PID controller help

[bryanmc]

I installed a PID controller on my presto pot yesterday and have a couple questions for folks that know about them.  I ran it through the auto tune procedure (twice) and have the following glitches.  On heating it will only get within 3 or 4 degrees of the set point before the temp starts to fall.  Once the temp start falling, it may drop 20 degrees or so before it starts to climb again.  For example, with a 335 degree set point, the temp will swing between 332 and 318 for the entire run.  I know the unit is capable of higher temps since I can set it to 356 and it will take the plastic up to to 350.

Any ideas what settings I need to look at and what type adjustments to make?  So far I'm pleased since I don't have to constantly mess with the thermostat (it has been removed) to keep the temp usable during an entire run, but it would be nice if this thing would keep the temp within a degree or two of the set point like it's capable of.

 

Here's a link to the manual.  They may as well have written it in Chinese...  My unit is the TA8, but it's the same manual.

http://blog.uvm.edu/cwcallah/files/2016/04/Mypin-TA4-manual1.pdf

 

Thanks

Edited April 19, 2017 by bryanmc

[fshng2]

I looked at the manual.

A tighter temp range is accomplished by adjusting the hysteresis of Out2/AL2.

Factory setting is 1 and normally about 20/30 seconds.

Setting Out2/AL2 for less than 1 will cause the relays to turn on and off more often and provide a tighter temp range.

 

 

Edited April 19, 2017 by fshng2

[bryanmc]

That makes sense, but I think it needs to be done to out1/al1.  It says out2/al2 is for cooling and isn't connected to anything in my application.  Do you think that could in some way affect it never reaching the set point?  From what I've read, it should vary ~2  degrees either side of the set point when cycling. 

[JimP]

Temperatures are some the hardest loops to tune because to the time lag.  You apply a heat source and some time later depending up on the capacity of the vessel, relation of the temp probe to the heat source, insulation, and and the amount of energy applied as heat. 

Is would be normal for the indicated temperature to cycle above and below set point and your tuning is to flatten the cycles as much as possible. 

If you can "AutoTune" with derivative  turned off I'd try that. Derivative might be taking you out in anticipation of reaching set point and is probably not needed in a small vessel.  Otherwise since you have already completed an auto tune simply turn it off and test run again.  Make yourself some detailed notes and record you settings along with response time and the amount of overshoot.  Only change one setting at a time, and keep notes.  No one can give you a valid number.  Way too many options.

You will probably see a different response with low level in the pot compared to a full pot. So it might be helpful to try and maintain a somewhat constant level in the pot and expect to see the loop to cycle again if you add a large amount of plastic at on time.

I have several years of industrial instrumentation experience but have been away from hands on for a while.  Good luck!

[mark poulson]

Can he use any other liquid to run his tests, so he's not ruining plastisol?

[bryanmc]

I wondered about that too Mark.  My guess is that all liquids have different properties as far as heat cool time but maybe someone who knows will clear that up. 

[SlowFISH]

43 minutes ago, bryanmc said:

I wondered about that too Mark.  My guess is that all liquids have different properties as far as heat cool time but maybe someone who knows will clear that up. 

 

I'm a beginner on these devices as well.... my boss is a tech/circuit board junkie and uses a PID controller for a small DIY furnace  (he make knife blades).  He can control his furnace within a 2-3 degrees fairly reliably - but that's because the heat source is fairly uniform and the furnace is a fixed size.   He explained to me I have a couple things "going wrong" with my setup - biggest issue being the plastsol level is constantly changing.  The pot can heat 2 cups faster than 1 gallon.... so the auto-tune will never really be able to compensate for this difference - all it can do it shut off as the temp rises and kick on when low.   If I tune it with 1 gallon - it will learn to turn on/off at a specific point to keep it steady - but this point will be different with 2 cups of plastic (catch the idea) If my plastic level stayed the same - then things would be easier to control.

 

Due to lack of competence in this area and him speaking way over my head - I've resulted to tuning the unit with about 4 cups of plastic (I usually make small batches) - this puts me in the "zone" for when I use more or have less.... and I live with the fluctuation to some degree - i see about 5-10 degrees +/- typically once I start pulling plastic from the pot. which not ideal - isn't the end of the world for me considering it's still 10x easier than trying to control a batch in a microwave at different levels or a pot with no controller.

 

Using my analytic side... I'd think covering the pot as much as possible could help (keep temp more uniform?)  I also used a long 7" thermocouple and have it sensing as close to the bottom as possible to insure I don't over cook - figuring I pull plastic from the bottom and if need be I can up the temp at the start of a batch if it isn't hot enough.

 

  J.

