DIY Heat Treat Oven Build

Status
Not open for further replies.
I just finished all the wiring, thank you all so much for this post!

If anyone has info on how to set the inkbird PID, I would appreciate it. Haven't tested at super hot yet, but 500F fluctuates about 30 degrees each way even with autotune.

Also, is the temperature setting supposed to max out at 1999F? I wanted to get up to 2100F for some special steel I found and I thought this was rated for 4000F...
 
Tarkus said:
I just finished all the wiring, thank you all so much for this post!

If anyone has info on how to set the inkbird PID, I would appreciate it. Haven't tested at super hot yet, but 500F fluctuates about 30 degrees each way even with autotune.

Also, is the temperature setting supposed to max out at 1999F? I wanted to get up to 2100F for some special steel I found and I thought this was rated for 4000F...
Click to expand...

If I’m not mistaken, there’s a parameter you’ll need to set for “max temp” that will allow you to raise the setpoint to 2100. I believe there’s also an “AT” button that you probably need to hold for “auto tuning”. This should tighten up your temp swings.

I have a PID similar to the inkbird, and you just let your oven reach your desired setpoint, then hold the AT button until the led turns on. It will “learn” for a few seconds, then the led turns off and the tuning should be set at that point.

edit: I just noticed you already tried auto tune. Are you waiting until setpoint is reached? You may have to auto tune a couple of times. If that doesn’t work, it could be a hardware problem, or in other words, your oven isn’t holding heat well enough.
 
Oh ok, that max temp suggestion sounds promising! Thank you!

On the inkbird, auto tune is a general setting that stays on until you turn it off. I was running it but one idea is the insulation makes it take so long to drop temp it feels like it's not working... I was sat there for over half an hour and it only cycled twice lol I'll try at a higher temperature to see if it can even out!
 
There does not appear to be any temperature limits in the settings. The only thing I could find was temperature alarms... I currently have it going through auto tune cycles at 1450F
 
It appears I can only go above 2000°F by switching to Celsius and even then the limit is 2000. There are also consistent temperature swings of +/- 20°F even with auto tune or manual adjustment.

Does anyone have resources on how to program and adjust these PIDs?
 
Tarkus said:
It appears I can only go above 2000°F by switching to Celsius and even then the limit is 2000
Click to expand...
What type of thermocouple are you using (and programming the unit for)? I am not sure which Inkbird you have ... but according to the docs for the ITC-100, if you are using a K type thermocouple, you should be able to go up to 1300C (2372F).....
 
I'm using a K type and have the PID set as such. I have P=35 I=650 D=670, but I have also used other settings to no great affect. I'll give that thread a read now, thank you!
 
Tarkus said:
I'm using a K type and have the PID set as such. I have P=35 I=650 D=670, but I have also used other settings to no great affect. I'll give that thread a read now, thank you!
Click to expand...
what model controller do you have?
 
Wow, in ONE phrase you have completely changed how I thought about these things! "[The proportional value is relative to the heat loss]". I'm gonna keep reading but that alone helps a huge deal! I have a well insulated oven so it doesn't lose heat very fast unless it's REALLY hot. You have given me a fundamental understanding on the first step to tuning, thank you so much!

I'm using the same Inkbird that the original poster linked!
 
Tarkus said:
It appears I can only go above 2000°F by switching to Celsius and even then the limit is 2000
Click to expand...
just looked at the manual .... according to it, with a K type thermocouple, you "should" be able to program to 1300C. It does not let you do so??? (this inkbird manual is not the best manual out there......)
 
Tarkus said:
Wow, in ONE phrase you have completely changed how I thought about these things! "[The proportional value is relative to the heat loss]"
Click to expand...
Yeah .... with a really well insulated oven, a small proportional band will give you overshoot .... and then with the integral function active, will then give you undershoot (keep reading - hopefully it will make sense). With that well insulated oven, you will want a pretty large proportional band .... but pay attention to just how that is programmed into your particular controller (they do differ in terms of how the proportional band is represented.....)
 
Using your example, I set P=5 I=1(minimum) and D=0. Immediately had it sitting <+/-2°F oscillation. Because I like poking things, I set P=3 and got a 7°F overshoot... Any thoughts?

Edit: I'm dumb. The low P means there's less heat to maintain temp so it dropped and then overshot Trying it at 6 now.
 
Ummm ... plus minus two degrees is pretty darned good. I’d call it a day and go get a beer....
 
That's what I was thinking! Then I noticed it became unstable much like in your explanation later in the post. It will now go no less than 3° below, hold for about 10 seconds then way overshoot by ~30°. I tried flipping the I value from 1 to 9999 but it didn't seem to change anything.

Now I'm waiting on results of adjusting the P value slightly... Just watched it hold temperature and then way over shoot again. One more cycle and it did the exact same thing.

For holding temperature, I could easily set it 15° low and fluctuate at that level... But I'm wondering if the overshoot comes from an inaccurate thermocouple reading. I've heard about issues with sleeves frying at higher temps.
 
Could be ... but could be the integral term playing havoc. When it gets to temp, try turning the controller off then on ... that hopefully will clear the integral (WEO’s idea). Am not at home now ... when I get back I will take a look at the manual
 
I saw that idea, thanks for the reminder! I'll give er a go and let you know.

Edit: it's still doing the exact same thing. For now I'll adjust down and perhaps change my thermocouple later. Thank you so much for all your help!
 
Last edited:
how well insulated is your oven .... and what is the wattage on your element? (likely is above ... but I am being lazy)
 
2.5 inches of K23 fire brick with fireplace cement on all seems except the door opening which uses a layer of kaowool for snug fit. The entire thing is shelled in 1/8" steel except the back where the element comes out. The element is 3000 Watts in a 0.41 cubic foot oven.

Surprisingly enough, opening and closing the door enough to move steel in or out doesn't seem to affect the temperature noticeably.
 
Tarkus said:
That's what I was thinking! Then I noticed it became unstable much like in your explanation later in the post. It will now go no less than 3° below, hold for about 10 seconds then way overshoot by ~30°..
Click to expand...
Ok ... I am home, and was able to refresh myself on the programming for the Inkbird. The Inkbird appears to very strictly use the "standard form" of the PID algorithm (look for that in the PID thread).

THAT MEANS:
  • to minimize the effect of the integral term, set the integral time to maximum value (9999)
  • to turn off the derivative, set the derivative time to zero (you did this)
I THINK what you are mostly running in to is that you have a whopping strong heating element (3000 watts???? !!!!) for a small oven .... to minimize the effect of the heating element blasting on full when you are just a few degrees below setpoint (and thus overheating the oven), you need to minimize the strength of the heating when you are near the setpoint. do this by setting the proportional band to a really big number. right now you have it set at 5 degrees .... so when the temp falls a little below the setpoint, the element comes on full blast.

if you are running temps like 1400F, try setting the proportional band to something more like 150 .... then if your offset is too large, try something like 100....... the smaller you go, the more you risk the instability you are seeing....
 
Status
Not open for further replies.
Back
Top