Some Principles of Radiotelephony Part II — Plain Talk About A.M. Fundamentals BY BYRON GOODMAN, * WIDX © Part I appeared in the May issue. Al- though Part II is complete in itself, it is highly recommended that Part I (and ‘Technical Topics in this issuc) be read hefore the second installment. — ED. Modulator Power Apparently, one of the most confusing points ubout a.m. radiotelephony is the neod for modu- lator power. Since the Year 1 amateurs (and commercials) have been looking for a means for feeding a low-powered audio signal into a big transmitter and getting out a husky ‘phono signal. ‘They're still looking. We can think of only two general clusses of modulation systems, One would be where rf. power fed into the modulated stage is controlled, and the amount of rf. appearing in the output is dependent upon the instantaneous value of the modulating signal. This could be represented as in Fig. 7A, and a typical example is a diode- modulator circuit. ‘The modulated stage is not an amplifier of any kind —it is simply a point where rf. power from another source can be controlled by a modulating signal. ‘The other general class of modulation system would be one where @ modulated amplifier stage is involved. Here a relatively small rf. signal is amplified by the stage to deliver a large rf. power output. Any such device will, of course, have a d.c. power supply associated with it, since the increased power must, come from some- where, and the amplifier actually only transforms the d.e. power into r. It docsn’t do this job completely — some of the d.c. power is used up by the stage in the process. A high-efficiency amplifier might transform 75 per cent of the d.c. power into ré. power, while an inefficient one might deliver only 30 per cent. If the efficiency of the modulated stage is constant over a given range of d.c, input voltage, we ean modulate the stage by using the modulating signal to control the input voltage. This is illustrated in Fig. 7B. Uf, however, the efficiency (and input current) of the stage can be varied by the modulating signal, then we can modulate the output by using the audio signal to control the efficiency (and input power) of the stage. This is illustrated in Fig. 7C. ‘The first example of a modulator, shown in Fig. 7A, is seldom used except. at low power levels, and is of little more than academic interest in this discussion, The other two examples are the ones commonly used in transmitters. “+ Assistant Technical Raitor, QST. June 1954 Let’s examine the example of Fig. 7B more closely. Since we know (even without showing an actual circuit) that the modulated amplifier is taking d.c. power from the d.c. voltage source, we can represent the amplifier by a resistance capable of dissipating that power. Further, we have agroed that the efficiency of the amplifier is constant, so we also know that this resistance iss constant one (not changing in some way with applied voltage), or else the conirol of the Modulated Stage RE Power Madulated ‘input —»——| |» Pomer Output (A) * BIE ten se, iy Modulated Amplifier Eticlency, Moditated Aniplifier : |» FPomer Signal Outout Gia De. Power {Shire AF Hodatat + otis 3 Fig. 7—Three possible methods for modulating « radio signal, (A) ‘The modulating signal controls the passage of energy through the modulated stage, A. diode modu lator otago isan example of such « method, ‘modulating signal varies the power input to a constant-fliciency rf. amplifier. (6) "The modalating signal varies the eficiency and input current of a variable-efligieney r,f, amplifier. 13input voltage by the modulating sigual must also vary ina similar manner. While this might be done, it would be au unnecessarily eompli- cated system. Thus we have 2 constant resistance, a dic. source nnd aur u.c. source connected in series, as in Fig. 8A. We have omitted any mention of r.f. now because we know that the r.t. output voltage is some constant percentage (the efficiency) of R Emod th Fac (a) — vous Ema “across +E | Emin Time @) Wig, 8— Fig, 7B looks like (A) above to the de. power source and the modulating-signal source. ‘B) For 100 per cent modulation of the voltage ap- ied to Rin (A), the applied voltage will vary with time in this manner. Enrx and Enis represent the voltage swings above and below the steady Ese. the input voltage. Our problem is to find what power, if any, must be furnished by the audio souree, Med, for the maximum-permissible mod- ulation percentage of 100.1 ‘The d.c. source pute a voltage Hite ueross the resistor J, und the audio source, Brot, also de- velops its voltage across R. Maximum modula- tion will occur when the voltage across 2 is being swung up to a value equal to twice Bas aud buck down to zero, as in Fig. SB. ‘This is obvious, of course, and it can be secu that the maximum and minimum peak swings, Bmax and Brig, ure equal in value to Esc. "To sce what all this means in the way of power in B, let's first review what we mean by power. Tn a de. circuit it is, of course, simply 2A or Bt +R. I and # have steady’ values, so it's easy to know what numbers to use. In an a.c. circuit, it’s a little more complicated, because the current or voltage is not constant but is changing rapidly, as in any of the sinc-wave representations we have showa so far. In scientific circles, “power” is called the “time rate of doing work.” It might be consid- ered a spocial kind of average. ‘To illustrate this double-talk, lot's look at Fig. 9 In Fig. YA, a resistor fis connected to 2 voltage source, Ha. We know that the current, through the resistor is Hie + R, aud it will be constant with time unless we change Hus or 2. "Why tho msximum-permissible