llbaker2 wrote:You are repeating what I just said except for the part about setting Artisan to log only (which is what I normally do) Guess that means you agree. Will proceed with testing as planned and see what happens. May be a while before I'm able to do that many test roast. Will report back with results in case anyone is interested.

Thanks again for your time and input Randy G.

I didn't want Artisan in control mode when it was reading ET and assuming it was BT (and vice versa) to alleviate any situations where the coffee could be burned (if that would have been possible). But after the switch, if the spike happens on ET then it will indicate a BT thermocouple problem (which is quite rare).

Since the "+" model came out, in all the time I was working for Hottop, I never saw the problem you are having. The thermocouple plug reversal on the control panel is the easiest to do to get started with.

Understood, you make a good point. I normally use Artisan in record mode but you didn't know that. Do you think running these test without beans would be OK? Im thinking of trying without for most testing and maybe throw in some beans too after doing a few without. Thing that makes this hard is the intermittent nature of the spike problem. Some roast run just fine, some prematurely eject due to spike. Might not be a problem with machine. Might have something to do with a power surge or the like. As you pointed out the fact that the spike happened at the same time that the fan went to zero is kinda interesting. Maybe the change in load had something to do with it? The profile Ive been using does have multiple (2 or 3) fan speed and heat level changes within it.

llbaker2 wrote: Do you think running these test without beans would be OK? Im thinking of trying without for most testing and maybe throw in some beans too after doing a few without. Thing that makes this hard is the intermittent nature of the spike problem. As you pointed out the fact that the spike happened at the same time that the fan went to zero is kinda interesting. Maybe the change in load had something to do with it? The profile Ive been using does have multiple (2 or 3) fan speed and heat level changes within it.

I would usually tell customers to try "empty roasts" to test for problems or to verify solutions before roasting beans.

You might want to go back over our saved graphs to see if the previous spikes happened when the fan's speed was raised. The main fan can cause early ejection, but I was never sent a graph to know if that sort of spike was evident. It could be a fan problem because when the temperature spikes like that the control system may see it is having reached the user-programmed target temperature or that the programmed maximum temperature allowed has been achieved. If you had posted, "My roaster sometimes ejects the beans early without me doing anything," and there was no graphic evidence, I would have stated that the cause was very likely a failing main fan. But it is odd that it happened at shutdown when the fan went to zero.

Sorry its been a while but I only roast once a week or so. Since my original post I have done 9 roast. Two of the nine were no bean test roast. Out of those 9 roast I had 4 spikes. All 4 spikes actually occurred after roast was ejected. One spike occurred only a few seconds before I ended the recording so only the beggining shows on the chart. One of the 4 occurred almost immediately after the end and posed a potential hazard because the beans only partially ejected. The machine shut down in the middle of the ejection process due to spike. I ended this roast in the middle of second crack and the beans that weren't ejected continued to crack. The spike caused the machine to shut down so I had to resort to use of the emergency eject knob. Bean that did not eject got very dark by the time I remembered the emergency eject (30 sec or so after shut down). I have included images of all nine roast below.

Now that I have proof that the spikes are not just a fluke but a recurring problem I will do as suggested above and switch the leads on the bean and environment thermocouples and do a few more roast to see how the switch shows up in the recorded profile. IF the spike shows up on the environment thermocouple rather than the bean (showed spike before switch) this would be a pretty good indication that the panel is bad. If the thermocouples still show spike on bean this would indicate the thermocouple it bad.

I attempted to post 9 profile images but they didnt show up so gonna try posting 3 at a time.

OK, see what I did wrong. I did not click "place inline". Here are other 6 profile images