kiln PWM parameter for PID

[coppercone2]

I don't know. Then I reset it and it gave me EE3 but it looks like everything is all good. It was flashing EE3 but the temperature was being measured and when I power cycled it, clearly the heater was OK


hmm maybe I should check the SSR for a short, but I have a feeling its a bogus error.


Is this possible because I have a very thin temporary thermocouple and its not like losing heat fast enough because the thick one that is like 16 awg wire will conduct heat out the wires and maybe vary the temperature a little bit?



Previously I did the auto tune at like 350C. It completed OK. The next day I set it to 850C and did it again and its going on about the errors

[Poroit]

If your Controller Auto Tuned at 350°C then that proves that your SSR is functioning OK.

EE3 is Loop Break Alarm so check your thermocouple.

Thermocouple.
How is your thin temporary thermocouple constructed?
Has it shorted somewhere behind the Hot Junction?

Controller Settings.
What Operating Range have you selected?
What Operating Mode have you selected?
Are you also using Setpoint2?
Output 1 Cycle Time?
What are the existing PID settings?

Sorry for all the controller questions but there are a few parameters that can screw up the controller operation if not set correctly

[coppercone2]

I need to print the manual because I don't like running this device in the house.


I replaced the thermocouple with the proper one.


The temporary thermocouple was very thin, probobly 28 awg single strand wire, which I folded over and put into a 22-18 crimp. It seemed OK but its possible that connection was bad. Now the thermocouple should be good since its properly terminated.

I can try to get this information,I just need to do one more disassembly because I wired the new one in backwards lol


I kind of like the Chromalox PID better, because they try to communicate better through the display... but it was too big. This one is 100% cryptic codes, so I need the manual near me

[johansen]

The temporary thermocouple was very thin, probobly 28 awg single strand wire, which I folded over and put into a 22-18 crimp. It seemed OK but its possible that connection was bad.

If you had provided any of that information previously otherwise i would have told you the thermocouple is non linear with temperature and that's why it ran out of time.

you can weld your own thermocouples but crimping them together doesn't work beyond maybe.. soldering iron temperatures, and even then the corrosion is going to set in and the wires go open circuit.

also the kiln ceramic insulation becomes slightly conductive the hotter it gets and you can have 60hz leakage affect the temperature.

[coppercone2]

no their welded at the factory I mean the crimps to the machine, it needs to be crimped to fit on a 250 spade. I just removed the yellow connector and put crimps on instead.

I know your not supposed to mess with the hot end. I have a spot welder for that part too

for the input of the thermocouple the C9000A PID literarly has 2 brass spades, the same as for AC power and the SSR.


It did pass the yank test. Anyway thats gone now


folding the wire into a over sized crimp on the hot end is mad noob haha. klingon wiring
https://static.wikia.nocookie.net/memoryalpha/images/e/e4/Opti-cable.jpg/revision/latest/scale-to-width-down/301?cb=20121204030820&path-prefix=en

[coppercone2]

Well I don't get it. I got a good thermocouple in there and it reads EE3 again when it gets to like 500 degrees on Autotune.


It still measures temperature. The heater still works.


What the hell does this thing want.


I read the section about the LBA.

Is it possible that if its set to 2x the integration time, and the integration time is mad wrong, it thinks there is a loop break when there is no loop break because its just too slow?


When it cools off tommorow I will try to get all the values it has.

[Poroit]

Have a read of Section 4.2 of the Omega Manual.

It will be a Time issue that is bringing up EE3.

Look at your setting for SP2 Operating Mode as it can also effect EE3.

Lets see your parameter settings when available.

I see a few of you posts where you are concerned about the spade terminals on the controller.
Mainly used a space savers....Not my preference but have used them successfully in applications.

The Cold Junction is where you introduce a dissimilar metal to the thermocouple so in your case it will be at the rear of the controller where you have the spade connectors.
The controller will look after the Cold Junction from there so don't be concerned.

"thermocouple non linear"...this is not an issue as controllers have built in Linearization Tables for all the Thermocouples, RTD's and Thermistors they offer.

[coppercone2]

Controller Settings.
What Operating Range have you selected?
-not sure what you mean. Its type K with a limit to 1150C (the instrument manual says 1200 but I gave it a bit of margin)

What Operating Mode have you selected?
-Autotune AT

Are you also using Setpoint2?
-No

Output 1 Cycle Time?
-0.3 seconds. I read that high cycle time can loosen the bolt on the SSR and that if possible use the fastest you can for a solid state application unless its a problem.

What are the existing PID settings?

