Process controller as a thermostat?

Compressors need a timer that limits the on time and off time to a minimum (typically something like 2 minutes each). Most process controls don't have that feature, but you could look. That kind of timer supplied externally will also fool with the integral control- it's a form of dead time- so the controller would have to be significantly detuned for stability.

Best regards, Spehro Pefhany

That's what I was afraid of. These things have an 80 second maximum cycle limiter in their programming but these things are still to smart for their own good I guess. I really don't need them sneaking up on the set point.

Trash: Stuff you throw out. Junk: Stuff you put in the attic until you figure out what to do with it.

See if one of the available output modes is simple on-off with deadband. You for sure don't want time proportioning. However, if you're stuck with time proportioning PID, then set the proportional as low as possible (or gain as high as possible), integral and derivative to 0. This will make it act like an on-off controller, but without deadband.

As a matter of fact, this will even work with an analog output (0V-10V) if you can get a gain of about 100,000 to 1,000000. You just have to use the output to drive a relay.

Deadband is a lot nicer, because there's no chance of flipping back and forth right around setpoint. I'd guess a deadband of 5-10 deg. would be fine if available.

Pete Keillor

Well, I found a "Users Guide" but it reads like it was translated from Japanese to English by an Italian and proof read by a Polish physist.. Very hard to make heads or tails of it. I do see an on/off- proportional timecycle setting. Also see Proportional band/Gain or Hysteresis, Intergral time/Reset and Derivative time/Rate parameters but no indication of the choices.

Think I will hook one up to a peltier chip and play around with it. At least that way I will not be giving a $300 compressor a nervous breakdown. :-).

Hey! I know the work of that team. They wrote the manual for my Chinese lathe. Best bit of Chinglish I've read for years. Best part is that they keep referring to the operator as "the jockey" Oh, maybe that explains why the cabinet is so short!

Tom

You've answered your own question, Glenn. Compressors need minimum cycle times so PID control doesn't fit. Why use electronics dumbed down to function as a thermostat rather than just using a good thermostat?

Contact life in a marine situation perhaps? Omega, Watlow, Honeywell and other electronic controllers that accept thermocouple inputs do have bang-bang conrol modes. I don't know about Omron but I'd expect they do as well.

Hysteresis = deadband. The point about requiring minimum cycle times by the previous poster sounded like he knew what he was talking about. Might be able to rig that with Agastats or something, but it'd be getting too complex compared to just finding the right controller. Deadband set high enough would do similar, depending on heat capacity of fridge and cooling capacity, but not as positive.

Pete Keillor