Forward/Reverse Combination PID

[RBPLC]

Are max cooling & max heating about the same capacity, BTU-wise?   If not, and if you have trouble getting gains that work well in both heating and cooling, you might move the neutral point, so that it's roughly equal BTU per CV%.

For this application, capacities are roughly equal. I suspect splitting it down the middle will work fine for this application with same parameters for each side. My intent though is to make them independent by having different tuning parameters that are used when heat/cool, cool/heat transition occurs. 

[Controls Guy]

No. Output gets assigned at the top of the calcs to the pTerm + Bias and then offset by other terms as appropriate.

OK, so I think we're saying the same thing.   But if that's the case, then presetting Output to some value and zeroing Bias won't have any effect past the first cycle, correct?

[BobO]

OK, so I think we're saying the same thing.   But if that's the case, then presetting Output to some value and zeroing Bias won't have any effect past the first cycle, correct?

That should be true, yes.

[Controls Guy]

That should be true, yes.

Weird.    @RBPLC claimed that doing so kept the CV from going to 100% even long enough for the PID to wind back down.   Maybe it didn't actually do anything and it integrated fast enough that it appeared to be OK?

Also, does Bias freeze on a disabled loop?   So that if you start a loop that's been off, first-scan CV will be (P + saved Bias)?  Or does Bias clear if you disable the loop?

[BobO]

Weird.    @RBPLC claimed that doing so kept the CV from going to 100% even long enough for the PID to wind back down.   Maybe it didn't actually do anything and it integrated fast enough that it appeared to be OK?

Also, does Bias freeze on a disabled loop?   So that if you start a loop that's been off, first-scan CV will be (P + saved Bias)?  Or does Bias clear if you disable the loop?

Either loop in manual or sample time not yet expired should leave it untouched.

Bias does freeze when in manual and is bumped on manual-to-auto based on bumpless mode setting.

[RBPLC]

Maybe it didn't actually do anything and it integrated fast enough that it appeared to be OK?

It's not "integrating fast enough". Even with Reset set to zero, the behavior is the same. On transition to from manual to auto, Output=Output+(SP-PV)*Gain.

[BobO]

Maybe it didn't actually do anything and it integrated fast enough that it appeared to be OK?

It's not "integrating fast enough". Even with Reset set to zero, the behavior is the same. On transition to from manual to auto, Output=Output+(SP-PV)*Gain.

Bumpless mode pushes Output into Bias.

[Controls Guy]

Ah, so that's the one cycle of 'retention'