Proportional : 2.0.  - This means 2 % I think
Derivative : 2 - this means 2 seconds I think
Approach Control : 2.5. - multiple of 2.5 proportional band
Integral : 1.1. - does it mean 11 minutes?


To get those settings I needed to disable Autotune because it just keeps trying to Auto tune ever time I boot the controller


Maybe I set the LBA for 30 minutes? If I understand it correctly it set for 22 minutes right now. maybe thats too fast.

[coppercone2]

I put the timeout at 30 minutes and now its ee5 ing me again


what the hell is this time limit. time limit is only mentioned in the document once under the definition of EE5.


Where are the auto tune time limits in the manual? I don't know what you mean


could it be unhappy because its empty? maybe I should run it with a crucible or something?

[coppercone2]

I let it warm up to 750C with out the auto thing, then I turned on the Push to tune (02) option, instead of full auto tune. it almost immediatly gave me EE5.


I guess this has something to do with the temperature change of the kiln being wayyyy too slow for the controller?


unless anyone has any insight, I guess I need to buy a thermocouple adapter for my graphing DMM so I can manually tune this sucker. I think I will anyway, because if I see EE5 again this kiln might get re-entry crash landing qualification

[Poroit]

How long did your kiln take to reach 750°C?

[coppercone2]

not sure like a few hours I think.

I bought a adapter from fluke so I can get a high temp probe in there to graph it on a fluke 289 because I am not going to write it down by hand like a graduate student. The probe I have right now that can connect to banana jack is like 300C max... so next week I will get a graph


A long time ago I saw my tax money being spent to get two kids to sit infront of a gieger counter manually logging it for 10+hours. I walked by in the morning for work they were there. I walked by at night after a lab lecture they were still there. I decided I will try to avoid getting into that pickle. And it was graph paper, with full scholarship physics grad students 



It was studying the decay of a particle accelerator target. I know those samples were very important to them.

I actually thought it was some kind of hazing ritual, the way they were sitting on stools in the middle of the room looking at the sample. I wanted to throw some paper balls at them. I may have even seen them at some other time, like a few days later.

In a sense its the ultimate "watching paint dry" experience. A number of particles evaporate into gamma steam.

My experience would serve as a excellent into to one of the Stargate episodes about a time loop.  "weren't you guys here.. doing the same thing.. way before?" Is it some kinda hologram ? Can this be real?


For all I know they are still there, reliving the same 5 minutes of time over and over and over again. Perhaps they made a naked singularity in there. I wonder what would happen if you threw a paper ball in there, it would probobly get stuck in mid air half way into the room.

[Poroit]

That is slow and would explain why you are seeing Timeouts.

I assume you have your Heating Output Limit(Output1) set at 100%.

Forget Auto Tune for your higher temperatures and set the controller as Proportional only.

Start with 2% Proportional as it sounds like this Kiln will limp to the Setpoint.

You can use Manual Reset to adjust any variation between the Steady State and the Setpoint.

[coppercone2]

Yeah thats what I am thinking since its a low power one for science. I still bought the adapter for my fluke grapher so I can try the manual procedure since someone already wrote it here.... why not


I can try to run it with a chunk of copper in there too to see how it changes.

Maybe this is so slow because its not exposed elements?

[Poroit]

I suggest that it is just a lack of power.

9A 120VAC =1080 watts.

Most higher temperature Kilns/Ovens start at around 4000 watts.

Good luck with the Manual Tuning.

[coppercone2]

I suspect it won’t matter unless you are worried about a really small over shoot but I wanna see

The chromalox did auto tune on my pcb cooker toaster, but that has poor insulation and bearly goes above soldering temp on max

[coppercone2]

Well I am trying to do the manual tune but its not easy. I realized the thermocouple hole is plugged, so I had to lay in a wire in there


There is also a big difference (100C) in measured from the unit TC and the Fluke 289.


I suppose to do it properly I would need to like monitor from the front panel. But I will see what it looks like anyway

[coppercone2]

@700C on/off
Frequency came out to be 145s
amplitude = 25.2Cpp

pulse time = < 7s
proportional band % = > 3.43%
dt = < 7.25s
∫ = > 145s
DAC = 1.5x default

times up, lets do this

[coppercone2]

Ok the little thermocouple that is a bit lower then the big thermocouple sees a small bump (15 degrees) and it drops down pretty flat with a slightly slower decay. I think it might be the DAC rate control, it fits the diagram the best from being too small.


However the big thermocouple I don't think over shoots, its linked to the wall. The small one seeing 15 degrees difference does not mean much since its like 28 gauge wire.


