Artisan PID Control Problem

[AlexC1107]

Hey guys, I am currently using the build in PID controller to roast, and limited the output(0-85%)(connect to the Heat slider). I have just found that in the later stage of the roasting, I actually don't need the full 85% of heat. Is there any command that I can use to change the MAX of the output limit? Like 0-85% in the beginning, after certain time, the limit will change to like 0-65%.

The Heat slider is controlling the Phidgets RC Servo which will turn the knob for the heat input.

Thank You all for getting me some help!

[hermetheuscoffee]

I don't know of a way to dynamically do this, but you CAN change that value in realtime during a roast, so perhaps that would work for you? With a well tuned PID, you generally shouldn't have to lower that.... the PID itself should be turning down the heat at the end if your temps start to ramp up. (more aggressive kP values can help in that area.)

[MAKOMO]

Could you show us one of your automated roasts incl. the heater adjustments done by the PID? As Jason, I wonder where the need form adjusting that limit during the roast raises from. A proper tuned PID should demand only the heat that is required. I also wonder what type of roasting machine requires less heat at the final stage of the roast vs the begin of the roast.

The output limits of the PID cannot be changed by Artisan Commands as we don't think there is a need to do that dynamically. Convince us from the opposite and we will add a corresponding command for this.

[jkfrfr]

I just watched this and learned a lot about PID tuning, I highly recommend it https://www.youtube.com/watch?v=FstdgnbaM2k

[hermetheuscoffee]

Yes, that's my video. :-) But I will be taking it down soon and replacing it with an updated version. Specifically I made a gross error in my recommendation of DoM vs. DoE. If you try using DoM with the D values that I show (around 70), you'll never hit your target temp. So to be clear, please use DoE mode in the PID's "CONTROL" dialogue. Otherwise, I stand by most of the rest of it.

[jkfrfr]

I assumed as much :D Thanks for the video, it is very helpful. I wonder if the kd could use some more attention? On my roaster (150g/1400W fluid bed), I found low kd values to be most helpful. I found if kd was 'relatively high', the response time of the PID slowed too much. I guess a reason for the difference could be that my machine is much smaller and 'faster' - there is almost no thermal mass, and I use inlet temperatures, which respond extremely fast to burner adjustments. Just food for thought.

[hermetheuscoffee]

Attaching a completed roast would be helpful. I'm not sure the kd value NEEDS more attention... you can pretty much nail roasts with a well-tuned PI system. I'm really not convinced that D play a necessary role in it. If your I value is mild enough so that the "baseline" level of heat is established slowly over time, then the overshoot will be minimal (if at all). Then your P value can effectively deal with deviations from your background curve.

[AlexC1107]

Sorry for the delay guys. We have watched the video and try to adjust our pid setting according to the video. I also putting a few screenshot, let me know if you can't see the screenshot.

[Terracotta-6]

I also putting a few screenshot, let me know if you can't see the screenshot.

Looks great!
Please provide a full screenshot of the PID Control UI and indicate whether a segmented PID control method is used.

[AlexC1107]

We are trying to have a command for turning on the limit is because the overshoot mostly happened in the later part of the roast(after first crack), but in the earlier stage, we want need the full heat for the roast.

[AlexC1107]

And thanks for everyone's comment, we will keep trying and will keep you guy updated if we success or need more help😝. Thank you AGAIN!!

Mesdames et Messieurs, Joyeux Noël et Bonne Année!