pid tuning problem

[skysurfer]

I'm trying to use the PID.vi to control a heater. I initially set i=0 and d=0. I obtained a fairly stable temp. with some offset. When I give a value to the integral term, the control turns into an on/off type of control and the temperature really oscillates. Any suggestions?

[SteveChandler]

From your post it sounds like you tried only one value. What value or values did you try? You should make it a very small value. You might have some luck with autotuning.

[skysurfer]

Thanks for the reply Steve. I've tried integral values as small as 0.001. As soon as I enter any integral value, the output goes to max. until it reaches the setpoint and then goes to zero until it reaches the setpoint after the overshoot. It seems to me that there should be some output as it approaches the setpoint after the overshoot so that the temp. doesn't drop too far below the setpoint. More suggestions please.

[nathand · Accepted solution]

If you're using the PID toolkit, you probably need a LARGE value for the integral, because the value is actually the time in minutes before the integral output duplicates the proportional output.  The algorithm takes the proportional gain, divides that by the integral time, and multiplies that by the integrated error.

[skysurfer]

Thanks for the advice. Using a large value for i solved the problem.