Does turning off the contactor cause the PID to adjust itself?

Homebrew Talk - Beer, Wine, Mead, & Cider Brewing Discussion Forum

Help Support Homebrew Talk - Beer, Wine, Mead, & Cider Brewing Discussion Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Robhansen

Hoghead Brewing
Joined
Sep 25, 2012
Messages
25
Reaction score
0
Location
Monroe City
Hi, sorry if this has been covered, I couldn't find it. I brewed with my new setup last weekend. It's a 3 vessel system with one 5500w element and one pump for now. I didn't put in a contactor in the element circuit, but am going to before I brew again. My question is if I just turn off the contactor and leave the PID temperature setting the same, will the PID re-adjust because the temp is not responding?
As I used the water from the HLT, I lowered the temp all the way down thinking that the element wouldn't fire. It continued to come on a few times until I switched to manual control at 0%. Is that normal?
Also will this contactor work? http://www.ebay.com/itm/DEFINITE-PU...455?pt=LH_DefaultDomain_0&hash=item43b698ca7f
Thanks
 
If the process value from your sensor is below your set point on the PID, then it will close your output relay (allowing voltage to pass thru it). It does not sense whether or not the element is actually turning on. Its only input is the temperature sensor. It will continue to monitor the sensor, and once the process value is close to, or exceeds the set point, then it will open the output relay (turning off your element). Hopefully that answers your first question.

If I understand your 2nd question, your saying that you reduced the set point on the PID below the temperature that is currently in the vessel, and the element was still running?

This is possible, and the simple answer would be have you performed the "auto-tune" function on your PID?

If not, do so, it will re-adjust the factory defaults for the P, I and D gains in the controller to what it thinks works best with your enviroment. Keep in mind a PID controller uses an algorithm to determine when to turn on an output. It looks at the "current" error (Proportional error), sum of error over time (Integral), and the slope of error over time (Derivative).
When you run the auto-tune, it will calibrate those terms to provide you with a consistent ramp up/small overshoot/little steady state error.

What you may of seen when you made a sudden change (lowering the set point), the PID took some time to adjust, and waited for the error to stabilize

Hopefully this helps.
 
Yes, I ran the auto-tune. I was concerned that if I just turned off the contactor, and did not run the temperature down, the PID would continue to "learn" even though the system was not heating.
 
I definitely could see that the output was on, but whenever I turned contactor back on I didn't have rampant overshot or anything. I don't think I really had the PID temp up though and the element off much? Maybe just for a few seconds during like "oh crap forgot to turn a valve"
 
It's a valid question... I don't know how quickly the PID learns when it's not in Auto-Tune mode... but if the PID is telling the element to fire, and the temperature is not going up (because you have the contactor turned off), the PID is going to start to learn that your setup has some serious issues getting up to temperature. So say the PID starts thinking that your system is just incredibly laggy and it adjusts its curves accordingly. Then you flip on the contactor, and all of a sudden the element actually fires and the temperature actually goes up. But the PID has 'learned' that it won't heat up correctly - will you see rampant overshoot until the PID re-learns that it actually can heat water?

A simple way to solve this may be to set your PID to manual mode when you have the contactor off for any period of time.
 
I'm not a controls expert and I haven't had much PID use experience either (basically disregard everything I say until someone confirms). The only time the P/I/D values get modified is during an auto tune function or manually changing them in the programming mode. If your element has been manually disabled, the PID will follow its program to try to reach setpoint but will not modify hard coded P/I/D values to do so. Once you allow power to return, it should be business as usual.
 
I'm not a controls expert and I haven't had much PID use experience either (basically disregard everything I say until someone confirms). The only time the P/I/D values get modified is during an auto tune function or manually changing them in the programming mode. If your element has been manually disabled, the PID will follow its program to try to reach setpoint but will not modify hard coded P/I/D values to do so. Once you allow power to return, it should be business as usual.

You are exactly correct.

The P, I & D terms are only changed by running the auto tune function, or by manually changing them. The PID doesn't change the way it reacts to a system by itself.

This is also the reason why you should re-run the auto tune mode anytime there is a change to the system (relocating the temperature sensor, changing a pump, changing vessel size or any orientations, etc..)

You can think of the controller as a mini computer, running a program (basically a 3 sets of equations which contribute to the overall stability of the system).

And I bet you all never thought that calculus you learned way back would contribute to making beer, but it does with a PID controller!
 
So Kal's comment is what set me to my thinking... http://www.theelectricbrewery.com/control-panel-part-2?page=11

He states the PID is in a learn mode for some period of time when it's brand new, which makes sense. If that's the case and you keep you contractor off for part of that initial learning cycle, it could learn bad values.

Or, are we saying that the PID never does any learning on its own without being in Auto Tune mode?
 
So Kal's comment is what set me to my thinking... http://www.theelectricbrewery.com/control-panel-part-2?page=11

He states the PID is in a learn mode for some period of time when it's brand new, which makes sense. If that's the case and you keep you contractor off for part of that initial learning cycle, it could learn bad values.

