Ab pid problems

Hello everybody:
I need somebody to help me in setting a PID.
The problem, it does not follow the equation, when is approaching to the set point.
The output should be proportional to the error, but even when it reach the setpoint, it will continue to 100% CV.
So let say it is slow to respond.
I am using to control an oven temperature(is for an experiment at the college), but I have more luck in controllin it by using a CPT instruction than using the PID.
Any help will be welcome, since I have tried, different set up, for , K, Reset and rate.
The only thing it seem to work is when I limit the Cv, to about 20%
Any help, will be greatly appreciated, since I am using a lot of my time.

Thanks

jerry123 said:

Hello everybody:
I need somebody to help me in setting a PID.
The problem, it does not follow the equation, when is approaching to the set point.
The output should be proportional to the error, but even when it reach the setpoint, it will continue to 100% CV.
So let say it is slow to respond.
I am using to control an oven temperature(is for an experiment at the college), but I have more luck in controllin it by using a CPT instruction than using the PID.
Any help will be welcome, since I have tried, different set up, for , K, Reset and rate.
The only thing it seem to work is when I limit the Cv, to about 20%
Any help, will be greatly appreciated, since I am using a lot of my time.

Thanks

Click to expand...

The output is only proportional to the error, if all your using is P. The integral and derivitive changes that proportion.

Turn off the I and D, and start from there, then add a little I, then a little D.

First, I suggest you take a look at the following Urls. Trust me, they will be very helpfull.


http://www.ronbeaufort.com/

http://www.ronbeaufort.com/sample_lessons.htm

You also have your PID setup incorrectly. The Time mode should be TIMED, not STI, unless you move your PID block into an STI interrupt routine.

And as the above poster said, with I and D values, your CV may very well not be 0% at setpoint.

robertmee said:

And as the above poster said, with I and D values, your CV may very well not be 0% at setpoint.

Click to expand...

The control output can't be 0 at the setpoint. The control output must be high enough to offset heat loss.

That raises the first question. Are the heaters capable of heating the system to the set point?

What is the time constant of the system? The integrator time constant should about the same as the time constant of the system or longer.

Peter, I don't think robertmee said he should run with no integral. He said because of the integral and derivative the output may not be 0 when at setpoint. I think this was a direct response to this from the original post:

Originally posted by jerry123:

The output should be proportional to the error, but even when it reach the setpoint, it will continue to 100% CV.

Click to expand...

Keith

Peter Nachtwey said:

The control output can't be 0 at the setpoint. The control output must be high enough to offset heat loss.

That raises the first question. Are the heaters capable of heating the system to the set point?

What is the time constant of the system? The integrator time constant should about the same as the time constant of the system or longer.

Click to expand...

Yes Peter, the oven is very small and the heater element, warm it up, from 20 C to 80 C in about 3 minutes.
So no problem with the heater.
Thanks

robertmee said:

You also have your PID setup incorrectly. The Time mode should be TIMED, not STI, unless you move your PID block into an STI interrupt routine.

And as the above poster said, with I and D values, your CV may very well not be 0% at setpoint.

Click to expand...

Hello Robert: I have tried, both, TIMED and STI.
STI seem to respond better than TIMED.
Thanks

jerry123 said:

Hello Robert: I have tried, both, TIMED and STI.
STI seem to respond better than TIMED.
Thanks

Click to expand...

Stop "Trying", start reading. Robert is 100% correct.

PID's on that processor set to STI NEED TO BE IN A TIMED INTERRUPT ROUTINE, Not floating randomly in any convenient ladder file.

There is no sense even trying to do anything to that loop until you get your update time's set up properly.

Greetings jerry123 ...

and welcome to the forum ...

beside the timing issues already pointed out by my distinguished colleagues, you have MAJOR problems with your input scaling ...

you didn’t tell us what hardware you’re using – but let’s assume (gosh I hate that word) that your input module produces a signal which ranges from 3277 to 16384 ... (that’s a very common range) ...

let’s also assume that the input signal lands at I:3.1 ... you’re moving that signal to N7:2 ... (there’s nothing really wrong with that idea) ...

BUT ...

then you’re using an SCP to scale the signal into a range of 24 to 100 as you send it to N7:0 as your PV signal ... (that is NOT correct) ...

the PID you’re working with will not give good control unless the input signal it sees (that’s the PV) has a full-scale range from 0 to 16383 ...

going further ... it looks like your scaling of the output signal (CV) might also be off – but we can’t tell for certain until we know more about the specific hardware that you’re using ... let’s just say that the range “6242 to 31208” is a LOT more common than the range “6400 to 31200” that you’re using ...

first fix your scaling issues – and the timing issues - and then see where you stand ... if you need more help, I STRONGLY recommend that you post your entire .RSS file (you may have to ZIP it first) ... that way we can tell a lot more about your system hardware ...

good luck with your project ...

PS ... and thank you Sparhawk for the kind compliments ...

Thanks Ron, and all forum members

Ron: The key was to work with PLC units, instead of the scaled set point.Because the PLC units have a ratio of about 200:1, so the system performs, beatifully.
After I have everything wrap-up, I will post the program and the chart(temperatur/Time).
Thanks again:teacher: