SJM wrote: Can you elaborate just a bit on this, I don't quite understand.

Hi.

The real BT profile starts at room temperature, not at ~340F. The BT profile on the graph comes from reading the sensor. If you start logging but you don't charge the beans, the BT reading shows ~340F degrees. However, common sense tells that the beans are at room temp. Once you charge the beans in the roaster, there is a temperature drop from the meter. In your case, it goes down to 200F. However, you know that the beans are at room temp and that it takes time to catch up (warm up). With time, the difference between real and false gets smaller. All roast profiles done with meters have initial false readings because of this.

Cheers

Ah, of course, and I do understand about the false reading for the BT curve.
For some reason I thought you were referring to the Delta BT curve. ((

Thank you.

The initial deltaBT values of a profile are also false. However, ET values should be ~OK. Also, some roasters have the BT probe just above the beans, which would get somewhat less initial false readings as long as the person doing the roast is aware of the location of the BT probe and its practical interpretation.

Cheers

Hi.

To show what I just mentioned before, I created a quick example with Designer



Then I did an ln() function (in black) in Extras/Plotter to simulate a "best" theoretical BT curve with steadily decreasing deltaBT. The function I did to approximate the BT curve made with Designer was:
These are the descriptions of the values/variables I used.

x = seconds in the graph (time)
e = base for natural logarithm
100 = multiplication factor
0.19 = height of ln()
50 = time shifting of curve (last to do)

I could have made it better if I did more fine tuning. Also, I were to approximate a curve in C instead of F units, the constants would change.

Once you approximate a previous plot in artisan, you can convert it to a background plot. Then you can try a new roast and make your roaster match the ln() curve and see if it gives you a "best" roast.

Cheers

Prescott CR wrote: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.

This is precisely what I've found to be the case on my Quest, and is why I very often *add* heat to the roast about 10C prior to onset of first crack.

[creative nickname] wrote:So, I did my first experiment with this today. I'll post again if my conclusions change, but here is what I've got so far.

I did two profiles of the Kirinyaga Karimikui AA, which I tried to make as nearly identical as possible, except that one was flicked and the other was not.

So happy you did this, and i hope others follow suit. Having a genuine opinion based on you own experiences, so much more valuable than regurgitating what someone else said. And now you know on your set up, a flicked roast isnt really going to make or break the profile.

I'm having trouble wrapping my mind around this supposed issue from a profile versus a technique perspective. Does this issue with the flick surround the problem folks have with finessing their drum roasters post-1st crack or is there a problem with the flick when examining/evaluating the profile itself? As far as profiling is concerned, I have to admit that my flicks are there more or less on purpose, since I've roughly followed Jim's loose guidelines from yore for most (not all, of course) of my profiles: Steady drying; brisk to 1st; slow during 1st; brisk to finish approach. A transition from a slow 1st to (a more or less) brisk finish is basically the definition of a flick! Is this inherently an issue (at all) from a profiling perspective in a general sense? Removing the flick basically means that the 1st crack phase and post-1st development phase have the exact same ROR... and this is supposed to be a good thing?

On the other hand, if this is actually some kind drum roasting technique issue, then, if you're good at the gas and fan dial, you can finesse your way to a good roast with a flick profile in place. Luckily, I don't have to worry about this ever since I successfully PIDed my North. I have the PID track ET instead than BT. There are numerous advantages to doing this, but from this context, it does keep large temp swings in check, which can typically add to roasting defects.

As can be seen from the profile below from my last 2-lb batch roast, the flick is there as a consequence of my flattening ET curve (the BT curve is indirect - no forcing the BT curve outright into a flick). Note that during this particular roast, I was playing with PID tuning parameters and stressing the algorithm during this roast (e.g. playing with the "I" parameter mid roast caused the PID to loose its integral history at the 5-min mark and pushing the fan a bit much at 1st crack caused the power to cycle ) so there are two moments in the roast where ET and SV fell out of alignment . Still, the ET curve was well auto-managed overall (still need to work on some wobble, but few more iterations I should have this tuned out).



So basically I'm at the point where I can "design away" the flick on Artisan if I want, but I and some of the roasting community here need to be convinced that this is something that needs correction to begin with.

slickrock wrote:...

you need to cut the heat at 7min given everything before it. you are baking the coffee from then forward.

Perfunctory of you... I rather enjoyed drinking that Guat batch. Admitted though, I was too quick quick to the draw when a marking first crack due to a single outlier (it actually happened close to the 9 minute mark), ... the 193C temp and 3 minute 1st crack should should have given that away. But all of this is a distraction to this thread, and in particular, to the assertion of my post, which if you don't mind, I'd like to steer back to: Is the flick inherently wrong from a profiling perspective only? - that is, not taking into account roasting technique, which may less than graceful for some. Assuming one is a master at his or her roaster, Is a lower derivative 1st crack phase followed by a higher derivative finishing phase inherently a wrong roasting approach?

slickrock wrote:Is the flick inherently wrong from a profiling perspective only? - that is, not taking into account roasting technique, which may less than graceful for some. Assuming one is a master at his or her roaster, Is a lower derivative 1st crack phase followed by a higher derivative finishing phase inherently a wrong roasting approach?

The only way you can know is to try it. As for what Scott Rao and others (including myself) are saying, yes, it is inherently wrong, as it will create undesirable baked flavors in the coffee. If you're skeptical, you ought to try it yourself and come to your own conclusion, as you seem unwilling to believe those who have said as much.