1. Can you use a PV+ to set up the PID?

Yes, there are some sample faceplates in FactoryTalk ME 5.0 and later or you
can build your own screens. You would setup a screen to access the internal
members of the PID tag, like [PID_tag]. SO, that controls the output in
software manual mode. The instruction reference, starting on page 504,
shows all of the members, the link to the reference is:
http://literature.rockwellautomation.com/idc/groups/literature/documents/rm/
1756-rm003_-en-p.pdf
2. Why does the PID instruction need an update time if a timer with the same
value is required to run the instruction?
The PID was not designed to regulate itself or even know how often it is run.
Thats the way it was made in the PLC-5. The original PLC designers decided
to do it that way and the ControlLogix ladder instructions are modeled after
the PLC-5 to remain consistent in operation between the two families. If the
PID is not regulated, then the integral contribution will not be consistent and
scan time dependent. Under the hood, the PID essentially just does a multiply
and add for the integration, and this has to be programmed with the right
time and controlled at a consistent rate.
3. How does the PID loop time work in regards to PLC scan time? Say the PLC
scan didn't land on 0.5 seconds?
Let me answer that by posing a situation. Lets say the timer is changed
from .5 seconds to 1 second without changing the PID. The portion due to the
integral term would change half as fast. So going back, if the scan time were
to hover between .48 and .52 seconds, then the integration behavior would
change just a little ((.48-.52)/.50 = 8%). In most cases, an 8% variance
probably wouldnt be noticed. In other words, it depends on how much off?.
Under the hood, the PID essentially just does a multiply and add for the
integration every time the rung is true.

4. Will the bumpless help if the set point changes during the process (drops 30
psi)?
I am guessing you are asking, if I change the Proportional gain to be based
on PV instead of ERROR in the PIDE, will this help my equipment when I
change the setpoint in the PIDE?. It depends on what you mean by help. If
you want faster reaction from the PIDE, then it will not help. If you want to
smooth out the PIDE output to take it easier on the equipment, then yes, it
will help. If I missed the question, then please provide more detail.
5. Would you discuss anti-windup for cascade control with PIDE?
I am not sure exactly what about the windup you want me to discuss, but I
will assume the question How do I set things up so that the slave tells the
master to stop integrating when the slave is not in cascade mode or at a
limit? This is shown on pages 21 and 22 in the PIDE whitepaper on how to
wire the loops to achieve this, link:
http://literature.rockwellautomation.com/idc/groups/literature/documents/wp/
logix-wp008_-en-p.pdf
If I missed the question, please add more detail.
6. Please explain when you'd use .swm or .man pros and cons of each (for
ladder PID).
With SWM, you can program the output in percent, and you do not need an
external tag. With .MO, you can define the range of the tieback tag and have
an external tag that might be easier to use with an HMI. Theres really no pro
or con; its just a difference of an internal tag or external one and you are
free to choose either.
7. When entering a value in "Deadband Value" is this value in % of CV or PV?
Or is it in Engineering Units? Also, is this the same setting that is labeled as
"CV Zero-Crossing Deadband" in the Function Block PIDE instruction? What is
"Zero-Crossing Off"?
In the PID, the Deadband Value uses the engineering units range. If No
Zero crossing for deadband is NOT checked, then the deadband wont kick
in until the PV reaches the setpoint. If it is checked, the deadband will not
wait for the PV to reach the setpoint but instead will kick in as soon as the PV
reaches the deadband. The deadband in the PIDE is the same with the option
to disable the zero crossing.
8. Does the PIDE require a loop update time?
Not if it is in periodic mode. Only if you change the mode from the default to
oversample mode does it need this. The whitepaper covers when it is a good

time to use oversample mode.

http://literature.rockwellautomation.com/idc/groups/literature/documents/wp/
logix-wp008_-en-p.pdf

9. Will you be covering tuning in more detail? e.g., What does a "1" mean in the
proportional gain? In other words, is there a method to calculate gains vs.
trial and error?
The topic of tuning is, well, there are literally college courses dedicated to
such things and it is a very broad and sometimes complex topic. I can tell
you about the sparkplug of a car and how it creates a fire, but thats not
going to teach you how to drive . . . but I will tell you about the sparkplug, so
to speak. The PID stands for Proportional, Integral, and Derivative. They
operate on a thing called Error which is the difference between the PV and
the setpoint. The Proportional has an output proportional to the error. The
Integral integrates the past error, and the Derivative is a prediction of future
error. Calculating and adding all of these things together is what the PID
produces. Setting the gains just tells how much of each to use. Having a 1 in
the proportional gain means that its contribution is 1 x ERROR to the output.
In this case, if the integral and derivative were set to 0, the PID output would
simply be equal to the Error value (the difference of SP PV) scaled to the CV
range.
As far as tuning goes, one popular method that is sometimes good enough is
Ziegler-Nichols method (you can find this with Google). This does not always
work though in all cases and may not give the optimal results. If you did not
want to do some study or playing with the gains, you could use tuning
software to get good results, or hire a tuning expert. Try searching Google for
PID tuning and you will find quite a bit of stuff on the subject. Simply due to
the wide variety of applications and the complexity of the subject, that is the
best answer I can offer.
10.Is there a manual mode on PIDE?
Yes, there is a program manual request input call ProgManualReq for program
mode and OperManualReq for operator request of manual mode. These use
CVProg and CVOper respectively.
11.Why would you put a timer on to execute the PID when the Loop Update Time
does this for us.
Unfortunately, the Loop Update Time in the ControlLogix PID does not
regulate the PID itself like the SLC PID can with its timed mode. The Loop
Update Time is only used as a multiplier for the integral and derivative
portions. If the PID is not regulated with a timer or periodic task, these
portions can get magnified or decreased as the scan time changes. The timer
is not necessary if the PID is put into a periodic task and the loop update is

set to the same value as the task period.

