SYL-2352, still overshoot temps like crazy

[Squeeky]

Ok guys, I brewed again yesterday and still having issues with my PID.

Just to mention my system, I have a three keggle HERMS system, with my RTD probe placed on the HEX output. It is really weird because I'm able to bring my strike water up to temp, and it seems to use it's fuzzy logic and stop heating when I reach the desired temp.

I will then dough-in and set my PID to 154, running the mash water through the HEX and back into Mash. After awhile I look down at my PID to discover it reading temps of 160+, and still firing the element. I can not understand this!

I have emailed Auberin to ensure my settings were correct and they agreed, I also confirmed my enclosure was not the issue. They also did not think that could be the problem because it was not getting hot to the touch.

If anyone has any input please let me know! Unfortunately my brewery is located an hour away so I never really get the chance to play around when I'm brewing. However I don't want to do another batch until I get this figured out.

Squeeky

 

[audger]

what values for P, I, D, overshoot compensation, and filtering are you using?

im using something like this (if i remember correctly):
P= 5 (%)
I= 15 to 25 (somewhere in there; in seconds)
D= 0 (off, this means that once it hits a setpoint, it just turns off the power. when it drops it turns it back on. no compensation)
SouF (overshoot suppression) = 0.1 or 0 (low or off, overshoot is unlikely in our usage because of how slowly water heats. if you are using the HLT for a HERMS though, you might want to raise this slightly if you notice overshooting )
ot (control period) = 2 for use with a SSR, set to 10 if using a contactor (this is just how quickly the PID is alowed to cycle the power on/off)
filt = 0 (off, because, again, the system responds slowly enough its not needed)

 

[JRems]

Are you reciculating the water in the hlt? Without recirculating or stirring every minute I had overshoot problems. But my sensor is in the hlt. I did a test run and had problems with it in the mlt output so I put it in the hlt. I just need the hlt 3 degrees higher than what I want my mash temp to be. It's very consistant and doesn't overshoot at all now. I know others have had luck with it setup like you have it, but I did not. Try switching the sensor to the hlt.

 

[Squeeky]

I can't check them at this time, but Auberin recommended I resort back to default settings (other then manual control) , which I did prior to yesterday's brew. According to the manual, I have the following settings:

P = 500 (1 - 9999%) Would that mean it's 20%
I = 1000 (0 - 9999) Not sure the seconds, as I don't know the what the range pertains too
D = 120 (0 - 2000) Should this be set to zero?
SuoF = (unable to locate default setting in manual)
ot = 2
Filt = 0

Thanks for the quick reply, maybe this will help me understand what I'm doing wrong.

 

[rsklhm10]

Are you reciculating the water in the hlt? Without recirculating or stirring every minute I had overshoot problems. But my sensor is in the hlt. I did a test run and had problems with it in the mlt output so I put it in the hlt. I just need the hlt 3 degrees higher than what I want my mash temp to be. It's very consistant and doesn't overshoot at all now. I know others have had luck with it setup like you have it, but I did not. Try switching the sensor to the hlt.

This was previously my problem. I had my HLT probe installed through the sight glass and if I did not constantly stir, I would get really bad hot and cold spots, making my temperature look way off.

As for settings, for some reason mine got way off a while back, but I never figured out why. I since corrected them pursuant the default numbers in the manual and have not had problems since.

 

[samc]

Did you use the auto-tune feature? Ideally with auto-tune if you were actually to run it with grain in the mash you would probably hit the right settings. I know in my case using the 2362 and AT with water the settings got pretty close and just needed a bit of tweaking with real world brewing. Also the system sensitivity changed a lot with slight changes in flow & recirculation settings so that is one area to monitor closely.

 

[Squeeky]

I did attempt the auto-tune with my setup. It ran for over 2 hrs and never finalized. I realize there may be cold spots, but as my probe is reading 160+ why would it still be firing? This is when I wish my stuff was at my house so I could play around with it.

 

[samc]

How about the Celsius or Fahrenheit setting? Double check to make sure you aren't set for 154C.

 

[audger]

Are you reciculating the water in the hlt?

shouldnt really matter, if the measured temp > set point, the relay should turn off... its possible to get the polarity of the relay control wires wrong, it is DC. though i think the relay would be constant-off in that case, not constant-on... double check though.

and unfortunately, atleast on my PID, the auto tune didnt find any settings better than the defaults (which are far from ideal). default/autotune might get you to somewhere workable, but understanding what P I and D do helps to get an optimal balance...



if you think about the "system" we are working with; how much energy it takes to cause 1 degree of change and how much energy we can put into the water per minute, there is not much overshoot or compensation that needs to happen.

whereas if you had a "system" like an oven, where heating the burner quickly raises the temperature a few hundred degrees/minute, and opening the door quickly drops the temperature just as quickly; compensation is beneficial. instead of waiting for the oven temperature to fall -200 degrees over 30 seconds, if you see it quickly falling in the first 5 seconds, you can activate the burner and prevent it from falling all the way down. same with the reverse.

in our "system", the water temperature climbs relatively slowly, less than 2 degrees per minute. so even if you turn off the heat right as the temperature reaches the setpoint, its not going to go much farther. if you have a large thermal mass between the temp sensor and the heater, however (like when you use a 5-10 gallon HLT as your HERMS), it will overshoot more. if that is a problem, you could turn on compensation to kill or throttle the power some amount of time before it hits the set point. this is why you can set some of the parameters of zero (off).

you also dont need to be careful if the temperature is hovering right around the set point. if its 1 degree off, you dont need to throttle the power down to 105; you can run it full blast for 30 whole seconds (or more) and still not overshoot....

http://en.wikipedia.org/wiki/PID_controller
will tell you a little more about what the settings do. ill get my actual PID settings tonight and post them if i remember.

