How to find the perfect roast for new green coffee



What I would like is to attend a day at Ikawa and learn everything that is possible for a home customer to learn about roasting in 4-5 hours. Any chance? And would anyone else like this?


CRACK and more CRACK!
As a novice roaster, I have been afraid to ask this question because “the first crack” seems to be such and obvious moment in all I have read. However, in my limited experience, there is usually a resounding CRACK followed throughout the next 30 or so second with other cracks, some loud, others soft. Am I correct in thinking that a crack is made by a single bean? Therefore one would expect dozens of cracks during a batch roast? Kind of like popcorn? Since we are advised to stop the roast some 30 to 60 seconds after “the first crack,” is that literally the first resounding crack, or is “the first crack” to be conceived of as a cluster of sound events?

Case: I roasted half a pound of Sweet Maria’s Guatemala Xinabajul Floridel de Pena this past weekend, using IK’s Final Puerta Verde roast. The first resounding crack was consistently around 6.07, but vigorous cracks continued through to 6.50 when the profile initiates cooling. Sometimes I heard a few cracks after the cooling started. Is this unusual?

Taste note that may answer my own question: I am drinking some now (premature just 2 days after roast, but I have a hard time keeping up with my roasting…). It tastes a little underdeveloped, but I expect it to be better tomorrow.


There is a discussion of this on the “phases of roast” thread.
What you are describing is normal. The period between the first and last crack is variable but seems to me to average around 45 secs. Had one at 15 and 1m 15 and also some cracking just when cooling starts. The amount of cracks is also variable. One bean I only heard 2 but they had all cracked. Another can be a veritable cacophony!
No need to worry!


Mine is wonder, not worry. I am trying to figure out from what moment do I mark “the first crack” and then begin the ideal initiation of the cool-down phase.

The question: Is “the first crack” a moment or an imprecise range of time over some seconds?


@cooley hello :wink:
Well there are two things mixed together in the discussions, one is The 1st Crack, and another one is the first crack of the 1st Crack :slight_smile:
The first pop of one bean is the unimportant first crack of the more important 1st Crack as a whole continuous time when the beans pop for the first time (we usually measure the whole time span - I note the begining and the end, either including the very first and last pop or excluding them to cut away extremes). Then there is also a Second Crack but I have never experienced it as I do not roast that dark.


Well, my understanding is that is a phase or range of time and happens when the beans reach a particular temperature. Due to the beans being individual and perhaps the roast profile not all the beans crack or reach that temperature at exactly the same time.
Many seem to consider the first crack as an outlier and regard 1C starting when the first few go off at the same time.
At the moment I’m measuring it as from very first crack, as any other way seems inconsistent. Given the small sample size there will be less cracks heard I’d think than from more usual amounts. Also, tracking software will place 1C on a graph, so it must be working on a specific temperature?


It’s still marked by the person where first crack starts.

Also you mentioned seeing all the cracks. Those cracks are a form of tipping and not where the cracking sound comes from. You have to slice the bean horizontally to see if the bean cracked.


good question @cooley, it is ambiguous! Scott Rao recommends marking the beginning of 1st crack from the 1st sound of beans cracking, which would begin the development time. The exception would be when there is a weird outlier, the lone bean that cracks 30 seconds before all the rest for example. That’s safe to ignore.


Determining the end of 1st crack can be just as tricky. I determine the beginning and end slightly differently each time depending on outliers and when the greatest number of cracks occur. The hardest thing to get used to was the small number of overall pops.


Well … for me a my tiny little OCD, its very uncomfortable that the starts and ends of the 1stC are so fuzzy, and dependent on the beans that are slowest/fastest (so the farthest away from the average bean there). In the same line of thinking that brought me to using iphone camera to track colour changes, I think a microphone on the same device (audio from the same captured video) can be used to filter out the pops and do some automagic math to find the centerpoint. That should be representative of the most average beans reaching the 1stC and also be the most consistent. Its also not extremely important for every one of the roasts, once a profile is set and used repeatably I would probably not repeat all the measurements unless done automagicaly with the roast :)) wink wink …:)) but when experimenting and sharing the outcomes of that experiments, with other that may have a different perception or method to observe the same … it would be very nice to have some less subjective and less fuzzy method.
Just my little though about this …