Or, are we saying that the PID never does any learning on its own without being in Auto Tune mode?

I'm not sure where Kal got that information. I've never seen my PID change values from those it derived during the auto tune process. There isn't anything in the Auber PID SYL-2352 information stating it will learn or adjust outside of the auto tune operation.
 
I shot off a question on this directly to Auber. I'll report back when I get an answer.
 
Thanks for all the input on this! I thought I'd seen that the PID continues to learn. We'll have to wait and find out what Auber says!
 
I've read the auber instructions 20 times and I never fully understood what they mean about fuzzy logic. It's really vague and to be honest, if you're going to design a unit that operates differently than the industry standard PID functions, you should explain it in more than one paragraph.
 
Actually, I'm using the 2362 controller. I didn't think it used "fuzzy logic". Anyway, I registered on Auber's forum to ask them. Waiting on them to complete registering. I do thank this board for all the knowledge I've gathered throughout my build!
 
I received an answer back from Auber, and subsequently called them to confirm. In summary, the parameters will not change unless you run an autotune or manually change them. I have posted my question and Auber's answer below.

------------------------------------

Here's a question regarding the SYL-2352 that came up on homebrewtalk.com,
and I suspect you can answer it definitively.

Does the PID "learn" at all times, or only when running autotune? I.e.,
will the parameters change just from running, or only through an autotune or
through changing them manually? Are there any difference here between your
PIDs, for those that are well-suited for brewing?

The background here is that for safety one may put a DPST mechanical relay
between the SSR and the heating element (and between the other 240v hot leg
and the element), so when switched off there is certainty that the element
is off. In that mode, if the PID is "learning," it is "learning" bad
information as the temperature will not rise in response to its signals.

Many thanks.
-------------------------------------------

Here's the answer I received through email:

---------------------------------------------
Hello Jeff,
The PID only learn when autotune. It will set the parameter during autotune only unless you manually change it. So, it will not learning bad by idling. It you turn on the output after controller has been idled for a while, the output may be very high for a short period of time. But the fuzzy logic algorithm should be able to make quick adjustment. If that dose not solve the problem, you can turn off the controller and start it again to reset all the output due to integral windup.
---------------------------------------------
 
Thanks for getting the research done on that! I had myself convinced that I would have to back the set point down every time I was done with a heater.
 
I think what Kal was alluding to was a sort of break in period. A brand new PID might autotune to different values than a PID that has been used heavily. This is not really a function of the PID but more of the system achieving repeatable state. Think about it. Your metal is all shiny and new. The wires are freshly soldered. There's never been any consistent amount of current running through them. Your equipment is not exactly the same on the first brew as it is after 10 brew days. It's probably all insignificant but the PID could observe the changes and end up with slightly different settings.

If you look at the work Kal has done, you cannot deny he is a stickler for details. This is another detail.
 
Thanks for getting the research done on that! I had myself convinced that I would have to back the set point down every time I was done with a heater.

No problem. Is there a simple way to set the PID not to heat other than backing the setpoint down? Manual mode at 0%? Any other method?

It's still probably not a great idea to have the PID trying to raise temp with the element off BEFORE you are really going to heat, given the integral windup. Perhaps it would overshoot, perhaps not. In any event, no parameters will change, and a reset cures all.
 
I think what Kal was alluding to was a sort of break in period. ... If you look at the work Kal has done, you cannot deny he is a stickler for details. This is another detail.

First, I will agree 100% that what Kal has done for e-Brewing is miles beyond what anyone could possibly expect. I know I wouldn't have even thought of doing my setup had it not been for his website, and I would have spent a LOT more time asking stupid questions over and over if I didn't have his site to go back to.

That being said, no man is above making mistakes. I think the below quote from his website is, in fact, wrong.

Kal's Website said:
The first few times you brew the Hot Liquor Tank temperature may rise higher or drop lower than the temperature you set. This is normal. The Hot Liquor Tank PID will automatically learn the heating and cooling characteristics of your setup and tune itself over time to provide more precise temperature control.

By reading that, I thought that if I just left my PID alone, it would learn the setup over time. I could force the learning by using Auto-Tune, but it wasn't a required step as the PID would work it all out. We now know that this is not true, per Auber.

No ding on Kal - like I said, he's the outright Godfather of electric brewing. One bit of misinformation out of the veritable tomes he has written... that's not a bad average.
 
No problem. Is there a simple way to set the PID not to heat other than backing the setpoint down? Manual mode at 0%? Any other method?

The only ways I know are as you described. You could also unplug the RTD from the panel, which would cause the PID to go into alarm and stop sending signals. Unfortunately, the 2352 sounds an alarm tone if that happens (the SWA-2451 does not), so if you don't have an on/off switch on your alarm then that rapidly becomes a terrible idea.
 
BadNews,

I have a single PID control panel that I just finished with one RTD going to my HLT and one to my BK, I plan on unplugging the RTD to the HLT and plugging in the RTD to the BK after I finish mash-in. I've tested this with a simple walk-through with water in my vessels but ran out of wire and never connected my alarm (my next to-do item) -I don't have a switch to turn off my alarm -is there any configuration setting that I can use in the 2352 to disable the alarm upon disconnecting an RTD? -I still want the alarm to go off when my BK reaches 205F, but I don't want it to go off when I swap RTDs.

