HM v14 seemingly running out of control?


 

NickMV

TVWBB Fan
I recently upgraded to v14 firmware (HM 4.1) and am finding that my HM is now, on occasion, overshooting and simply refusing to turn the fan output off. And I don't mean in a "I built this thing wrong" kind of way -- I'm talking software.

For example, I currently have a setpoint of 250F. At 255F and climbing, it continued to blow the fan..........in fact, it was still blowing the fan a moment ago at 262F and climbing (12F above setpoint, with no end in sight).

When I say "fan output", i'm referring to a substantial 30-40% output. I don't understand why it's happening, because I've never experienced this prior to upgrading. I'm luckily able to get it to cut off if I feed in a substantially lower setpoint temp. So in my current situation, in order to get it to shut off fan output for 250F setpoint / 262F actual, I had to set it to 230F and at that point it finally went to from 30% to 0%.

Anybody else experiencing this? I'm stumped.
 
Are you sure your PID settings didnt get wacked in the upgrade?

Definitely sure, just checked and they did not change from what I screenshotted pre-upgrade.

I'm also trying some diff PID settings in the mean time to see if it makes any diff.

Here's the orig settings I was using (which worked fine on both RD3+22WSM as well as a 15CFMservo-blower+ large offset cooker:

P - 2
I - 0.01
D - 15

New settings I'm trying:

P - 4
I - 0.02
D - 5

nickbbq.ddns.net:8080
 
Last edited:
I can't remember anything changing in the PID control logic with regards to startup, overshoot, or windup so it seems like it is specific to your conditions. HeaterMeter will always continue to stoke the fire after passing the setpoint because that's how it works and relies on integral decay to bring the temperature back down. What could be happening is that your temperature is approaching the setpoint so slowly that the I term builds up and it takes a very long time to decay back to where it needs to be to maintain 250F (the value is non-zero but HeaterMeter needs to figure it out). I would say it almost certainly is not a new software problem-- I think rolling back to v13 will give you the same results.

Regarding your PID values, a low P (2) and a high D (15) can contribute to this. If your output is 30% at 262F then I'm betting your I value nearly capped out at 100% before crossing the setpoint, where it would be immediately halved to 50%. At 262F your P value is (250-262)*2 = -24%. Your temperature is stable so D = 0%. Total output = ~26% and your I value is decaying at 0.12% per second. If your temperature drops 1 degree in the next minute, your I is now down to 42.8%, P=-22%, D=1*15%=15% for a total of 35%. That's right, the output went UP when you actually wanted it to go down. This is because of the high D term, which tries to counteract temperature changes.

A high D is used for a very slowly reacting system so change happens very gradually which partially accounts for what you're seeing, that it takes a long time for the system to return to the setpoint after an overshoot.
 

 

Back
Top