4 Society Amateur Radio Astronomers
32.5
xx
46.1
Soft Solutions 42.5 solar collectors 33.4, 37.2 solar electric power plants 46.1 Solar Electric Power Systems 44.5 Solar Energy 28.4 solar energy breakthrough 46.1 Solar Energy Research Institute 28.4 solar panels 46.1 solar power 43.4 solar pump 28.3, 28.4, 44.2 Solar Radio Astronomy 32.5 Solarex 28.4 SolarJack 28.4 soldar bead 48.1 Solec 28.4 solid state cooler 35.3 solid state cooling device 25.1 solid state switch 31.3 solidified smoke 30.5 Solvonics 28.4 Sony 32.5, 40.6 source impedance 29.2 Southwest Windpower 28.4 spark 38.2 sparse data set 36.3, 42.3 spatial frequencies 32.4 Special Education crafts class 40.5 Specification Document 31.2 Specs in Secs 27.5 Spectron 30.4 spectrum analyzer 25.6 Speech Processor data book 29.5 speech synthesis 34.4 speed 38.2 speed reduction 38.3 speed sensors 38.4 speed-torque characteristics 38.1 Speleonics 48.5 spelling check 32.1 SPICE 30.5 spike protection 44.1 sports radar 40.5 Sprague 25.6, 29.4, 44.4 square pixels 33.1 squirrel cage 38.3 stability 38.1 stainless steel potato chip 39.4 standing waves 32.4
switched capacitor filters 34.3 swithcing bay 36.3 symmetric 32.1 symphony orchestra 33.3 sync separator 27.1 synchronicity 47.1 synchronous AC motor 38.2 synchronous counter 35.1 synchronous demodulation 39.2 synchronous demodulator 27.1 synchronous demodulator waveforms
27.2
SYNERGETICS 35.1 synthesis 42.3 synthesizers 32.3 system monitor 25.4 SX cartridge refill 40.3 SX engine 30.1
surface wave 34.3 Surplus Center 31.5 surplus dealers 37.5 surplus flyer 30.5 Surplus News 34.5 Surplus Traders 37.5, 40.5, 42.5, 45.2,
Hardware Hacker III television 34.4 television gear 48.2 Teltone 39.5 temperature compensate 41.3 temperature compensated varactor
41.2
temperature differential 25.1 temperature drop 25.2 temperature gradient 39.4 Temperature handbook 29.5 ten lamp display 37.4 termination 32.4 Tesla, Nikola 38.3, 45.1 Tesla Book Company 33.1, 45.1 Tesla coil 45.1 Tesla coil resources 45.4 Tesla Coil Builders Assn 33.5, 45.2 Tesla Society 48.1 Texas Instruments 25.6, 27.5, 30.5,
31.3, 32.5, 34.4, 34.5, 36.4, 38.2, 38.6, 47.5 text compression 32.1 thermal drop 25.2 thermal energy 25.1 thermoelectric coolers 25.1, 39.1 thermoelectric modules 25.1 thermometer 34.5
title 47 45.1 tone 31.4 tone generators 28.2 toner 26.3, 40.1, 40.2 toner cartridge reloading 40.1 toner recharging 40.1 toner recycling 40.3 Toni Patti 32.5 toolboxes 25.4 tooth 26.3 toothpick 44.3 Toroid Corporation 47.4 toroidal current transformer 47.3 torque 25.1 torque sensing 27.3 Torx bit 40.2 Toshiba 26.4 total lifestyle basis 37.1 total teardown 40.2
touch control blender 31.3, 31.4
tv pattern generator 42.4 TV Typewriter 39.1 20 marks 43.2 Type I Fonts 32.2
Thompson, Don 30.1, 40.1, 40.2 Thompson Nyliner 26.4 3-D graphics 32.5 3-D Systems 36.2 3-M 29.3 three phase generator 35.1 three phase power control 35.2 three-phase wye 44.3 threshold effect 25.4, 31.1 Thumb Electronics 35.1 Thyristor 27.5 tiling effects 42.4 tilt angle sensors 36.4 tilt switches 29.5 tilts 30.4 Time Code Handbook 46.3 time delay 34.3 time division multiplex 47.4 time domain 38.5, 42.3 time standards 47.2 time varying 41,3 time-to-market 36.2 tin oxide 28.3 tinaja quest 25.4, 26.4, 29.4, 30.5,
