Trying to sort out my PID settings.....


 
C

Chris Statton

New member
Hi all,

Firstly thanks to everyone involved in the HM & RD3 projects for all the hard work.

I completed my second cook using the HM & RD3 yesterday, and have had a bit of regression in temperature control success. The first cook went quite well once the PID values seem to be nailed with thanks to RalphTrimble for the settings advice. But yesterdays cook fluctuated much more than expected (see graph below). I'm wondering if the community could help me to understand how the settings will (or should) affect the graph in question, so I can go into my next cook armed with the knowledge I need.

First, I should explain my setup to give people an idea of what we're dealing with here....

Heatermeter 4.2 with Thermoworks thermocouple and 3x Thermoworks probes.
  • Version 20151013B
  • OpenWrt Attitude Adjustment 12.09 | Load: 0.41 0.11 0.04
RotoDamper RD3 connected to single 1" inlet located in line with bottom of the fire basket
Ugly Drum Smoker with Weber Lid (Weber top vent opened 1/3 of the way typically)
PID settings:
  • B - 4
  • P - 3
  • I - 0.005
  • D - 5
Fan settings:
  • min - 0%
  • max - 20%
  • startup max - 100%
  • Invert output - NO
  • On at max only - NO

So the PID settings and fan settings were suggested with thanks by RalphTrimble, and as I said, they worked great on the first cook. After a quick look yesterday, RalphTrimble suggested I open the top vent up a little more and that's what can be seen with the amplitude lessening at about 13:00 on the graph.

TUsF6if.jpg


Hope I haven't missed any info, but if there's anything more I need to tell please ask. I'll be ever so grateful for some advice.

Thanks!
 
Last edited:
Have you read the WIKI section on PID tuning? That would be a good place to start.

There is a lot more info there that is helpful in understanding PID tuning, but the manual PID tuning method goes like this:

1)From the LinkMeter configuration page, set the controller to Proportional-only mode by setting B=0, I=0, D=0.

2)Increase the P until the output of the loop oscillates, then the P should be set to approximately half of that value for a "quarter amplitude decay" type response.

3)Then increase I until any offset is corrected in sufficient time for the process. However, too much will cause instability.

4)Finally, increase D, if required, until the loop is acceptably quick to reach its reference after a load disturbance. However, too much D will cause excessive response and overshoot. A fast PID loop tuning usually overshoots slightly to reach the setpoint more quickly.

So in P only mode you adjust to get a stable oscillation (not diminishing or ramping upward), then set P to half that value which should make the oscillation decay rapidly toward the setpoint. Then you tweak the other PID values.

From looking at your graph it seems you may have excessive response, so a reduction in D may be helpful.
 
Last edited:
I was all excited yesterday that I finally had my HM and RD3 ready to go, so I loaded up my kamado's bowl, torched it for about 15 seconds in the middle and put my deflector plate and grill grates back on as if I were doing an long indirect cook at 225, no meat though.

After plowing right through my setpoint, I realized I had left the default PID values in there and had to adjust. I quickly grabbed Bryan's BGE settings from the PID Tuning Wiki, and tweaked the max and on above.

Defult:
Parameter Value
B 0
P 4
I 0.02
D 5

Bryan's BGE
Parameter Value
B 0
P 2.5
I 0.0035
D 6
Min Fan 10%
Max Fan 100%
On above, initially had this at 90%, then tuned it down to 60%.

I started with my 1/16th inch high temp TC probe inserted through the side. Then later fed it through the top vent with the vent closed enough to hold it firm. I also had a ET732 on a clip at the center of the grate.

So, I think if I read the wiki right, the overshoot on my next start up should be lessened by taking my P value from 4 to 2.5 as I have.

Once it got back down to 225, it was stable from 5:30pm yesterday until about 1am this morning. The oscillations grew, but seemed stable from 1am until about 11:30am today until the pit seemingly ran out of fuel. Would raising the I value and lowering the D value temper these oscillations more than what I saw during this time, or is this expected in the later stages of the cook?

I'm going to try some steaks tomorrow, so that will be a good time to tinker. If I see the servo and fan output staying high as it did yesterday when approaching my setpoint, would adjusting the P value be the first change? Will it correct quickly enough or would I need to adjust the fan manually? What oscillations am I looking for in Ralph's reply above if the graph is still on the rise?

Thanks!

Here are my charts:

