PID controlled, by ET.

The ET sensor is obviously connected to the PID. The BT sensor is connected to a temp meter, which is connected to my laptop during roasts. I monitor BT, and use it decide when to end the roast.

Randy

Good set up for a P1

If the graph's jaggies are real, not just sizing artifacts; it may be worth detuning the PID parameters by 1/3 to see if that gets rid of the them. You'll get a few seconds lag between the target and actual, but that's better than overshoot and then the controller turning down the heat too far

I put a higher resolution picture here

http://drop.io/rt_roast#

You can still see here, and there are still jagged edges.

I will try playing with the PID settings a little for the next roast.

By detune the PID parameters, do you change all of them by 1/3, so slow down the response?

Randy

Yeah, each a 1/3 less response. For the D and the I, that means lower numbers. For the P, it's a higher number if it's in degrees or percent of range (like Fuji controllers), a lower number, if it's in gain (like the old US built ones).

I added a couple more graphs to my drop page, one with more insulation on the chimney, which I normally use, and one with less insulation, which was my old configuration.

http://drop.io/rt_roast

If you look at them, they look pretty close to each other.

With both roasts, I have tried adjusting the PID parameters a little, but I still don't get a smooth ET, but more insulation does give a much smoother ET, which seems to be the biggest difference.

The average delta T is 80F for less insulation, and 72F for more insulation.

I guess these results still don't make sense, but they make more sense then before. I guess the other thing I am getting from these measurements is that the added insulation for the chimney is not making a too much difference in the delta temps, but the added insulation is making the ET graph smoother.

Randy