I do believe IBTS may be more suitable for small batches, but maybe due to my previous experience with larger roaster (I have worked on Giesen 15kg and others), I hit the YP and the FC almost around same temperatures as the bullet when I use the Bean Temp.
So both of those recipes failed in that own ways, the P9 caused an error code and stopped the roast. Had plastic melty smells during the preheat. (to be fair, sometimes going with P8 too long trips an error, but P7 has always been fine, i’m on 120v model)
The other recipe, my machine did very little to push the roast along. preheat and started out looking good, but fell behind, and eventually got to a negative RoR.
I don’t think it is fault of the recipes, just that some setups are different than others.
There was a recipe from the 350g thread that played out really well.
You might want to try cleaning the glass on your IBTS. I suspect that will fix your readings. Typically when you see crossover in the graph, IBTS is dirty. I clean it pretty much every other session to make sure it stays clean and accurate. Since the machine relies on IBTS for preheat temps, it’s important it stays clean for repeatability and consistency. I suspect you will also find it more accurate/reliable at measuring color later in the roast as well…
I have a weekly cleaning routine, I do it to keep my roasts consistent. I do not have any color issue eith my roasts, I can roast 3 different profiles and end up with the same color of the roasted beans (thats what you practice in SCA basic roast). IBTS suitable for small batches, mostly because they are rotating below the thermal coupler probe. But for larger batches, definitely I will prefer the BEAN Temp. I explained previously, but I will put more details to make it clear scientifically.
There are two major transmissions in roasting, Endothermic and Exothermic, when exothermic occurred, the beans release energy, that will increase the overall energy in the roasting chamber. This extra amount of energy, IBTS sensor doesn’t read it, though it will impact the roast process.
If your IBTS stayed above the Bean Temp, thats happened either you are roasting small batch or you didn’t preheat your machine enough.
Did a quick tasting this morning with a couple recipes.
Keep in mind, that even though i am taking directions from others, that my experience in no way reflects the same flavors and end result that they get, this is simply to see how well recipe’s translate.
Coffee used was a mid grade Costa Rican coffee that has a mellow apple acidity and a slight baked spice and subtle chocolate finish typically. But it has aged out a bit, and lost some of that sparkle.
350g by @jondemir
“Hoosian” i tried something loosely based on some things i had seen on Rob Hoos’ profile. I dare call it ‘Hoosian’, but for lack of a better descriptor at this time.
The third roast was something somewhat more generic in terms of what i’ve seen being common approaches lately. moderate charge around 205C, moderate fans to start, and stepping up in fan and down in power as the roast goes. 1.5min dev.
Brewed a smaller 30g in Chemex, 1:15 ratio with a 3.5minute target.
The cup impressions:
Jp’s was fairly solid. This roast had a more muted flavors, which considering how the roast began to fade in energy towards the end, as i let it ride out, that makes sense. I assume my fan setting or power settings are effecting the overall energy of the roast, and i might push it a little more. That said, it was pretty great how well his recipe translated to my roaster, and how well it brewed.
hoosian brewed about 15seconds faster than the other two roasts, i brewed it twice, and both times it was pretty much exactly 3min15seconds. At this point my palette is over worked, so i’m not going to redo it. Something to note anyway. I felt like this one was the cleanest overall flavor, and brightest acidity (without being harsh). I am assuming the high fans were the reason for this. EDIT: as it cooled to roomtemp, i felt it was a little unbalanced, perhaps a combination of too high of fan and a short dev. It had some unevenness looking in the bean, however on the exterior it seemed more even than the others. That said, was clean and bright, and shows potential.
Generic approach, while i don’t think this was the best version of the Costa Rican coffee, i sort of felt it was the best of the three. a bit more depth and rounded acidity than the hoosian hi-fan test of 2. This one may have also had the most even look, cracking open a few beans from each roast. i kept the fans high enough to promote even roasting, and went a but higher in temp then the other two. so, not like all things were equal. But certainly was not as clean tasting as the ultra hi fan hoosian concept. just goes to show that the conventional wisdom on this forum is on the right track, perhaps.
Depending on the coffee or mood, anyone of these could be preferable to the other. I find the hoosian approach very interesting and may try it on something more fresh and bright.
i feel that Jp’s 350g recipe could be a great starter toast because of how well it translated. With a few small tweaks, it could work perfectly on my bullet. that’s pretty amazing. i’ll add his to the list of small batch recipe up top of this thread. Thanks @jpndemir !
Let me clarify that the recipe I suggested Is just a variation of Morten Munchows “get to know your roaster”-thingy
I found that I could roast pretty much hands off(when I had a computer) and get results I wanted by just tweaking the recipe here and there after cupping. I use the same strategy for all batch sizes and usually fiddle with fan settings if roasts get too “cloudy” or “smoky”.
I let the beans ride, tweak for next roast and let the beans ride again.
Currently everything is happening manually but here are 3 roasts where I tried a new bean for the first time and had access to a computer:
I think the most dramatic shift in how I viewed the bullet, was seeing the Sweet Maria’s roasts. Made me rethink what is going on, and what is possible.
But the main impetus for reviving this thread is the Andrew Coe curve. He seems to have embraced the more conventional approach, which is nice to see. After so many weird curves and experiments, it is nice to see a grounded approach that looks logical.