modulation percentage in (00 was explained in Part I~ Ep, 14 Fig. JB shows a similar ease with an alternating voltage source, Exc, und the problem is to find at peak value of By, will do the same work a8 Eae's doing. Now the only work that the eurrrent does in flowing through the resistor is to heat it, so here is our common factor. We recall that the power dissipated in the resistor for the d.c. enso, and hence the power supplied by the d.c. souree, is FR or BR. We can draw a curve for this as in Fig. 9C. When we try to do the same thing for the a.c. cireuit, we're stumped. At least we are until wo treat, each little period of time separately. Then it ien’t, too difficult, All we have to do is to take the value of eurrent at that instant, square it, und inultiply by the resistance. When we do this we get 2 new curve that looks like Fig. 9D. This turns out to look like @ sine wave of twice the frequency of the original. You could call this a graph of the “instantaneous power" (but be careful how you use that expression, “instan- taneous power”). What we want to find is how tho a.c, of Fig. 9B relates to the d.c. of Fig. 9A when they both have the same heating effect on R. It's obvious that if the frequency of the a.c. were very low, the heat changes in R could be deteetod if we were brave enough to touch the resistance, but any a.c. we would be working with would be of a frequoney high enough so that the heat changes during a eyele wouldn't be apparent. So at any practical frequency it must reach an “average” of some kind. ‘The mathematicians will tell you that the “aver- age” of a curve like this ean be obtained by taking the area under the curve (shown shaded) or a given time interval and dividing it by the ime. The answer is a singic value that, work- ing over the sume time interval, would give the same total area, But. the value of our power for the dic. case is just the same thing — it is » figure that, over a given time interval, xives the area under the curve (shaded portion of Fig. 90). You can tell by just looking at the two power ‘curves that the peal “instantaneous power” for the ac, case is a high valuc compared with the cc. ease. Tt all works out, if you dive into the mathe- matics of it, that an n.c, with a peak-to-peak swing of 2.828 amperes bas tho same heating effect as a ce. of 1.0 amperes. With a peak-to- peak value of 2.828, the peak value is half of this, or 144. If the peak value is 1.0, the dc. equivalent value is 0.707. This dic. equivalent value is called the “effective” or ‘rams (root-mean-squared”) value ~ it’s what an mmeter indicates for you, There ure devices known 2s ‘peak meters” that can indicate peak values for you— they're useful for measuring sc. that ia made up of more than a pure single frequeney, where the relationship betwoen peak value and r.m.s, value is not known us it is in this case. Now let's get back to that modulator-power requirement. Referring aguin to Fig. 8B, we can recall that in this 100-per-ceut-modulation ease, Fas = Eaux = Enix. With our new-found knowl- OST foredge about a.c., we see now that the effective value of Emoa is 0.707 of the peak value, and xince the peak value is Zmax (or Enix), the ef- fective value of Hoa = 0.707 Bmax. Without knowing about the a.c., we knew that the dc. source Hj, was delivering power to the load Jt that can be computed by Power supplied by Hae de. power supplied, Independent of this, the source Enea furnishes power computed by (0.707 E, Power supplied by Broa = Since Ene has the same value as Hye, we seo that, in 2 constant-elliciency modulation circuit, with 100 per cent sine-tave modulation, the modu- lator must furnish power eal to one-half the d.c. power supplied to the modulation eirouit. ‘There is one point you shouldn’t overlook. We said nothing about the efficiency of the mod- ulated stage except that it was coustant. It might be only 40 per cent or it might be as high as 75 per cent, but we would still need the same amount of modulator power for 100 per cent modulation. We need power to swing the voltage around on the modulated stage, and there is no way around it. Of course, if the modulation per- centage is less, we will require less power from the modulator. If the modulating signal is a complex one, like voice, in which the rms. value is lesa than 0.707 of the peak value, then we will require less modulating power. But the modulating-power source should be capable of & aio ©) times (B) An ac, gource and load R, with a plot of eurrent vs. time. ease (C) and for the a. Notice that each eyele in (B) gives tro eyeles in (D). plot of 7 os ime for the a w) June 1954 (> Fig. 9—(A) A dc. source and load K, with a plot of current rs. delivering the power necessary for 100 per cont modulation with a sine wave, By now you're itching to ask about the other modulation system, the one in which the efi- ciency is varied by the modulating signal (Fig. 70). This looks like pay-dirt territory, beeaus. conceivably the unnamed “ellicicney-control” circuit has a high effective resistance, and it, won't tale much modulator power to swing it all over somebody’s half acre. Quite true, Such cireuits ean be made in which the modulating voltage can be developed across’a high resistance and the power involved is low (control-grid modulation is an example). But you still pay a price. For a starting point, remetber that we must swing the efficiency up and down about mean level. It is obvious that the maximum eficieney that we can swing up to is the maxi- mum the tube is capable of (maybe 75 per cent, with your fingers crossed), So our meun level, or “operating” point, will he down to half of this, or 37 per