31.5, 33.4, 34.5, 35.3, 36.3, 37.5, 38.4, 39.4, 46.5
Touch Memory 46.4 touch memory 46.4 touch plates 31.4 touch screen 39.1 touchtone 34.4 trade journals 26.3, 33.3, 37.1, 37.2 Trade Winners 44.5, 47.5 trailer hitch works 39.4 transform 42.1 transformations 25.5 transient protectors 44.2 TRANSIL 44.2 transition temperature 39.4 treble control 34.1 triac 31.1, 31.3, 48.1 trig functions 42.2 trilateral conspiracy 48.1 Trimble Navigation 48.5 Trimmer Primers 37.5 Triplett 47.3 Trovicel 29.3 TrueType 39.2 truth table 36.2 truth-in-advertising foods 39.4 TRW 34.4 TTL Cookbook 29.5, 34.5, 36.4, 38.6,
40.6, 44.5, 46.5
ultraviolet flame sensor 28.2 UMI 33.3, 42.1 underdamped 34.1 Unibit 40.2, 40.3 unintended currents 48.1 Unitrode 31.5, 34.5 universal ac-dc motor 31.3 universal motor 38.1 UNIX 39.2 unusual tools 37.5 upconverters 41.3 update 39.3 Upper Sonoran desert 33.3, 41.1 US Government Bookstore 45.1 US National Bureau of Standards
47.5
tunability 34.4 tune 41.2 tuning 41.1 tuning dial 34.1 tuning voltage 41.3 Tupperware 28.1 tutorials 41.4 tv frame 32.1, 48.3
xxii
vertical scale 42.2 vertical shift 42.2 VGA adaptor for MAC LC 48.1 VGA monitors 48.5 VHF 40.5 VHF channels 41.2 VHF prescalers 35.1 Vice Grip Unibit 40.3 video 32.4, 37.5, 41.5, 43.4 video camera 47.5 video compression 42.3 video compression chips 40.6 video display standards 33.1 video distribution amplifiers 32.4 video driver 32.3 video gain 32.5 video image compression 36.2 video inverter 27.1 video links 40.5 video modulation 48.3 video monitor 32.4 Video Products Guide 42.5 video time codes 46.2, 46.3, 47.2 video quality 32.1 videotape 25.6 vinyl clad materials 30.3 vinyl clamshell 30.3 virtual ground 27.1 virtual transform 42.1 VISA 30.1 visible lasers 33.5 visible laser diodes 33.5, 38.6, 39.1 visible light photopolymer 36.2 VITC 46.3 VLA site 32.5 VLF radio 47.4 vocational education 43.4 Voice Synthesis LSI 32.5, 42.5 voltage noise 29.2 voltage spike 44.1 Vortec 25.3, 35.3 vortex cooler 35.4, 37.5 vortex nozzle device 25.3 VW microbus 28.3
wavelet theory 39.1 wavelet tutorial 48.1 Wavelets 48.1 wavelets 38.5, 42.1 wax 36.2, 42.5 weasel words 28.1 weather fax reception 40.5 weather patterns 25.6 weather radiosonde transducers 26.4 Wedge Innovations 30.4 wedge shape 39.3 weigh scales 27.3 weird science 31.5 Well, The 28.4 wenge 30.3 whine 38.4 white write 40.1 whitewash job 28.1 JC Whitney 44.4 Whole Earth Review 28.4, 31.5 Whole Works 46.5 wholesale distributors 26.3 wholesale surplus sources 37.5 Wholesale Tool 39.5 Wicker, Guy 26.2 Widebus 30.5 wilderness areas 33.2 windmill rotors 28.3 window detector 39.3 Winston 33.4, 37.2 wiper 40.1 wiper pad 40.1 wood routers 27.5, 47.1 wooden end rail 30.3 word processor 26.3 write 30.2
Zomeworks 28.4
genie
EMERGOP.TXT 39.2 FASTFERN.PS 25.6 FUZZYFIT.PS 45.4 HSTOGRAM.PS 42.4 MYFAVOR.TXT 40.4 WAVELET.PAK 38.6
Xicor 38.4
X-rays 43.4
yellow 29.1
yellow book 28.1
yellow dye 29.1 Yellow Front 30.3 Yellowjacket Solar 28.4 Yen, Henry 47.2 Yuppus dementus 28.4