BPID settings for recovery

Hi,
I use the HM on my cabinet smoker (door instead of a lid) and during a cook, I send the lid open signal to cut the fan for a couple of minutes while I add meat / wrap, move around or take out meat.

Sometimes a door opening event is less than a minute, and other times it may be more, longer lid open events require more heat to recover back to cook temp.



Ignore the noisy "Fire" temp in grey, its not right.

A quick recovery is on the left (pulled out a brisket, pit dropped to ~112) and a slower recovery on the right (photo opp, wind blowing, moved a turkey, put the cleaned probe into a cooking pork loin.... temp dropped to like 87 degrees)

The recovery on the left did not require the fan to go to 100%, but the one on the right took 100% for about 3 minutes, then it backed off and increased again, then slowly started cutting the fan until the temp hit the target temp, then it really cut back (but still kept pushing!)

Did I trigger startup mode settings or something ?

This is my first run with a servo, and the servo is working a damper on the exhaust side.

My servo does not move much until the fan is at 0%... is that the way it is supposed to work? The temp continues to rise when the fan is at 0 once the damper turns to the closed position. There is some area open for the exhaust even when the damper is "closed".

Here are my settings:


I have a 2x4 air inlet (1-3/4" x 3-3/4 ID or 6.5 sq in.) and use three 3" exhaust vents with about 1/4" opening around the butterflies when closed. That is 2.15" area each, so ~6.4 sq inches of exhaust vent when closed.

Using the three vents, I can make the exhaust come out of the cabinet low, middle or high. The butterflys are tuned to let a little more air flow out of the bottom, because I am trying to use a reverse flow pattern inside the cabinet.

Gary Graham said:

My servo does not move much until the fan is at 0%... is that the way it is supposed to work?

Click to expand...

The amount of movement of the servo is determined by the difference between the two SPD numbers, in your config above they are very close together and therefore the servo will not move much going from 0-100% If you want more movement put more spread between those numbers....

RalphTrimble said:

The amount of movement of the servo is determined by the difference between the two SPD numbers, in your config above they are very close together and therefore the servo will not move much going from 0-100% If you want more movement put more spread between those numbers....

Click to expand...

Thanks Ralph.
I found the PID explanation by Bryan at http://tvwbb.com/showthread.php?42779-Post-your-HeaterMeter-graphs&p=466141&viewfull=1#post466141 and worked through the numbers using my cook data.

My cabinet is an offset with the fire outside the cook chamber so that gives a slower response.... so it seems a larger P is in order. Doing the math from a drop to 150 degrees (typical door opening event) I could use this: 225-150=75

For the I term, I think I will hold at .005. The overnight portion of the cook was really flat with 1%-3% fan until the fuel started to burn out around 5:30AM.

For the D term, I am not sure if it should change... sounds like I can keep it low like it is so there is a bigger spread between P and D.

If I understand correctly the larger spread between P and D bump the servo a little towards the closed position (earlier...) as the fan duty cycle is dropping, rather than all at once when the fan hits 0.

Will try BPID = 0, 75, .005, 2


P.S. I think I may have fried my thermocouple....



Shows my gravity fed charcoal firebox, the server fan assembly and thermocouple in the heat transfer tube.

Unusual to see a blue-ish color for the glowing red metal on the firebox and heat transfer tube, but the combination of florescent light and infrared must make blue on the camera. I think this is the hottest I have seen the xfer tube and I am sure I cooked the TC.

My guess is I caught the burning of the "bottom of the bag" charcoal giblets. The small bits burn really hot, so that's why I use a TC to keep an eye on the fire. I think the lumps are what cause the bumps in the night in the uprights.

I think I now recognize that sorting lump charcoal by size is a best practice, and I should probably move the TC more under the cabinet, away from the grate area.

I was referring to the movement of the servo, not the PID tuning... The difference between you servo SPD settings is small, so there is going to be little servo movement (like you reported: "My servo does not move much until the fan is at 0%"). You would get more gradual servo movement if your damper rig allowed the servo SPD spread to be greater.


EDIT:
I also noticed that your SPD settings are higher on the left and lower on the right, which I had never tried. (I am running my current roto damper at 780-2100 and Inverted) I wasn't sure if having the SPD settings backward like you have would cause a problem or not, but I just set my servo SPD to run from a higher number to a lower number and it freaked right out, so no good! Reverse those numbers, make the lower SPD number be on the left and the higher on the right. If you need the servo to move in the opposite direction select INVERT OUTPUT, don't put the high number before the low number....

OK, will try it again in the next couple of weeks.

Thanks for doing the backwards number test.... Was your roto assembly binding with that arrangement ? My butterfly arrangement will tip 90 degrees in each direction, so I have it setup where "closed" is the centered position i n the full range of motion. I wanted the butterfly to tilt the other way, so that required me to reverse the numbers. I can re-position the splined arm so "closed" is at one end of the movement range and use invert as suggested.

I am using a smaller sweep on the servo because of the way the area increases a lot with a small movement of the butterfly blade. To increase the sweep, I think I may try to block off one of the 3 exhaust tubes (vents) and then use more of the SPD range.

If I use a plug inside the cooker, that would be the easiest to change if needed, even during a cook.

Thanks for the comment and test.

Gary Graham said:

Thanks for doing the backwards number test.... Was your roto assembly binding with that arrangement ?

Thanks for the comment and test.

Click to expand...

Yes, my servo was binding when I set the numbers backward, as if it wanted to spin all the way around backward or something. There was very little motion then a rapid spasm of motion, the motion I saw was in no way coordinated with the small movement that should have been happening, I was afraid the one trial might have killed my servo but it lived.... I would correct it right away, the servo should move in nice slow stepping motion when you raise and lower the blower speed, not a rapid spasm like that...