cent, in the best possible ease. And because the efficiency ia low, und the tube it~ self is dissipating most of the input, power, the input to the stage must be lower than what could be run with higher efficiency. Otherwise the tube would overheat. The net effect is that the carrier output power is about one-fourth what would be obiained from the same tube and a constent-el= iency modulation system. So hack to the “constant-efficiency” modula~ tion systems. You probably have recognized by now that plate modulation falls under this head- ing. The plate-modulation case is clear-cut: The modulator swing applied to the plate circuit must have a peak swing equal to the d.c. plate voltage, and this means a modulator-power requirement, of one-half the de. plate power souree. The tube can run at its best eflicieney of as high as 70 or 75 per cent, and we can get out a husky and fully-modulated carrier, if we will supply the necessary (and perlaps ex- pensive) modulator power. How about screen modulation? ‘The sereen circuit doesn’t take much d.c. power, s0 the modulator power require- ments are low. Quite correct. But how do you go about getting that necessary “operating point,” about which the sereon voltage will be varied? If we make it the normal screen voltage for the tube used a an rf, amplifier, we're going to swing it up to twice this voltage on peaks, Two things can happen, The tube ean bum up Lecauso it’s being over- Joaded. Or the output can increase with- out hurting the tube, showing that we weren’t getting as much out in the first place as we could have got. Hoy,. then, cau we modulate this amplifier-via the screen grid? The oly’ way is first to find what the tube ean do aa a straight r.f. amplifier, and then cut ‘the sereen voltage back to about one- (Continued on page 118) Tone 1SMLL Te ee Pre-Wound Coils for the VEST POCKET BEAM as shown in QST, Moy, 1954 BELOW: Close-up — ABOVE: Installation TUNED—READY TO INSTALL Set, consists of three coils (reflector, di- rector and driven element) wound on $4” polystyrene bars complete with coil covers and end plates. Driven element coil in- cludes 5-turn link for 52-ohm coax line. All coils are wound and tuned by WOQFG and W@VZC, designers of the VEST POCKET BEAM. Coil assemblies are complete and ready to install, SET OF $ 95 3 COIL . ASSEMBLIES a CONVERT YOUR 10 METER BEAM Since the element lengths for the VP Beam are about the same as those for a 10-meter beam, it is a simple matter to convert {o 20-meters provided the inside diameter of the center section is 34”. Bring your 10-meter beam out of moth balls and start working on 20. COMPLETE VP BEAM KITS AVAILABLE Kits include pre-wound coils, elements, clamps, insulators, lement panels, hardware, 12 boom and mounting assembly. Wall details on request. BEAMS BEAMS VEST POCKET BEAM CO. 6635 Delmar Blvd., Room 324, St. Louis 5, Mo. 112 Radiotelephony (Continued from page 18) half. The plate current drops to one-half (ap- proximately) of its previous value. If then we think of the sercen cireuit as the # of Fig. 8A, we sce that we ean swing the screen from this value up to # seie pexk value or hack down to zero, for 100 per cent modulation. But here's the unhappy part. When we modulate the seroen, we have an eficiency-modulation system quite parallel to control-grid modulation, and. the output obtainable with e sereen-grid-modulated stage runs just about the same as it does for control-grid modulation of the same stage. Sereen- grid modulation may be a little easier to apply and adjust in some cases, but it is no end-of- rainbow pot-of-gold deal.’ Clamp-tube modul tion is simply a resistance-coupled version of screen-grid modulation. You may run into “cathode modulation,” which is a combination of grid and plate modula- tion. Here again you don’t get something for nothing -~ the more modulator power you can supply, the more nearly the system approaches plate modulation and maximum output for a given rf. amplifier tube. Peak Power Input Some time you're going to run into a sage who will mention that the “‘peak power input” to your 100 per cent plate-modulated ‘phone rig is 4 times the unmodulated d.c. input, and as proof he will point out that on positive peaks the voltage is doubled on the modulated sta; (this is the same as the peak of Emax in #' 8B.) He will show that the doubled voltage is equal to 2 Hy. and that therefore the power ia 2 Bae)? +R = 4 Ea? + R. Then he'll try to eon- fuse you by ssking where this extra power comes from since your modulator and d.c. supply to- gether only furnish a power equal to 134 Had + R. ‘Don’t let him enow you ~- you just remember back to those earlier paragraphs about “instan- tancous power” und ask him where the power is hiding on the negative peaks! "The power delivered by the a.c. source, Bue is an average figure as measured by any practical measuring equipment. In other words, the meter looks at the various “instantaneous powers” over a cycle, averages them (the values in Fig. 9D) und comes up with an answer, ‘The a.c. souree, Ey, is working hard some of the time and coasting some of the time to come up with the figure indieated by the meter —~ your heckler is picking one small fraction of the time and trying {to confuse you with it, Linear Amplifiers As a lust resort in your search for power, you may wonder about modulating a low-powered stage and then building up the power level in a following (linear) amplifier. If the modulated stage is really low-powered (5 or 10 watts ri. output), and the following linear amplifier has high sensitivity (requires little driving power), | you can sometimes gain from this system in | (Continued on page 114)