Perhaps I wasn't clear in the above. I noticed the difference after tasting the cups unblinded. After that, I ran three more blinded triangular tasting trials, during which I correctly identified the flicked cup in two out of the three tests.

Okay, thanks for the clarification. Would be interesting to see if others also notice any differences between flicked and unflicked roasts.

The only time I've noticed the flick is when going a good bit past first crack and close to or into second crack. So if one is roasting lighter then flicking is not so much of a concern, correct?

If I don't take proactive measures to prevent it, I often see an uptick in ROR just after first cracks are dying down. I wouldn't worry about the flick for Nordic roasts, but it can easily arise (at least in my experience) in roasts where you are aiming for a city or city-plus finish.

Proactive measures? I turn off the gas but beyond that what can you do? When you see the "flick" it's too late.

I notice after about 90 sec post C1 need to start turning the fan up and at about 30 sec post 1C turn the gas down..

I started following info in Bold Java's video and slowly turning gas down and air up as roast progresses..
I wish they would go over the roasting theory with out roasting and not streaming so you could hear it well..
I'd like to know Time or Temps when is the best places to make adjustments..
I remember they said make adjustments before changes but not sure when the changes happen because they said they were playing catch up because they turned the gas on late..

thusband wrote:Proactive measures? I turn off the gas but beyond that what can you do? When you see the "flick" it's too late.

Well all things are relative; even if you start to see a flick, you might be able to stop it from accelerating too much by cutting gas and increasing air. Presumably larger and more sustained flicks produce larger variations in taste.

But by proactive, I mean anticipating a potential flick ahead of time. How far ahead will depend on your equipment. If you have a perforated drum and lots of airflow you may be able to turn on a dime; my USRC has enough thermal mass that for a typical charge I need to plan ahead by 45 to 60s. So to avoid a flick, I'll usually start progressively cutting gas in small increments starting around 45 seconds into first crack. Sometimes I go all the way down to 0", sometimes less, depending on how much momentum the roast had entering first crack.

As you can see from the graphs I posted, the timing is quite important. I generated my "controlled flick" by cutting gas too soon, but not quite far enough, resulting in an ROR dip during first followed by a rebound.

Thanks for that, creative nickname. I guess it's practice, practice, practice. I have a solid drum Huky so I think I'll try heading into first crack with a little more momentum and then cutting way back on heat a little earlier while increasing the air.

The methods of managing the heat in a roaster in order to get desired results are going to vary wildly between different types of roaster. My IR12 at 5000' elevation is going to be way different than a Probat L5 at 1000'.

Add to that how you approach FC and load amount and it's all over the place.

So it's best to find someone who has the same roaster you do and understand how s/he gets to FC and what comes next. There's also the coffee as a variable too!

From what I've see the flick might also be a symptom of the cliff before it. Where ROR dies quickly then bounces back (most likely from me trying to stop the drop). When I get enough heat at the right time into the beans to prevent the drop I can cut the heat and coast with a much better ROR curve. The challenge here is to no charge too much heat in the system that the ROR stops declining.

Thank god for data logging, I can't imagine trying to conquer all this without it.

Hi.

A decreasing deltaBT would theoretically approximate a natural logarithmic curve since ln() function has a positive slope that decreases steadily as time increases. In artisan, you can "see" the slope (deltaBT) with the linear projection enabled. Also, you can see the numerical value on the LCD (degrees per minute units). That would be the geometrical interpretation of the slope (graph). Then the physical interpretation (ie. behaviour of the beans in the roaster) is the speed of BT just like for example the speed formula on physics.

distance X
velocity = X/Time

BT curve = BT profile
deltaBT = derivative of BT curve (or slope) = Delta-Temperature/Delta-Time (where DeltaTime = sample period)
NOTE: the BT profile has an initial drop from the sensor that is not real as the bean temperature starts at room temperature (ie. not at 300C)

Then,

*best theoretical bean temperature curve BT = ~ln(bx+c) would need a multiplication factor too
*best theoretical deltaBT= derivativeBT = d/dx(ln(bx + c)) = b/(bx + c)
*best theoretical second derivative (deceleration) of BT = derivative of deltaBT = d/dx(b/(bx+c) = -b^2/(bx+c)^2
NOTE: x = Total increasing time; b and c = constants
* an idea that could not be true for all cases and which depends on other factors and experimentation

Cheers

NOTE: the BT profile has an initial drop from the sensor that is not real as the bean temperature starts at room temperature (ie. not at 300C)

Can you elaborate just a bit on this, I don't quite understand.
Here is a profile that maybe you could use to explain if you would.


What is real and what isn't real for the ROR?

Thank you