PID keeps overshooting

[Dgonza9]

Hi everyone,

I'm wondering if there's anything left to try to get this PID to work properly. I've autotuned many times, played with some of the values, and even autotuned with a mash tun full of grain to fully simulate the actual use of my system. I'm using a RIMS tube to recirculate to a square cooler I mash in. I make ten gallon batches.

After transfering the strike water to the cooler, I recirculate and dial in the PID. I usually wind up a few degrees lower than the mash temp and try to use the PID to bring temps up without overshooting. Typically, the PID overshoots several degrees. In fact, when it hits the set temperature, it doesn't seem to change what it is doing, but keeps firing at about the same pace it had been firing at when raising the temp.

Now, if I turn the PID off right when it gets to the set temp, then turn it back on, it holds the temperature perfectly. But if I just let it ride it will overshoot.

Another odd thing is that when I autotune it the PID seems to work fine right afterwards. If I try and raise the temp after autotuning to a higher set temperature without first turning the PID off, it doesn't overshoot. But when I come back on brew day and use it to hit my mash temp, it overshoots.

Any thoughts? I'm ready to throw this thing out the window.

Cheers.

Dan

 

[Dgonza9]

Been doing some more reading and I'm wondering if you have to wait until the temperature of the probe equalizes before turning on the PID.

When I first dough in and start recirculating it takes a few minutes for probe to heat up and the temperature to equalize. Is it possible that these rapid fluctuations confuse the PID and alter what it does later?

Cheers,

Dan

 

[alien]

It sounds like the auto tune is too aggressive. As you know, kettles heat up faster than they cool down. This means that the tuning needs to be highly "damped" to avoid overshooting.

There's probably no substitute for fiddling with the gain parameters manually in this case. I suspect all 3 parameters will have to come down.

 

[kevink]

Where exactly is your temp probe located?

 

[processhead]

Where exactly is your temp probe located?

+1

If your temp probe is in the wrong area of the process, you can get some overshoot because of the time lag.

I've had best results with the probe in the top 1-2 inches of the mash.

 

[alien]

If the problem is lag, then Dan should be seeing a whole lot of dead time after switching on but before seeing any rise in temperature. Is that the case?

 

[Dgonza9]

+1

If your temp probe is in the wrong area of the process, you can get some overshoot because of the time lag.

I've had best results with the probe in the top 1-2 inches of the mash.

Interesting. Mine is at the outlet of the RIMS Tube.

My main concern is that the PID doesn't seem to start backing off on firing when it reaches the set temperature. I would think it would slow down some, but based on the light that shows the element firing, mine keeps firing at the same rate.

But what's odd is that if I turn it off, then on again when it's at the set temp, it will hold it perfectly, generally within one tenth of a degree.

I think I'm going to try giving the mash temp some time to stabilize before using the PID. I may also set the derivative to zero and see what I can figure out. But thanks for the replies.

I'd be curious to try putting the probe in the mash. Although I would think that might create hot spots or confuse the PID. I just want to be able to recirculate water that's at the exact mash temp I'm using.

 

[processhead]

I don't see the overshoot being the result of the probe in the outlet of the RIMS tube.

I think I would focus on the PID settings. For most processes its not necessary that the derivative be set to anything other than zero.

 

[bdh]

Interesting. Mine is at the outlet of the RIMS Tube.

My main concern is that the PID doesn't seem to start backing off on firing when it reaches the set temperature. I would think it would slow down some, but based on the light that shows the element firing, mine keeps firing at the same rate.

But what's odd is that if I turn it off, then on again when it's at the set temp, it will hold it perfectly, generally within one tenth of a degree.

What's the value of the I parameter relative to P and D? It sounds like I is way larger than P or D so the integrated error that builds up while the system rises to the set temperature is overwhelming the control signal. Then, if you reset the PID the integrated error gets wiped out so the system behaves like it should (or at least the effect of I is negated since there isn't a large error signal for it to integrate).