Note: I do have an on/off switch for the PID controller itself. If I could just turn off the PID, then swap the RTDs and then turn it back on that is a workaround that would work for me, but I'd prefer to disable the "disconnected RTD" alarm.


Adam
 
Apologize for the thread hijack; Badnews's post just instantly concerned me while I'm only 7 days away from brewing my first batch on this system.

Adam
 
I'm still baffled by the whole "fuzzy logic" terminology. It sounds to me like another way of describing the inherent function of PID. In other words, what makes PID smarter than a simple on/off controller that happens to have a hysteresis setting. When set point is a long way away, we expect the PID to run the element balls out until it gets close, then ease back on the power cycles as it gets closer and closer. . Autotune essentially tests a few different cycle times out to see how fast the system responds and then sets the PID values appropriately.

Auber seems to suggest this unit is different than typical PID function but I'm not smart enough to understand how.
 
I'm still baffled by the whole "fuzzy logic" terminology. It sounds to me like another way of describing the inherent function of PID. In other words, what makes PID smarter than a simple on/off controller that happens to have a hysteresis setting. When set point is a long way away, we expect the PID to run the element balls out until it gets close, then ease back on the power cycles as it gets closer and closer. . Autotune essentially tests a few different cycle times out to see how fast the system responds and then sets the PID values appropriately.

Auber seems to suggest this unit is different than typical PID function but I'm not smart enough to understand how.

Bobby, I'm not smart enough either, but I did long ago understand fuzzy logic. It is really just a different method for expressing an algorithm, in an analog fashion in more or less natural language, rather than a digital one in equations. The advantage for a continuous control system should be that is easier for human beings to implement and optimize the algorithm, thus more likely (or less costly to get the algorithm there) to result in faster convergence to the setpoint and a tighter error band. I'm pretty sure that there is always an equivalent "traditional" algorithm to the fuzzy one, and vice versa.

While grossly oversimplified, the traditional algorithm would know that error = -10 def F, while the fuzzy one might have defined a function for "negative error" where anything greater than or equal to the setpoint would yield 0, anything less than 100 deg F below the setpoint would yield 1, and everything in between would yield something in between 0 and 1. Then the fuzzy algorithm could be specified with statements like "If error is negative then do x." It really just provides a layer of abstraction allowing for specification more "in words" than "in numbers."
 
I really think we're over thinking this. Auber makes no mention of a "break in period" which is what this sounds like you're talking about. You autotune the device and so long as during the auto tuning you're running your pumps at the same flow rate as you would when you're pumping wort, the PID parameters will be set pretty darn close. The PID learns how much energy your system can put in and how much it losses. If you change that somehow then you need to re-tune.

Regardless, once set, it doesn't change those settings. I haven't seen my PID's overshoot their set temperature by more than 0.1*F.

I too am baffled a bit by the fuzzy logic statements. It tends to imply the values used are more of a possible range of values. Perhaps this is to offset some of the integral windup?
 
I really think we're over thinking this. Auber makes no mention of a "break in period" which is what this sounds like you're talking about. You autotune the device and so long as during the auto tuning you're running your pumps at the same flow rate as you would when you're pumping wort, the PID parameters will be set pretty darn close. The PID learns how much energy your system can put in and how much it losses. If you change that somehow then you need to re-tune.

Regardless, once set, it doesn't change those settings. I haven't seen my PID's overshoot their set temperature by more than 0.1*F.

I too am baffled a bit by the fuzzy logic statements. It tends to imply the values used are more of a possible range of values. Perhaps this is to offset some of the integral windup?

Agreed regarding the over-thinking, but at least we answered the original question. Parameters change on autotune, or when you manually change them. End of story.

Now, back to the over-thinking on fuzzy logic. Yes, Auber's answer to me certainly implied that one thing the fuzzy logic algorithm does is detect an acceleration of the rate of change and offset the integral windup. Not that one could not do the same with a non-fuzzy algorithm, lol.
 
By reading that, I thought that if I just left my PID alone, it would learn the setup over time. I could force the learning by using Auto-Tune, but it wasn't a required step as the PID would work it all out. We now know that this is not true, per Auber.
Interesting. I did indeed think that the PID would (over maybe half a dozen or so batches) "learn" the system without having to manually force auto-tune. Based on this thread it appears this is incorrect and that you should force an auto-tune the first time the setup is used and every time you make any changes to the setup.

I'll go change it on my website just to be clear that auto-tune should be run. It'll be updated in a few minutes...

No ding on Kal - like I said, he's the veritable Godfather of electric brewing. One bit of misinformation out of the veritable tomes he has written... that's not a bad average.

It still bugs me. I don't like mistakes.

Kal
 
It bothers me that I used "veritable" twice in the same sentence. I fixed it.

But now we're all smarter in the eBrew world!
 
Back
Top