How to create IT/BT curves for PID control?

[jkfrfr]

For some time I've been happily using Artisan PID to control either the inlet temperature or bean temperature to follow background curves on my DIY air roaster. I've also been using time-based events to control my fan levels. I've seen IT/BT curves on Roest machines, and read that the Artisan PID can now follow IT/BT curves, but I can't for the life of me figure out how to set that up.

I'm guessing I need to set Profile Background>Playback Events>by BT and Ramp>Burner, but I see no place to create "IT events", if they exist.

This post (https://artisan-roasterscope.blogspot.com/2025/04/artisan-v312.html) mentions "IT SV background events". Where do I find them? If anyone has any experience with this, I would appreciate it! I'm using 3.4

[jkfrfr]

Hmmm, maybe I have it. In Events>Config, rename an event type (3, Damper, in my case) to IT-SV. Then in the Event>Sliders tab, check "IT-SV", change the Action to Artisan Command and Command to pidSV. Then with a background curve open, go to Roast>Properties>Events, click Add, change type to IT-SV and give it a value. I not going to test this tonight as I'd like to get some sleep :D

[jkfrfr]

Looks like Artisan Command might need to be pidSVC to hit a given temperature

[MAKOMO]

Bravo! You got it.

pidSV({}) should. It sends its argument unmodified to the PID infrastructure while pidSVC({}) assumes its argument in C to be automatically converted to the temperature mode selected by Artisan. If temperature mode is C both have the exact same effect.

All this is somewhat experimental. I am using mostly replay of SV target for my inlet temperature by time and haven't tested the by BT much. Maybe the BT/time mode is more adequate. It is somewhat described in the v3.2.1 release post.

I am sorry that this, as many other more experimental things, are not (yet) properly documented.

The Santoker Cube Bluetooth PID setup is what I am using for IT/time profiling.

Note that you can move background even targets during the roast to adjust things dynamically.

Take care and do not burn down the house!

[jkfrfr]

Thanks for the response! I'm happy to hear I'm headed in the right direction, I'm always happy to be experimenting. I will post some results after testing the IT/BT curves.

[jkfrfr]

So I did my first trial run with an IT/BT curve, using the 'by BT' and ramping functions.

First off, it worked quite well, which was surprising. In fact, it hit the DE, FC and drop targets more precisely than my normal profiles, which is surprising as this was a first try and not at all fine-tuned.

The second thing is that I don't really understand how it did that. I'd like to think it wasn't luck, but I don't get it. You can see that from TP to DE, the measured BT was well below the target background BT, and that the IT-SV events were delayed because of this (this should slow the whole roast, as I understand) - but then somehow it pulled it back together to hit the DE target exactly.

Then, if you look closely, you can see that from minute 5 to FC, the measured BT was slightly higher than the background BT, and so the IT-SV events were moved forward it time. It looks like this keeps the BT curve higher than it should be, but then hits FC through DROP pretty much exactly where I wanted them. So was this luck or do IT/BT curves actually hit their targets more accurately?

[MAKOMO]

Could you please attach your Artisan settings as .aset.txt (menu Help >> Save Settings) and those two profiles as .alog.txt (note you need to rename those files to end in .txt to be able to drag-and-drop them here to a comment field. I will take a look.

[jkfrfr]

Here you go, thanks for looking. The background profile is BT FC 7.7, the roast curve is Test Roast 143

Test Roast 143.alog.txt

artisan-settings 3.1.aset.txt
BT FC7.7.alog.txt

[jkfrfr]

Also, while roasting, I couldn't really grasp what it was doing, so I took a screenshot.

The blue vertical line is crossing the IT background profile at about 195-196c, but the IT-SV slider says 190c and the SV slider says 198. What confuses me more is that the measured BT was running lower than the background BT, and you can see the IT events were delayed in time. The most recent IT event (purple dot) is about 180c, so is the IT-SV valued taken from the line between the that point and next IT-SV event, and the blue vertical cursor line - which looks to be about 190c?

Is this how the IT/BT curve corrects itself, because the slope of the line between the most recent and next IT-SV events is steeper than than the corresponding points in the background IT curve?

[MAKOMO]

Yes, I think this is what is happening here. You should use an extra device of type "PID SV/Duty %" to have the actual duty being recorded. You will have to repurpose one of your extra devices as the current limit is 10 (which we thought would be A LOT!).

As said, the Playback by time/BT should work better.

[jkfrfr]

After some more testing, I think the "by BT" function is not working. I repeated the above roast (TR143) with the exact same beans, etc, changing only to the "by time" function. I ran the roast (TR145), and compared them. When overlayed, the two roasts are almost identical. The Duty% for the burner is on the bottom of each graph, and it too is practical identical.

Test Roast 145.alog.txt

I also took a screen shot mid-roast, about the same time as TR143, and you can see that the IT-SV values are different from TR143.
One can also see that the IT-SV events are being happening at the same time as the background curve, indicating that they are being calculated by time rather than by BT.

My guess is that the red background IT curve is taking precedence over the purple IT-SV event curve, which is why both roasts came out the same. One can see that the IT-SV curve is being calculated (based on screenshots), but I think they are being ignored for the red IT curve.

How do I deal with that?

[MAKOMO]

Are you running the PID in manual mode (pidmode(0)? I cannot see this from your settings where the PID is configured for "Ramp/Soak".

If you have the PID in background follow mode, it will do this, follow the backgrounds IT curve. With the PID in manual mode, the pidSV({}) of the IT-SV slider will take effect.

[jkfrfr]

No, that must be the issue, the PID is in background follow mode. I will give it another test soon.

[jkfrfr]

I set it to manual, and gave it a quick test. The IT SV wouldn't leave the preheat temperature - I'm pretty sure I have something with the slider events setup wrong. I tried to figure it out, but wasn't successful. I'm sure its something simple.

I set "max" to 300 because I thought that was the maximum celcius temp it should output, but probably thats not correct. And I'm pretty sure I need something after pidSV, but couldn't get it right.

[MAKOMO]

The Command for IT SV should be pidSV({}). The argument {} is a placeholder variable for the slider value to be set as PID SV.

[jkfrfr]

Thanks! I was just testing that. It didn't work at first because I had turned off the SV slider in the PID Control. I thought maybe they were interfering, but no, its necessary!

Now it looks like its responding as I would expect.

[MAKOMO]

Don't forget to record the PID SV with a corresponding extra device configuration.