Rickpatbrown wrote:There is what I usually refer to as a plateau that can proceed the crash. This can be a small rise in ROR, if bad enough. We should have a specific word for it.

Most crashes and flick are a direct result of that pre-1C plateau. I agree, it should have a name. The problem is it is not consistent across all coffees.

From my experience, some coffees do not crash not matter what I do. Some coffees crash no matter I do. Other coffees I can force to crash if too much heat is carried into 1C.

If constant heat is applied from charge through 1C there will inevitably be a rise (or plateau) in RoR just before 1C. Then the release of energy at 1C causes the RoR to drop and...crash. Mitigating the crash involves reducing the energy in the roast environment well in advance of 1C.

Right now I only have 3 coffees that are crash prone. A decaf, a Kenya and my arch nemesis, the infamous Sumatra. The Kenya and decaf crash can be controlled with heat reduction prior to 1C. Nothing works on the Sumatra except blasting it with heat for a minute starting 15s prior to 1C. Then the trick it to reduce the heat so it doesn't flick.

Almico wrote:Most crashes and flick are a direct result of that pre-1C plateau. I agree, it should have a name. The problem is it is not consistent across all coffees.

I think of it as many (not all) tend to have a pre-first crack hump, then crash, then flick. So in my mind it's Hump-Crash-Flick.

I am not sure where I heard the term hump but that's what stuck in my head.

also for some reason the chance of having a bad plateau or hump right before first crack seems to be WAY higher in naturals and experimental processes for me, though i have no idea what causes that difference

I have not found that to be the case. I roast a lot of natural brazils, ethiopias and centrals and those seem to be the most stable.

shrug yeah could just be me inferring causation from what's really just random chance. or some other variable i can't really account for at home

Seems pretty hard to target a declining
ROR if you dont monitor it......... Somebody going to have to explain that one to me....

Actively attempting to manipulate it....is totally different from monitoring it

Yeah, the first few minutes sets up the whole roast. And you're only going to have a few key adjustments after that point usually. But it's different for different roasters. Different amounts of mass, different amounts of heat retained. Different amounts of airflow.


I don't think anybody would argue that heavy commercial roasters with thick drums and high mass to the weight of beans being roasted dont seem more thermally stable than smaller roasters. But then again even the size of the thermocouples can affect the interpretation of what's happening....... By dampening the response


I have an impression...... Just from reading..... That some people roasting on the bullet may attempt to control their batch temperature more actively than people using conventionally heated roasters.

Not sure if ending up with declining RoR is that hard without monitoring....below graph is just that, a graph showing what happened, all I did was keep air constant, full power until 3:20 and gradual 0.5 power decreases...no monitoring involved until the very end

I bet a dollar that you learn to make those incremental reductions by monitoring it as you made them......... You probably didn't decide on them , or the initial heat, or charge, or airflow ......blindly.

Once reasonably dialed in, yeah you could do without it. In fact a few times I've started artisan late (due to stupidity ) after I dropped beans.....for some reason artisan didn't pick up the ROR at all.... Had to run the batch without it. No estimate of first crack , development, etc. Seems to be some kind of bug in my version. Like flying blind. Much prefer having the ROR and the estimated times etc

Mbb wrote:I bet a dollar that you learn to make those incremental reductions by monitoring it as you made them......... You didn't decide on them , or the initial heat, or charge, or airflow ......blindly

Sure, that whole 'protocol' was made up using empirical data a.o. from the RTDs, and delta BT, BUT, the result is that it appears that (using a Huky) loading power forward and methodically tapering off power works to end up with a steady declining RoR without monitoring delta BT, that knowledge can probably be transferred to other Huky users.

Mbb wrote: In fact a few times I've started artisan late (due to stupidity ) after I dropped beans.....for some reason artisan didn't pick up the ROR at all.... Had to run the batch without it. No estimate of first crack , development, etc. Seems to be some kind of bug in my version.

Not a bug, but an FAQ: I can not see the rate-of-rise curve during roasting?

Same for all versions. RoR is only drawn after the CHARGE event has been set.