 

[borny]

Have you checked your ssr? Maybe you have a bad ssr. This happened to me and the heating element was always energized...



Just a thought,
Dan

 

[P-J]

Squeeky,

The easiest fix for you would be to move the temp probe to the HLT instead of in the HEX output. Then as you recirculate the mash, also recirculate the HLT water (Or setup a stirrer in the HLT) to keep the water moving across the coil. You will only lose 1 or 2 degrees between the HLT and the mash with this method. Plus auto tune works very well with this setup.

We need to talk.
Ha!

P-J

 

[Squeeky]

Thanks for all the input guys!

I guess the biggest confusion is with the PID itself. I want to make this clear that the PID is reading the correct temp. However if it has a set point of say 154, the temp will read 160 and keep firing. Due to the little OUT led, I can see it is still firing with a set point of 154, even when the probe is tell it that it's at 161. It is also no constant so I can assume it is trying to compute something.

Next time I go to my parents in-laws I'll run some water test and even take pictures if I can't determine it by then.

 

[CidahMastah]

I did attempt the auto-tune with my setup. It ran for over 2 hrs and never finalized. I realize there may be cold spots, but as my probe is reading 160+ why would it still be firing? This is when I wish my stuff was at my house so I could play around with it.

the PID is reading the correct temp. However if it has a set point of say 154, the temp will read 160 and keep firing. Due to the little OUT led, I can see it is still firing with a set point of 154, even when the probe is tell it that it's at 161.

I had this exact same experience with my PID but it was resolved with autotuning. Don't mess with the factory settings (i didn't) and it now works (has been working correctly for for quite a while). Of course I made changes like:
t=2
sn=21
dp=0
etc,

But that was it. I never messed with the P, I, D, SouF, ot or Filt settings.

next time you can try to reautotune. I had this happen too and it was because I was not giving it enough time to autotune before it hit set value.

Start autotune as soon as you start heating. I did mine like 10 and 20 degrees in front and it would never resolve itself.

To Autotune:
1. Press "set" to get to "at" - use increment and decrement buttons to set value to "2". It will flash and automatically start autotune. To do this correctly you want to autotune with your typical strike water volume and you have to be over 10F away from that target (start the autotune early, it seems to work better IME).
2. Once it is done on autotune you need to go in and reset the a/m back to "1" so you can use auto and manual heating modes. Autotune pushes the A/M setting back to 2 (manual mode prohibited).

 

[P-J]

CidahMastah,

Thanks for that. Excellent advice.

Cheers!

P-J

 

[CidahMastah]

CidahMastah,

Thanks for that. Excellent advice.

Cheers!

P-J

Glad I could (safely) contribute to the eforums

Actually you helped me out with some of that stuff on my ebuild thread, then I sorted out some of the final details - so thank you as well!

 

[kal]

+1 to not changing the factory integral settings. Just do the auto-tuning.

If the PIDs not able to auto tune itself then it's because it's unable to control the heating which probably means something was wired incorrectly.

Kal

 

[adam01]

One interesting thing, reading the Auber manual. It says that the fuzzy logic works as long
as you aren't changing temperature by more than 1deg/minute.

I don't recall the OP's element size or the HLT size, but a 2kw element raises ~5gals
at 2deg/minute.

I'll be running auto-tune this week to see how it reacts, if it works at all.

 

[CidahMastah]

One interesting thing, reading the Auber manual. It says that the fuzzy logic works as long
as you aren't changing temperature by more than 1deg/minute.

I don't recall the OP's element size or the HLT size, but a 2kw element raises ~5gals
at 2deg/minute.

I'll be running auto-tune this week to see how it reacts, if it works at all.

When I autotuned mine (successfully), I filled my HLT with 15G of water and set the PID to 165F and hit autotune right at the start. Has been good ever since.

I will have to check the manual for the context you mention above. But I can say my HLT is 5500w for 15G and my well water might be in the 50F range (delta of 115F in ~20 minutes). It probably takes me ~20mins give or take to heat to strike and the autotune works.

 

[adam01]

Well, if auto-tune doesn't work, maybe CidahMastah can post his settings ...

 

[CidahMastah]

Well, if auto-tune doesn't work, maybe CidahMastah can post his settings ...

my settings are posted above. But in short, factory integral settings have not been adjusted. The only setting I changed were listed in the above post. I can double check if needed (I have a sheet at home that has all my settings noted on the manual for reference). Let me know.

 

[adam01]

I was thinking that after auto-tune, the P, I, D parameters from the auto-tune process are
now reflected in what used to be the factory values.

How long did it take auto-tune to run (mins or hours) ?

 

[CidahMastah]

I was thinking that after auto-tune, the P, I, D parameters from the auto-tune process are
now reflected in what used to be the factory values.

How long did it take auto-tune to run (mins or hours) ?

Ah gotcha. I would imagine any settings that change are pot size/strike volume specific.

I would say it probably only took 20-30 minutes (but I didn't time it). Basically it was done by the time my strike water reached its temp and the temp held accurately from then on out. When I did it with only 10-20F in front it kept flashing for over an hour, then I gave up on it. That is why I was recc that you start autotune right at the beginning of warming up your strike water.