Temperature Controller

[bassinbc]

I have a temperature controller that controlls relays and works ok. I would like this controller to not change temperature so fast. Maybe a PID controller could help? The controller just turns on and off a pair of relays. Any ideas on how to make this controller not be so spontaneous? Maybe have a delay so when the temperature is right on the limit the controller doesn't cycle on and off so fast? Any ideas?

[bassinbc]

I am using a cDAQ 9172 and temp module 9211 and relay module 9481 with labview 2009.

[vt92]

Yes, you need PID to modulate (PWM) the relays.  Feed the ouput of the PID to a PWM generator.  Your PID % output = duty cycle.

[jack47]

A PID controller would help.  I haven't really seen on/off control set up for a system with cooling and heating.  You might have to split up your temperature limits for the heating and cooling systems to give you a deadband where the controller is idle.  For example have it heat up until it is in the middle of the high and lo limit.  Don't turn the cooling on until it hits the high limit, but cool until it is in the middle.  The same for the heating, don't heat until it hits the low limit, but keep it heating until it is in the middle.

[bassinbc]

A little background. This system really only heats or cools. It doesn't do both at the same time, I just wanted to write on vi to handle heating/cooling. The way we would use this vi is that the unit under test is cooled down to say -50 deg F. with electrical valve on a nitrogen tank supply. The valve cuts on and off the nitrogen tank supply. For heating it is the same idea except the we use a heavy duty heat gun to force in hot air to get the uut up to 275 degrees. The relay in this case controls a larger relay that can handle the current of the larger heat gun. Anyway, we have some pid controllers we use for this, but I want to develop a labview program that will take there place. Anyway a PID controller should work for this, I will just have to spend the time to figure it out. Jack47, your deadband idea might work as well.

[jack47]

If you have commercial temperature controllers with RS-232 or TCP/IP, I would be tempted to use them and have the LV program setting setpoints.  Building a PID controller in LV will take time, though buying the toolkit would speed things up.  However, there is a lot of development in controllers.  Auto tuning, ramp rates, etc.  That and they don't lock-up.