I am not sure how much utility there is trying to adjust the system with the small TC, when its painfully slow. I can try it again setting the DAC to 2x instead of 1.5x.


But with a substantial load, i think this might really be splitting hairs... i think the point was made pretty clear that this kiln is really damn weak and small, so for any of this to matter, you would need more watts.


I can still do a disturbance test, see what happens when I put in a empty crucible, because theoretically what I see can be explained by either derivative approach control being too small,or derivative time being too short.... because I don't think you can tell them apart without doing a load test. If I had to guess I would say its derivative approach control, because its the only setting that they expect you to guess on... the other ones are all based on calculations.

[coppercone2]

I monitored it and this time, with the setting set to 2.0x DAC,

On first setpoint it reached 712 degrees C

After a while it dropped to 699 degrees C and then rose up to 700 degrees C.


I jostled it by opening the door for a bit until it dropped to like 640 degrees C.

When it came back up, It got to around 703 degrees C and it dropped back down.


Since the magnitude of the error caused by the disturbance is 3 degrees C, and the turn-on over shoot is 12 degrees C, am I correct in thinking that the deritivie approach control setting is too small, and that the deritative time setting is also too short?


or should I leave the derivative time alone, try the approach increase one more time and call it good?


How accurate do you think I can get this? Right now the over shoot is ~1.8% and the disturbance compensation is 0.5% of the setpoint.

Increase derivative time from 7 to 10 seconds? My calculation came out to 7.25. I have it at 7 right now. 10 seems like alot and going against the rules, but there seem to be no real rules for the DAC control.

[dietert1]

When reading this thread i didn't find the motivation to use a PID controller.
In case the aim was getting a faster ramp-up be warned that successful autotune will result in oscillations during ramp-up, reaching temperatures much above setpoint, like +30 %. Exactly what you can't allow as it will destroy the furnace or the thermocouple.
I'd guess this control requires an extremely conservative approach with slow heat-up. The important term to observe is D.

Regards, Dieter

[coppercone2]

Well the formula I used for manual tuning in the manual resulted in 12C overshoot and 3 C overshoot after a disturbance.

I thought that the over shoot at start up and over shoot from a disturbance would be similar if it was purely a derivative issue.

Is it a mixture of DAC and derivative, since the initial over shoot is bigger then the later over shoot if you open the oven for a bit. they draw the wave forms like that in the manual that looks like if it was just derivative it would be same over shoot at start vs disturbance ?


Not sure if that is a good disturbance to open it. I thought maybe I should put an object in there, but I don't know the size I need for a disturbance.

[dietert1]

No, the initial overshoot is proportional to the initial temperature error Tset - Tambient. So 30 % of 750 °C would be 250 °C (still safe). The autotune features i have seen strive for a near critical adjustment for best steady state response. Better adjust the PID manually. As far as i remember the experts translate D as "damping" instead of "differentiator".

Regards, Dieter

[coppercone2]

The autotune fails on this, I did it manually using the manual and a fluke 289 with graph mode to figure out the wave form


The problem I have right now with the numbers I listed is that increasing the DAC control and increasing the proportional time do not seem to effect the over shoot, which is 12C. Like If I open the door for maybe 1 minute (turned off) and let it warm back up, it will consistantly over shoot with little effect from the following changes:


*****From********
@700C on/off
Frequency came out to be 145s
amplitude = 25.2Cpp

pulse time = 0.3s
proportional band % = 5%
dt =  7s
∫ =  3 minute
DAC = 2x default


*********To***************

derivitive time - 7S -> 10S
DAC Control -> 2 -> 2.5


Since that change had no effect on overshoot,

It makes me think the problem might be with the integral or the proportional band %. I could try a longer integral time or a wider proportional band.

I thought to Keep the derivative time @ 7 seconds (like the calculation from graph), the DAC control at 2.0, and then try to vary either the band % or the integral time.



Maybe for a small kiln I should just make the proportional band % wider since its smaller, then it would have more time to do everything? I am imagining this might be like landing a plane on a runway that is too short (aircraft carrier). I thought maybe to up that from 5% to 10%. I am imagining generally everything lands better if the runway is longer?

[coppercone2]

Ah I think I read the table wrong.


They give you a hysterisis % on the right and a band % on the left. I think I need to take that oscillation I found, which is 4.3%, round to the next higher percent, which is 5%, then find 5% on the right side of the table, and input the number to the left of it (10%) into the machine.


So I think my calculations were put in correctly, other then the proportional band, where I put in 5% instead of 10%. I think this basically put the controller in a impossible approach, so my plane keeps ending up in the swamp.