Artisan delta BT curve not smooth

Discuss roast levels and profiles for espresso, equipment for roasting coffee.
Posts: 1
Joined: 10 months ago

#1: Post by KerfyNird »

My delta BT curve is not smooth in Artisan. I'm not sure why this is the case. I can get an idea of the trend as it is roasting, but would be better if it didn't seem to have so much noise. I am using a Hottop 2k+ on Artisan 2.8.2 (linux version). I'm using the standard setup with 2 second sampling.

User avatar
Supporter ♡
Posts: 1759
Joined: 6 years ago

#2: Post by mkane »

Thats normal. Add smoothing if you like

Posts: 470
Joined: 17 years ago

#3: Post by rmongiovi »

Well, strictly speaking it might be normal or it might be electrical noise. There's no way to tell without seeing what it actually looks like and investigating.

But as Mike says, even without a problem it's normal for the delta ET and delta BT curves to wobble. There's inherent noise in the probes where the readout isn't the same from sample to sample. Additionally, for the BT probe you're measuring an ever changing mix of beans and air which are at different temperatures so the readout is going to vary depending on how much of what the probe is in contact with. A thicker probe will alleviate this but the downside is that it always returns the wrong temperature since the thermal mass of the probe requires time to react. Using Artisan's smoothing has much the same effect. Smoothed values add delay to the results because they incorporate values from a selected number of past readings.

So you've got a tradeoff. Put up with the wobbles and eyeball it yourself or smooth the results to get a nicer curve but accept that you're losing details. In any case, the readout of the probe isn't actually the true bean temperature. It's just a representative sample. Accept that it's only an approximation of what's really going on.

User avatar
Posts: 3612
Joined: 10 years ago

#4: Post by Almico »

That doesn't look like noise to me. It looks like poor probe/bean mass contact.

Posts: 470
Joined: 17 years ago

#5: Post by rmongiovi »

Remember that the thermocouples that come with the 2k+ return only integer temperatures. That naturally contributes to a bit of inconsistency in the RoR.

Posts: 331
Joined: 14 years ago

#6: Post by rgrosz »

For my B-2K+, I use a delta span of 5 seconds and smoothing of 6. Here is a link to ALL my settings for artisan 2.4 ... index.html
LMWDP #556
Life is too short to drink bad wine - or bad coffee

Team HB
Posts: 5517
Joined: 16 years ago

#7: Post by ira »

I would think if you're only taking readings every 2 seconds, you be much better off taking them much more often and averaging them over the past 2 seconds. As I recall when I wrote the code to monitor the temperature of my Behmor, I read it a hundred times a second or more and averaged it over the last 1/4 or 1/2 second. Just seems like that would get you smoother results without sacrificing accuracy.