Controllogix PIDE problem

pauly · Jan 15, 2008

I am using the PIDE function in version 16 of RsLogix 5000 for a job which I am in the process of testing. The problem I am seeing is with the loop in auto and with larger changes of PV the output drops to zero and then ramps back up again. This is happening with a setpoint of 100% and the PV changing from say 30-32% i.e. a 2% change. With smaller changes there doesn't seem to be a problem. I also notice that the CVHALM is coming on.
Any body has any ideas I would be very grateful.

Peter Nachtwey · Jan 15, 2008

pauly said:
I am using the PIDE function in version 16 of RsLogix 5000 for a job which I am in the process of testing. The problem I am seeing is with the loop in auto and with larger changes of PV the output drops to zero and then ramps back up again. This is happening with a setpoint of 100% and the PV changing from say 30-32% i.e. a 2% change. With smaller changes there doesn't seem to be a problem. I also notice that the CVHALM is coming on.
Any body has any ideas I would be very grateful.

It would be best to see a trend. Other than that I would say you have found a flaw in the implementation. There is no reason why the control output should go to 0% when the SP is at 100%. So let me guess, you are using the velocity form of the PIDE and all the gains are multiplied by the error and not the PV.

pauly · Jan 15, 2008

I can't post a trend at the moment but when I go back to the office I will take some. Peter you are correct, I am using the velocity form of the PIDE. The cv doesn't always dive to zero but sometimes drops off by 10-20% then gradually picks back up. The application switches between the override mode and automatic, are there any initializition bits that need setting when doing this maybe? I am pulling my hair out here!!!

TConnolly · Jan 15, 2008

Just a guess:
Tag.CVEUMIN and Tag.CVEUMAX are backwards?

Peter Nachtwey · Jan 15, 2008

pauly, I bet the problem goes away if you configure the PIDE so the derivative and proportional terms are mulitplied by the PV only. The real problem is the derivative term acting on the error and not the PV. Get back to us if this fixes the problem. I will explain why then but for now you should know some forms of PID do not like step changes because the output swings into saturation due to the 'derivative kick'.

pauly · Jan 15, 2008

Off the top of youre head do you know which setting in the PID block changes this?

Peter Nachtwey · Jan 15, 2008

No, I don't. I don't use the PIDE.

I would have to look it up but the last time I read the documentation there were many more options and I am pretty sure the PIDE has a I-PD or PI-D mode. I am sure someone else that really uses the PIDE will help soon.

John M · Jan 16, 2008

How to set PIDE to calculate Derivative on PV

You just need to select the proper calculation in the General Configuration Box.
It is about halfway down on the left side Calculate Using and allows to choose E (Error) or PV (Process Variable) for both Proportional and Derivative Term.
I think Peter is asking you just to try changing the Derivative Term first.
I have attached a screen shot of what you should see.

I hope this helps.

Controllogix PIDE problem

pauly

Member

Peter Nachtwey

Member

pauly

Member

TConnolly

Lifetime Supporting Member

Peter Nachtwey

Member

pauly

Member

Peter Nachtwey

Member

John M

Member

Similar Topics