12.Is the simulation process valid for other systems
The point of the simulation was to show PID operation. The simulation was
just a simple filter instruction and would almost certainly not compare to any
real process as far as a tuning behavior goes. I.E, the tuning gains that would
work for this simulation are unlikely to work as well in any other application.
Each application typically has to be tuned for the best gains for that process.
However, if what you are really asking me is can you change the word water
flow to temperature or pressure then sure, they would generally all have
the same basic behavior that if the CV increases, then the PV increases.
13.Which is the main advantage of using PIDE over PID?
The PIDE offers more input fields. If you needed to have the PIDE controlled
by an operator station and the program at different times, the PIDE has this
built in. For the ladder PID you would need to write extra code to do this. The
PIDE is also bumpless when changing the proportional gain (PID does not).
The ladder PID timing needs to be managed, the PIDE has the periodic mode
that does not need to be managed. The PIDE includes a ratio mode. If you
dont use any of these features, then it doesnt matter which instruction you
use. If you do need these features, you might enjoy the PIDE more.
14.When would you use dependent vs. independent gains and vice versa?
The dependent/independent setting changes the equation the PID uses. If
you gave me one set of gains for dependent, I could calculate an equivalent
set of gains for the independent setting that will give the same PID output. It
comes down to a matter of what you are familiar with tuning wise. The
equations are shown on page 514 of the instruction manual.
http://literature.rockwellautomation.com/idc/groups/literature/documents/rm/
1756-rm003_-en-p.pdf
If you are not familiar with tuning either one, then you can learn by playing
around, or search the internet for PID tuning, or sign up for course that
teaches that subject.
15.Could you briefly explain Tieback again please?
Tieback tag holds the value that the PID feeds to the output when in manual
mode, i.e., you can directly control the output.
16.Would you please explain the Deadband function in the PIDE. Is it similar to
the PID?
Yes, the Deadband in the PIDE freezes the output just like in the PID when the

PV is close enough.
17.Is it valid to use PIDE oversample mode, with a timer controlling the time
trigger, with a PIDE in a periodic task?
Yes, the oversample mode in the PIDE makes the timing behave like the
ladder PID and it needs to be regulated the same, by a timer to Its enable bit,
or by being in a task that scans as fast as oversampleDT.
The PID whitepaper page 13 discussed when it is usually used.
http://literature.rockwellautomation.com/idc/groups/literature/documents/wp/
logix-wp008_-en-p.pdf
18.When converting a PLC-5 PID instruction, would you choose PID or PIDE?
I generally would choose PID, as it is almost a twin of the PCL-5 PID, but
thats not to say you couldnt use the PIDE.
19.Can you explain the autotune feature briefly. Does the Ladder have
autotune?
Ladder does not have an autotune. The PIDE has a built in Autotuner with
some RSLogix 5000 packages that bumps the PIDE output in manual mode,
and watches the system response. It then generates gains based on the
response. The Autotuner getting results guide has a little more detail, link:
http://literature.rockwellautomation.com/idc/groups/literature/documents/gr/p
ide-gr001_-en-p.pdf
20.We use PIDEs in AOIs, when I open the PIDE within the AOI in runtime, there
are question marks in most of the configuration fields. Has this behavior
changes with versions of RSLogix 5000?
I put a PIDE into in AOI in V19 and I dont see any question marks. I think I do
vaguely remember this in earlier versions. However, in either case, to see
the current values actually being used, there is a drop down box at the top
called data context. If you choose a tag, then the question marks should
disappear and show you the currently used values. If you dont get very far,
let me know.
21.What if you have ladder PID and function block PIDE running in program?
I am not sure if I understand the question. Let me know if I get it wrong, but I
think you are asking if you can use both in the same project. Yes, each
instruction is independent. If you put in several PIDE and several PID
instructions, the controller has no issue with this.
22.Will Rockwell Automation be going away from PID to PIDE, or keep both?

