Originally posted by JavaB link=1167177851/30#30 date=1167788268
I did a bit more fiddeling with my HG to understand the control better. Temp is controlled with a large dial with has many "teeth" (can be seen on the the pics I posted earlier).
Couple of observations from measurements (TC in front of nozzle, outside the bowl) in the 60 .. 75% power range:
1) when stepweise increasing power by advancing by one tooth, temp rises quite consistenly by 10 deg after some time lag
2) its impossible to map these dial position to **absolute** temps. Eg when running the HG for some 15 min temps were 10 .. 20 deg higher (gun heats up, maybe the control Triac circuit is is temp sensitive, resistance of TRIAC may chnage, 1000 effetcs etc)
3) when changing control direction from eg up to down , there seems to be some hysteresis before the HG responds with appr. 10 deg down steps. You have to dial back 2 .. 3 teeth for that. I also read somewhere that certain Triac "dimmer" circuits have such hysteris behaviour
4) when giving over some time up and down commnads , the temps come always out a bit different. However, smaller **relative** temp up/down changes are consistently possible with these caveats
I tend to believe that in my case building external TRIAC dimmer circuit makes not much sense. The HG appears to already have a basic dimmer circuit type of control, although with only one potentiomenter moved by the dial. So I already get roughly 10 deg steps control, just like at above mentioned Steinel HG and incidently like your external circuit (power change with 10 deg at 250 deg operating temp = 10/250 = 4%). And different control circuit doesnt help against the thermal lag and need to "plan ahead" with the temp settings, which, in this case I suppose means that when during the exotherm phase of 1st crack the beans produce heat and increase temp you should probably *increase* the temp by one notch in anticipation of the heat absorbing phase coming in 30 sec later or so , etc.
Now being better aware of these issues you mentioned my last roast went roughly like this (sorry no nice log file yet):
start of 1st crack around 7 min @ 199 deg,
8 min: 209 C
9 min: 216 C
10 min 220 C
11 min 224 C sencond crack starts
12 min 228 C, roast off
So between 1st & 2ncd crack I have initially a slope around 10 deg/min, then down to 4 deg/min until the end. Does this look reasonable ?
I used some peruvian beans, is endothermic behaviour very bean specific, ie are some beans more prone to stall than others after 1st due to high heat absorption ?
Anyway, I am almost exclusivley using these beans, its a workmate who runs a specialty coffee as 2nd business in his "free time", so its very convenient for me to get the beans from him (and they are good for espresso roasts). So maybe with some more roasts I will learn to manage the temp behaviour of these beans together with all the thermal lag in the control loop. Well, thats the main fun of this whole home roasting, cranking the HG controls around like mad and yet having in the end a decent cup and being able to do roast profiles which even a professional Probat roaster owner cant do (with larger charges) and all this with a heap of kannibalized home tools for a few bucks ...

Sorry for putting you all asleep with these tedious ramblings, every Coretto-roaster wonders probably since hours why the heck arcticroaster is not backing off his HG, there you got your nice continous control of flow an heat without any fancy TRIACs or whatever, but the sad truth is in my setup backing off the HG is not an option: the lid is in the way of the nozzle and I would loose all temp readings as I would be pulling out also all TCs as well ... :

Cheers, Peter


Leave a comment: