Howdy All,

I roast on an electric drum roaster, Toper Cafemino Electric (1kg). A number of folks have expressed interest in discussing how to roast with consistency, in particular how airflow can impact the crash/flick in roasting. First is how I manage the profile and particularly the crash and flick, followed by discussion of heat input and drum speed. My results are simultaneously a bit roasty and sometimes a touch too acidic/underdeveloped tasting. I will post some profiles from Artisan tomorrow after the storm blows through and I can get out to the roaster to grab the images.

The roaster has been modified with a SSR to control heat input, and a linear actuator to control airflow via the damper. Both were through a custom program to read Artisan serial commands and translate into duty cycle and damper setting via Phidget devices. Now I'll use Artisan to directly control those devices.

My general roast approach is a 10 -11 minute roast, ala Rao mostly. Without the mods, you get a roughly flat RoR til drop, even with airflow adjustments. I start with closed airflow (0%) at charge, and then go to near full power or full power at 30-60s into roast depending on profile and charge weight. After turning point I start increasing airflow to about 40% at dry point, and then start decreasing duty cycle and increasing airflow to 100% by FC with duty cycle relative to ambient air temp, charge weight, etc. This yields a consistent and reproducible profile without crash and flick for me. If I couldn't decrease heat input via duty cycle I don't think I could get a steadily declining roast profile.

I couldn't control the crash and flick until I added an inline exhaust fan to increase airflow through the roaster. Now I can hold more tension between heat removed and heat input so the beans less effect the temperature inside the drum. I also remove any gases emitted by the beans more quickly, no idea if this has any impact on the roast at all. Now I can control the crash and the flick pretty easily, but my roasts are as mentioned before, a touch roasty and some a touch underdeveloped, even with 25% DTR. I have even increased roast time to around 13-14 minutes without conclusive changes.

I have not experimented with drum speed control, but I have not noticed charge weight to make much of a difference in the roast profile except for charge temp. In other words, if I adjust BT to compensate for charge weight changes, and I hit dry at 4:30 into the roast and the profile is the same from then on, I get identical roasts. This leads me to believe that drum speed wouldn't have enough effect to mess with controlling it during the roast.

Any thoughts, input, comments, questions, concerns or just general distaste for me and/or my writing style are welcome. Please feel free to share anything, I have thick skin and I think discourse helps folks like me who would really like to be good roasters but don't have the time to test every single idea.

Thanks in advance!

Very well written! It appears you've gotten a repeatable degree of control over your profiles, no easy task.

I ran one profile on this COE winning Colombian washed Geisha about 10 days ago and feel it's a tad undeveloped. There's likely more oomph left to be found. I purposely kept the post 1st crack development time to a minimum which is common for a light floral geisha. But I'll hit it again with a bit more heat up front and see what pans out.

The reason I mention it is we often dial in a profile based on bean morphology and traits like density, moisture content, previous experience, etc. But my preposition is that while one profile might be the best profile based on previous experience, or utilizing the above info, it is not always the case and might not yield the best cup. That's where experimentation comes in and breaking down some of the rules we generally follow. My previous experience has taught me that most geishas with their wider open seams and longer tapered shape require gentler heat and slightly more time, since very few would risk letting the beans develop much beyond the onset of first crack, lest they loose those delicate jasmine notes and other bright features in exchange for roast flavors, there's not a lot of segments to play with. You're not likely to just randomly add 2.5 minutes to your drying phase, etc.

So I think we're left with a belief that tells us to experiment more that we're used to. If one could shape a profile that defies norms and yet is still absent any roast defects (very important), then it should help us find the tastiest cup out of what we're roasting, rather than following someone else's suggestions of what to avoid.

I will follow this thread with great interest!

I have the same roaster and absolutely recognize your flat RoR / airflow adjustment statements. I typically charge 830grams at 190degC, leave the exhaust at 50% throughout the roast and switch off the heater right before FC. That way I get somewhat of a declining RoR without crashes. My roasts take approx. 14 mins.

Last week I tried two extremes in terms of airflow: 1) Start with a closed valve, then open up before FC vs. 2) Start with an open valve and the close before FC. I could not really taste the difference in the cup! This lead me to believe that I'd better use charge temp and bean mass to control the roast, rather than fiddling with the air valve.

What is your typical charge temperature and mass?
I would be very interested to see your SSR mod and some of your profiles.
Will post some of mine too

Rush wrote: I couldn't control the crash and flick until I added an inline exhaust fan to increase airflow through the roaster.

I would suggest trying to manage the crash/flick with heat rather than air. Pulling lots of air through a roaster can trick the TC into lower readings due to the high airflow around the beans. Airflow should be used only as a last resort to stop an imminent flick, not as standard practice.

+1. Also too-high airflow can strip some aromatic notes out of a roast, resulting in a duller, less nuanced cup of coffee.

I agree with the "don't touch the air" approach. All of the roasts where I have just barely flicked the air up in the last part of FC induced a plateau in the RoR before declining again. Like Mark said a well, the resulting cup was dryer and duller than others. Since then I have had no issues getting a controllable FC dip with no flicks by only adjusting the gas and leaving the air at the same setting the whole roast.

Here is the example with the tiny flick of air at the end (5V increase), the RoR flattens for about 5s then goes down which I think that just made things worse. Obviously there were other issues with this roast but the air at the end was one thing I stopped doing after that batch.



When I left the air the same and just tried controlling the RoR with heat, things were easier to predict and my FC's have been consistent since then.

FWIW, I don't actually advocate "set air and forget it." I just prefer moderate changes that avoid the extremes. I generally start my roasts at about 60% fan power and bump it up to about 85% at BT=300F or so. The exact numbers will vary quite a bit based on fan power and the design of your toaster though, so you really need to experiment with different settings and cup the results to zero in on what works best on your own equipment.

I see, my RoR usually ramps hard for a bit with air increases during any point in the roast which annoys me so I just give it one ramp after a 1min, no-air w/low-heat soak, post charge. The convective shock is the only major ramp showing in my RoR. However I have never tried altering my settings pre-DE, I just have a routine that has been working for me to try and alter time in FC to get a desired weight loss. I cup all my roasts expect my "warm-up" roasts, those beans are blegh.
I have been wondering what changing my pre-DE routine would do though...

Almico wrote:I would suggest trying to manage the crash/flick with heat rather than air. Pulling lots of air through a roaster can trick the TC into lower readings due to the high airflow around the beans. Airflow should be used only as a last resort to stop an imminent flick, not as standard practice.

I tried literally 353 times to do so, couldn't do it. If you want to show me how I'd gladly pay any reasonable fee for that pleasure!

Went ahead and roasted a few other batches to see what came of the changes I made pre-DE. Overall, both roasts were faster and seemed like they could've gone longer into DT without any flick, even with another slight fan increase. That wasn't the goal but the concept of having more room to develop if the coffee needs it is always nice. The pace and feel of the first roast was nice, and the air settings I think make more sense (20V fan @ charge-->30V @ TP-->35V @ 300F until DROP or late FC).



2nd roast is a SM Rwandan I have and was a little more choppy at the beginning and not as hot but the basic principles stayed the same. I did one less air adjustment (30V fan @ charge-->35V @ 300F until DROP). The dip in RoR pre-DE was eminent but after that things went ok. The slight flick-up pre-FC doesn't seem like a big deal but will cup that to see. The beans are nice and brittle too so it seems the DT wasn't too short (12% moisture loss on another batch was like drinking green tea in a bad, underdeveloped way). In case anyone is wondering, rough calculations for what the voltage of my fan equates to in total % of fan power: 20V=15%, 30V=23%, 35V=27% (40V=31%) etc.



Seems some low airflow @ charge going into TP does a good job taking some more moisture out of the beans early which makes sense. That seems to take care of the plateau in the 1st roast with no signs of a flick and with a higher MET in the 2nd roast I would wager the plateau would be nil as well. Lastly, when you compare the affects in the second roast to a different roast of the same Rwandan I did a few days ago, you can really see the difference leading into FC: