ControlLogix PIDE With Time proportional Control

[macg 0]

I'm look for some suggestions for two problems. 

1. I'm working on an environmental chamber that I only need to control the temperature.  It's range is -50C to 200C.  Cooling is accomplished with CO2 on a single valve.  Heating is made up of 3 heat levels; high, med and low.  I'm using a ControlLogix processor with a PIDE and SRTP instructions.  I originally monitored the output percentage of the heating portion of the SRTP instruction to determine which level of heat to use.  However that confuses the gains of the PIDE when switching between the heating levels.  Any suggestions on how to determine which heating level and how to tune the PIDE?

2. A second issue with the PIDE instruction is that the temperature input is not very stable while operating.  A blast of CO2 into the chamber of course drops the temperature quickly which the sensor reflects however when the valve is off the temperature is warming back up.  So the D portion of the PID instruction is going crazy.  I've tried averaging to try to smooth it out and that wasn't helpful.

Ideas???  Quit my job and be a Walmart greeter?

[Michael Lloyd 85]

I'd get rid of the D first and run PI. When you typed "temperature input is not very stable while operating.  A blast of CO2 into the chamber of course drops the temperature quickly which the sensor" the lights in my office flickered "Don't USE DERIVATIVE" in morse code. Derivative is used in systems with a lot of lag time. Like minutes. Lots of minutes

Turn Derivative off and start with a Gain of 1. Maybe three minutes in the I term. See attached "rough guide".

Set the Equation type to Dependent so I units are min/repeat.

Walmart is a very dangerous place to shop. Idk... I think I'd rather tune a controller

 

Initial Settings For PID Controllers.pdf