Startup to Stable: 11/15/15 from 1:30pm - 5pm

http://imgur.com/5Exjbk1


Stable to burnout: 11/15/15 8:35pm to 11/16/15 8:35pm
http://imgur.com/BwVxsvJ
 
Last edited:
I see a low and slow with a lot of initial overshoot
So, you either need to start a smaller fire, or perhaps you need to clamp down the top vent a bit (at least for the forced air portion of the process), or have air leaks elsewhere in the grill, or perhaps need to back down on your blower speed for low and slow cooking.
There are a lot of possible variables in play, with different grills and different HM configs etc.
The way I do a low and slow is with the fan on at Max Only, 100% Startup Max, around 20% Max and 0% Min. This way the fan stokes the pit temp and then lets the damper control the majority of the cook, only kicking in (at 20% speed) if the HM output hits 100%.
One thing to keep in mind, a pit driven by forced air flow (blower running) is much different than a pit run by natural draft, the major difference being the top vent setting.
A forced air flow scenario works best with the top vent closed fairly tight, the blower will force the air out of the smaller holes and when the blower kicks off the top vent will tamp down overshoot fast.
A passive damper controller pit works best with the top vent opened a bit, allowing the natural convection flow to stoke the fire.
So I typically start a small pile of lump coal with a weber starter cube (or even half), let the starter cube burn off (air out that dirty smoke) and then close the lid with the top vent just cracked open. The blower will stoke the pit will and smoke pours out the vents. Then when the target temp is hit the blower cuts out, the damper chokes down, the pit settles... Now it is time to put the food on and also adjust the top vent open a bit for the passive section of the low and slow. Adjust the top vent so the RD runs about 35% open in the beginning and it should be able to run a nice long cook.
 
RalphTrimble said:
I see a low and slow with a lot of initial overshoot
So, you either need to start a smaller fire, or perhaps you need to clamp down the top vent a bit (at least for the forced air portion of the process), or have air leaks elsewhere in the grill, or perhaps need to back down on your blower speed for low and slow cooking.
There are a lot of possible variables in play, with different grills and different HM configs etc.
The way I do a low and slow is with the fan on at Max Only, 100% Startup Max, around 20% Max and 0% Min. This way the fan stokes the pit temp and then lets the damper control the majority of the cook, only kicking in (at 20% speed) if the HM output hits 100%.
One thing to keep in mind, a pit driven by forced air flow (blower running) is much different than a pit run by natural draft, the major difference being the top vent setting.
A forced air flow scenario works best with the top vent closed fairly tight, the blower will force the air out of the smaller holes and when the blower kicks off the top vent will tamp down overshoot fast.
A passive damper controller pit works best with the top vent opened a bit, allowing the natural convection flow to stoke the fire.
So I typically start a small pile of lump coal with a weber starter cube (or even half), let the starter cube burn off (air out that dirty smoke) and then close the lid with the top vent just cracked open. The blower will stoke the pit will and smoke pours out the vents. Then when the target temp is hit the blower cuts out, the damper chokes down, the pit settles... Now it is time to put the food on and also adjust the top vent open a bit for the passive section of the low and slow. Adjust the top vent so the RD runs about 35% open in the beginning and it should be able to run a nice long cook.
Click to expand...

You are good.

I did have my top vent wide open for most of the startup, so I will close that off more tomorrow. I definitely don't have any air leaks. My pit is about as well sealed as could be.

Does the HM software track the output of the RD like it does the fan? I see the live reading of it but was curious if there was a history like the blower output.

I'm loving that I've got it to this point and can start to have fun with dialing it in and getting all of the bells and whistles turned on with alarms, apps, etc....thanks again!
 
Nailed 225 in about 30 mins after lighting my fire. I used Bryan's BGE settings to start this time around on my kamado with the top vents open just enough for the Pit probe cable. Everything after the 17:45 mark is noise as I took out my deflector plate, through some steaks on, flipped them once, tried out some new probes, etc....

I'm now confident and excited for my next long low and slow!

http://imgur.com/xdT7CoR
 
Not bad for just getting your feet wet with the HM. It takes me typically about 15 minutes to get my grill settled on temp but I have a LOT of cooks under my belt and I've got the technique down pat...
 
RalphTrimble said:
...
...
The way I do a low and slow is with the fan on at Max Only, 100% Startup Max, around 20% Max and 0% Min. This way the fan stokes the pit temp and then lets the damper control the majority of the cook, only kicking in (at 20% speed) if the HM output hits 100%.
One thing to keep in mind, a pit driven by forced air flow (blower running) is much different than a pit run by natural draft, the major difference being the top vent setting.
A forced air flow scenario works best with the top vent closed fairly tight, the blower will force the air out of the smaller holes and when the blower kicks off the top vent will tamp down overshoot fast.
A passive damper controller pit works best with the top vent opened a bit, allowing the natural convection flow to stoke the fire.
So I typically start a small pile of lump coal with a weber starter cube (or even half), let the starter cube burn off (air out that dirty smoke) and then close the lid with the top vent just cracked open. The blower will stoke the pit will and smoke pours out the vents. Then when the target temp is hit the blower cuts out, the damper chokes down, the pit settles... Now it is time to put the food on and also adjust the top vent open a bit for the passive section of the low and slow. Adjust the top vent so the RD runs about 35% open in the beginning and it should be able to run a nice long cook.
Click to expand...

This is amazing advice, thanks Ralph. This was obviously a few years back, are you still using this method? I imagine this would prolong the blower life also.
 
Yes, I still use the same routine for running my pit these days. My blower and damper are still going strong after being out on the grill 24/7 for several years now, no problems....
 
Hi,
What/where is B parameter?

In configuration i see only P, I, and D...

I’m currently struggling to set up my wsm 18’(47cm).
 
I have 22, 18, and a mini built WSM

P- 4
I- 0.009
D- 5

Start 100%

Min 0
Max 50%



That’s what I use on every cook for every one and I have maybe a 5 deg diff. I’m ok with that!! Haha. I don’t think my ribs will care too much.

I may try 30% to see if that helps but it’s damn stable at those right now at lease.
 
You must log in or register to reply here.

 

Back
Top