 

Edited April 19, 2017 by SlowFISH

[bryanmc]

20 minutes ago, SlowFISH said:

 

I also used a long 7" thermocouple and have it sensing as close to the bottom as possible to insure I don't over cook - figuring I pull plastic from the bottom and if need be I can up the temp at the start of a batch if it isn't hot enough.

 

I mounted the k sensor (thermocouple) so the tip sits in a divot about 1/3 way through the bottom in the middle of the pot and in contact with the bottom.  I used computer heatsink compound on the tip so I get a good thermal contact.  My theory is the plastic can't get much hotter than the bottom of the pot (heat source) nor cooler than it.  I tried to figure out a way to actually get it into the plastisol from the top but my stirrer is the full width of the bottom and going all the way through the pot would leave me trying to seal the space between the sensor tip and the threaded coupling.  Does yours come all the way up to the set point?

[JimP]

By all means test with any liquid. Ideally something as close to the normal material as possible in density and thermal conductivity.  But do a test, lets say using water, record the results and adjust settings and repeat test.

Better response or worse?

Also place the temperature sensing probe near the heat source so the probe has minimum time lag.  If you could agitate the liquid during heating would also be helpful.

Edited April 19, 2017 by JimP

[Baitjunkys]

Did you let it learn in its entirety? The light quit flashing? You will never see the same temp as the liquid being on the pot itself, Only close. On a presto I would not put it near the heat source IMO.  Are you using a ssr relay? These things will keep learning, I suggest you teach it 1/2 full, this way it is under a few degrees full, and over a few degrees at low level. Teach it with plastisol, your wasting your time with anything else. You can use the same gallon till it is brown if you need to wont hurt anything. You need to adjust the puf value as well, Against a candy thermometer etc. My best guess as i have did about 40 of these so far, is your true temp will be off +10 -10, Just the heat sink compound alone will alter temp.

[bryanmc]

Hi Leonard...

I was hoping you'd chime in.  Yes, I let it run the full auto tune (at least the light went off) it took around five minutes or so.  Yes, I'm using an SSR.   The sensor is in the middle of the pot, so it's as far from the heating element as possible with my stirrer configuration.  I thought the heatsink compound (just a thin coat on the tip) would fill in the irregularities and give better contact and more accurate readings.  If that's not the case I can clean it off in a minute or two. 

I'll continue with plastisol as you suggest.  The temp indication seemed close to correct but I didn't check it against my digital instant read thermometer.  What exactly is the puf value? 

Any ideas why I'm not getting to the set temp?  It went past it a couple times during the auto tune as expected) so I'm at a loss for why it isn't getting to it during normal operation 

[Baitjunkys]

the t couple should be fine, It will just alter the actual temp.

Puf Pvf  the chinglesh letters on there, it adjust the offset of actual and reading. 

 

Learn it again. These things keep learning all the time. You can manually adjust the P I and D settings, but you really need to know what you are doing, I dont generally fool with them. 

[SlowFISH]

5 hours ago, bryanmc said:

 

I mounted the k sensor (thermocouple) so the tip sits in a divot about 1/3 way through the bottom in the middle of the pot and in contact with the bottom.  I used computer heatsink compound on the tip so I get a good thermal contact.  My theory is the plastic can't get much hotter than the bottom of the pot (heat source) nor cooler than it.  I tried to figure out a way to actually get it into the plastisol from the top but my stirrer is the full width of the bottom and going all the way through the pot would leave me trying to seal the space between the sensor tip and the threaded coupling.  Does yours come all the way up to the set point?

I made my setup to use with an injector - so my stirrer isn't the full diameter of the pot as I want to be able to lower the injector into it... so in my case I drilled a hole in the alum plate my stirring motor sits on and push the thermocouple thru it and down into the pot - close to the bottom - the sensing tip is probably .125" (max) off the bottom... my theory was at that level I'm sensing the hottest plastic - and therefore (hopefully) not over cooking it which is my biggest concern.  I noticed it does a pretty good job keeping temp even if your not really blowing thru plastic too fast.  I did notice the more I used it the "better" it seemed to work.

 

 J.

[bryanmc]

Here's what I did and discovered today.    I turned off I and D with P set to 2.25 (that's where it was from the last auto tune I assume). The temp fluctuation with a 330 set point was 330.7 - 319.7 (turning off I and D allowed me to finally reach the set point ). I started lowering P until I ended up at 1.60 with a swing of 333.8 - 324. I tried adding I but even the smallest amount (.25) sent the upper temp into the 340s. I never turned D back on.  I just need to figure out how to have the controller turn the ssr full on sooner so it doesn't drop as far. I'm not sure if there's a way to do that but hopefully someone knows how.