AB or ISA on the PID PLC5

I have a basic question about PLC-5 PID,that is on the configure there is Equation AB(Independent) or ISA(Dependent).
When we use AB and when we should use ISA?
What is the meaning of standard loop tuning methods when we use ISA gains?
Thanks in advance

Some users find the ISA form easier to tune. If you think about the PID equation in terms of phasors, the P term is at 0 degrees, the I term is at -90 degrees, and the D term is at 90 degrees. The three vectors are summed to determine the output, and this sum has a magnitude and a phase angle associated with it. With the P gain factored out of the equation as in the ISA form, when you adjust this gain, it affects all three terms equally, which only affects the magnitude of the output vector, not its phase angle. Obviously, adjusting either the I or D gain will affect both magnitude and phase. Anytime you are affecting both magnitude and phase with only one change it makes it more difficult to determine the effect on overall stability.

As far as "standard loop tuning methods" go, you need to make sure you're calculating the gains for the right equation with your tuning rules. Many rules calculate for the ISA (dependent) form. Note that you can convert the gains between forms with the following equations:

Kp = Kc, Ki = Kc/Ti, Kd = Kc*Kd

Greetings Deer,



in general, both equations (if properly tuned) will give identical results ...



specifically, suppose that you properly tune an AB/Independent equation ... and then you graph and print out the response of a system ... then suppose that you calculate the proper tuning parameters for an ISA/Dependent equation and use that setup for the same system ... and then you graph and print out the response of the system ... now you should be able to lay the two printouts on top of each other and hold them up to the light ... the traces will be almost identical ...



some people consider the AB/Independent equation easier to tune - because a change in the Proportional setting will NOT affect the amount of Integral or Derivative action ... that’s why we call it the “Independent Gains” equation ... specifically, the Proportional, Integral, and Derivative actions may all be adjusted independently of each other ...



tuning the ISA/Dependent equation is usually harder for most people to get used to - because a change in the Proportional setting WILL affect the amount of Integral or Derivative action ... that’s why we call it the “Dependent Gains” equation ... specifically, the amount of the Integral and the Derivative actions depend on the setting of the Proportional action ... more specifically, an increase in the Proportional setting will increase the Proportional action - AND the Integral action - AND the Derivative action ...



quick analogy:



the AB/Independent equation is sort of like having the salt, the pepper, and the sugar all on the supper table in separate containers ... you can add a little bit of any one without affecting how much you add of the others ...



the ISA/Dependent equation is more like having the salt, the pepper, and the sugar all mixed together in only one container ... when you shake on more salt (Proportional) you automatically get more pepper (Integral) and more sugar (Derivative) ...



now ... even though it might be somewhat harder to tune, I usually recommend that students use the ISA/Dependent gains equation ...



major reason: the PLC-5 and ControlLogix platforms allow the AB/Independent gains equation ... BUT! ... the SLC-500 and MicroLogix platforms do not (specifically, they use ISA/Dependent only) ... therefore ... if a student gets familiar with the AB/Independent equation and builds up a “feel” for tuning this equation, then he will be at a disadvantage when forced to tune the more common ISA/Dependent equations for the smaller platforms ... specifically, he’ll have to develop a new “feel” for the settings ... this might not be a big thing - but why force yourself to relearn - and relearn - and relearn something each time you go from one platform to another? ...

What is the meaning of standard loop tuning methods when we use ISA gains?

Click to expand...

the term “standard loop tuning methods” refers to methods such as Ziegler-Nichols and other such “rule-of-thumb” systems for calculating settings for the PID tuning parameters ... since most of these were developed before the AB/Independent equation came into use, the numbers that they generate will not work properly - unless you do some math to convert them into values suitable for the AB/Independent equation ...



I’m pretty sure that the conversion formulas offered by rytko fail to consider the differences in units (minutes and seconds) between the settings for the two equations ... if you want to research this, you can find the conversion formulas listed on page 14-3 of this manual ... here are some examples that I think are correct:



if the settings for an ISA/Dependent system are:

Kc = 1.11 (Proportional setting)

Ti = 2.22 (Integral setting)

Td = 0.33 (Derivative setting)



then the settings for AB/Independent system would be approximately:

Kp = 1.11 (Proportional setting)

Ki = 0.008 (Integral setting) - note three decimal places

Kd = 21.98 (Derivative setting)



notice the big differences between the numbers for the Integral and the Derivative settings ... plus consider this:



with the ISA/Dependent equation, to INcrease the amount of Integral action, we DEcrease the Integral setting ... (use a smaller number) ...



with the AB/Independent equation, to INcrease the amount of Integral action, we INcrease the Integral setting ... (use a larger number) ...



since many PID systems are based on integer math (and not floating point) the AB/Independent equation’s “bigger-number-means-more-Integral-action” rule makes it capable of providing a much larger amount of Integral action than the ISA/Dependent equation ... in some applications this might be considered a big advantage of the AB/Independent equation over the older ISA/Dependent equation ...



I hope that this helps ...

Ron Beaufort said:

now ... even though it might be somewhat harder to tune, I usually recommend that students use the ISA/Dependent gains equation ...

major reason: the PLC-5 and ControlLogix platforms allow the AB/Independent gains equation ... BUT! ... the SLC-500 and MicroLogix platforms do not (specifically, they use ISA/Dependent only) ...

Click to expand...

one more reason: the ISA/Dependent gains mode is more readily understood by "seasoned" (not with salt & pepper) instrument techs since it mimics the terms used on older stand-alone analogue controllers (reset, rate).

Simulation for PID

Well, thanks all for your advice,
To learn more about PID, I think the best way is use the real world device for simulate function of PID.
I have a PLC5 training box from Rockwell,where in the box there are switches, analog input(POT) and analog output(Meter VDC).
PID need one analog input as PV and one analog output as CV.
CV can represent with DC voltmeter.
Problem that I face is I don't know how to simulate feed back or PV value.
As you know the PV is automatically run and not possible to use a POT that use manually.
After a few days, I get an idea to make a PID simulator but I don't know is it possible or not.
I have:
1) Water Bottled/Galon Dispencing System (usualy use in office,home,etc)
This dispencing have COLD and HOT valve.

2) Termocoulpe

3) Control valve

4) Water container/drum

5) PLC5 AI(IFE) and AO (OFE)

On the dispencing system I will modify the HOT and COLD valves and replace with control valves.
Is it make sense to running a PID simulator?
Again, thanks for your time

as for a simulator for PID ...

Greetings Deer,

read this thread and pay particular attention to the post by Peter Nachtwey ... I think that you'll find the ideas there to be helpful ...