Trouble With PID Controller

[brant2000]

I recently finished doing some mod's which included installing a 4-way valve to be used for return water temperature protection. The valve is a throttling plug valve, which operates from 0 to 100 percent recirculation using a 90 degree rotation. I have connected a proportional rotary damper actuator to this valve, which also provides about 90 degree rotation. I got and connected a very inexpensive PID controller (REX C-100) that provides 4-20 ma output signal, to set the position of the rotary valve actuator. Life seems good.

So I tested it out last night and everything seemed to be working correctly, until the temperature approached and then overshot the set point. I figured I may just need to adjust the setpoints, settings, cycle times, etc. to dial it in. After playing with it for hours, even when I had the controller set to a minimum setting of 0 C, it still wouldn't provide an output of less than about 12 ma. It wasn't even that the output was being limited, it seemed like the controller wanted to continue varying the output signal despite being approximately 70 deg C away from the set point. The controller also has alarm relays, which seemed to work fine.

Anybody have any experience with these or similar controllers?

 

[TCaldwell]

if you are sure the input and output scaling is correct, and properly displaying the process variable, setpoint and controller output value then I would check pid parameters. If you went with the default values, that could be the prob.
let me know what each value is and if it has a proportional band or gain, is integral seconds/repeat or repeats/sec, the derivative wont be required. the varying output is normal

 

[TCaldwell]

if scaling is correct and displays show what they should then try a pb=.4, I=100sec, leave ar at 100 it should respond quicker, have you tried the auto tune?

 

[brant2000]

I'm not really sure how to scale the input/output values. The controller allows you to select the type of TC, which I figured was the only way that you could adjust this scaling. The measured value (PV) was relatively consistent with an analog gauge. Might there also be another way to modify these scales? I would have just assumed that the device was always attempting to keep the displayed PV tracking the displayed SV, but that certainly didn't seem to be the case.

The next time I have a chance to build fire, I'll try adjusting to some of the values you recommend above. I tried many different options, including setting both the i and d values to 0. I also tried the auto tune, which also didn't seem to help.

 

[TCaldwell]

by setting the I term to 0, you effectively removed the strength of the output, leave the d at 0. The pb is the range or band around setpoint that the I works within. The i is the ammt of effort or output to bring the process value to setpoint.

 

[TCaldwell]

any luck, are you getting some controller response?

 

[brant2000]

I've given up on trying to use this controller. The operation I've seen from it is so erratic, that I'll have no faith in it working correctly. At the moment, I am operating it with full open/full close control using the alarm relay, and have ordered a Sestos controller to replace it. I've had great success and really like the capabilities of a current Sestos controller I have.