PID Tuning Temperature

I generally would avoid the ZN tuning method - it relies on pushing a loop to close to instability for my comfort

If you can I would perform a step test - if you need some guidance please ask

raffles said:

I generally would avoid the ZN tuning method - it relies on pushing a loop to close to instability for my comfort

If you can I would perform a step test - if you need some guidance please ask

Click to expand...

The closed loop tuning just needs an oscillating pv correct? Very much what he already has right? There is no real reason that the control has to be driven more unstable than what we already see is there? If he can set the Iterm to zero and pump up the Pterm until a similar oscillation is seen, isn't that enough?

g.mccormick said:

The closed loop tuning just needs an oscillating pv correct?

Click to expand...

No, the oscillation should be on the verge of going unstable.

Very much what he already has right?

Click to expand...

I doubt it

There is no real reason that the control has to be driven more unstable than what we already see is there?

Click to expand...

It depends, is the system on the verge of going unstable?

ZN is awful, it is an attempt to tune a system without understanding it. I posted a link to the Controlguru site that shows the steps on how to tune a heat exchanger using the step method and the techniques to determine the open loop gain, time constant and dead time. This is the way to do it if you don't have auto tuning software.

ZN results in a system that over shoots, under damped. What really gets me is that people try to apply ZN to anything and it just won't work except for simple temperature systems.

set the P and the I both to 1, if still oscilating switch off the I
next step is to put the I on 0.1 and slowly put it up until starting to oscilate.
use half of this point.

I would not advise doing that

P band of 1 = P gain of 100 - quite a difference

I (time constant, seconds) of 1 is a fair amount of I action! Reducing it to 0.1 is makes it 10 times more!

Is the equation standard (ISA / dependent) or Parallel (independent)? If standard adjusting the P gain is adjusting the I as well.

Different manufacturers use different equations and units - this is important to know before adjusting anything. If the controller is so old and unknown then tests need to be carried out to determine these things.

If not you're just guessing and testing [change this, ah #$%^ it made this worse put it back]

Peter is being very polite about the usefulness tuning a system with ZN. It seems that it is what is taught in college and university and well versed in academics, real world results don't back up its notoriety.

Peter Nachtwey said:

No, the oscillation should be on the verge of going unstable.


I doubt it


It depends, is the system on the verge of going unstable?

ZN is awful, it is an attempt to tune a system without understanding it. I posted a link to the Controlguru site that shows the steps on how to tune a heat exchanger using the step method and the techniques to determine the open loop gain, time constant and dead time. This is the way to do it if you don't have auto tuning software.

ZN results in a system that over shoots, under damped. What really gets me is that people try to apply ZN to anything and it just won't work except for simple temperature systems.

Click to expand...

I'm sorry to have incorrectly stated bad information. Please listen to Peter and others advice.