There are no plans to change anything at this time or the foreseeable future,
both are available.
23.On the PIDE, the deadband zero crossing does not stop the error term from
winding up. is there a way without adding a manual disable on deadbands to
the PIDE to stop it calculating?
I am not sure if I understand the question. The definition of the error value is
PV-SP or SP-PV. There is no deadband on this calculation. The deadband
applies to the change of the CV while the error is under the deadband. The
error is continually updated to know when the PV leaves the deadband.
24.Should the PID ladder be used in an STI?
Only the PLC-5 had an STI. However, the ControlLogix does have periodic
tasks. Sure, the PID can be used in a periodic task. If the task rate is
matched to the loop update time, then the timer instruction is not needed
either, which simplifies code a little bit.
25.How do you implement tracking of the PIDE output when not in control so
that its output tracks the active PIDE controller such as with a constraint
control application?
I am not absolutely sure of what you are asking, but I think you are asking
about a low select application. The PIDE whitepaper has an example of two
PIDEs, one active, the other one tracking. See page 8. I think the page is
missing the management of the SetCVPrevious bits, but I can discuss that if
you like. Link to the guide is
http://literature.rockwellautomation.com/idc/groups/literature/documents/wp/
logix-wp008_-en-p.pdf
26.How did you add the function block again?
I did a right click on the sheet, selected ADD ELEMENT, then typed in PIDE
and clicked OK.
27.How can you determine what the smallest scan time that can be used?
Generally, you would want to run the PID only as fast as needed to control
the loop. To determine this time, you would make a disturbance to the
process and see how long it takes to mostly settle out, then try to be about
10 to 20 times faster for the PID.
28.Can I change the P-I-D set points based on a set point band, also based on
the PV angle of attack and second derivative alter the set points of the
same?
The PID and PIDE instructions do not touch the setpoint, well, except for PV
tracking. However, anything can be written in ladder to implement this as the

setpoint can be changed from the user application if desired.

Also, for nonlinear applications like paper winders, it is also common to do a
thing called gain scheduling. This means the program is written to change
the gains with different values depending on the current PV value. The PID
instruction allows the setpoint and gains to be changed on the fly by other
instructions.
29.Depending upon application, can we calculate proportional and integral gains
rather than a rough estimate?
Generally speaking, tuning software is a good way to go. However, there are
also methods that can suggest gains based on measured system responses
to an upset. One popular method is Ziegler-Nichols that can be used to
calculate gains. You can search Google for the term or for PID tuning.
Tuning methods arent always perfect, but sometimes they are good enough.
You could also play with the gains to try to hone in to good control.
30.If using a periodic task, do you need to use the timer to control the PID?
In that case it is not necessary to use a timer. Instead, make sure the loop
update time matches the task rate.
31.Can you describe the procedure to tune a slow loop vs. a fast responding
loop?
Tuning is generally beyond the scope of tech support as it is highly
application dependent. However, the slower the loop is, usually, the smaller
the gains are. Generally speaking, tuning software is a good way to go.
However, there are also methods that can suggest gains based on measured
system responses to an upset. One popular method is Ziegler-Nichols that
can be used to calculate gains. You can search Google for the term or for PID
tuning. Tuning methods arent always perfect, but sometimes they are good
enough. You could also play with the gains to try to hone in to good control.
32.How does controlling the execution time work, if the task scan time does not
match the configured oversample deltaT? Say 0.5 second PIDE execution, in
a task that runs every 0.17 seconds (not a multiple of 0.5).
Under the hood, the instruction just does a multiply and an add. So if this is
done more often, the integration will be faster than the equation in the
manual says it should be. I.E, in the case you state, the integral will be .5/.17
or about 3 times greater then what the stated equation says it should. In
other words, the integral, and derivative for that matter, contributions get
magnified.
33.Do we recommend using Control station PID tune with Rockwell RSLogix
5000?
Depends on what you mean by recommend. Do we recommend a tuner vs.
hand tuning? Well, thats up to you, but tuners usually save time and can

sometime offer improved tuning. We are discontinuing RSTune, but a third

party solution like Expertune or Control Station could certainly be used. Does
Rockwell recommend them? As far as saying Use a third party tuner if you
need or want to, then yes in that regard.
34.Where can I send further questions concerning autotuning?
Typically, you would just contact the Technical Support center via the normal
means to ask the questions. However, at this time you can send me an email
directly at bakyer@ra.rockwell.com. I am usually pretty busy, so give me
some time to respond.
35.Sometimes when setting up a PID, the output will "lock" at 100% and stay
that way until Id dig into the data structure and find a tag with and
erroneous value. Clearing that to zero fixes it but I always wonder what
caused that.
The cause is that one of the input values had been erroneous at some point
and then that value propagates to the PID internals. Starting with V19, the
PID instruction checks all input values to be valid before using them.
Technote 41810 documents the issue.
36.How do I achieve a bumpless change from Manual to Auto? I want the auto
output state to start controlling from the manual operating point. I try this
but the auto operating point always remembers its last value.
The PID instruction does do bumpless transitions to auto automatically. I
suspect in your case that rung of the PID is getting disabled. Even if the
manual bit gets set, the PID will not see it until scanned true. However, if the
rung has been disabled, it wont see it and wont know that the manual bit
has been set. The suggestion in this case is to change the program so that
whatever you have setting the CV tag, to have it write to the Tieback tag or
set output of the PID instead and to keep scanning the PID when put into
manual. You can verify what I am suggesting by looking at the status area in
the PID setup. The output%, if the rung is false, will not match what the CV is
being set too, but is instead the previous value that you say the PID snaps
back to. If I missed the mark, let me know.