Coffee Data Science
Modifying the Caramelization Stage in Coffee Roasting
Nuance in the development of the roast
Previously, I explored modifying the fan speed during the Maillard phase of a roast to affect the way beans absorb heat. I had some good results, so I wanted to explore doing something similar for the end of the roast after the first crack.
I explored Fast Caramelization (Fast C) and Slow Caramelization (Slow C) using the fan to accelerate or decelerate the roast. A higher fan speed raises the bean temperature faster. All of these tests were on the Roest.
The main affect on the Rate of Rise (RoR) was at the very end. I used a baseline profile of Thermal Pulsing hence the variable RoR.
We can zoom in on the last 2 minutes, and that’s where there are major effects on the RoR. The baseline is offset from the Fast C and Slow C, but they have a similar shape.
Post-Roast Metrics
I looked at two sets of roasts: an Ethiopian bean and an India bean. In terms of weight loss and density, there were not large differences. Fast C did lose a little bit more weight.
In terms of moisture, there were not much changes. Water activity is still unknown in importance but reported.
Roast color is where the fun is, but the roasts were close enough (i.e. within 5 of the average color). The Fast C roasts tended to be lighter in color (higher Agtron).
Metrics of Performance
I used two sets of metrics for evaluating the differences between techniques: Final Score and Coffee Extraction.
Final score is the average of a scorecard of 7 metrics (Sharp, Rich, Syrup, Sweet, Sour, Bitter, and Aftertaste). These scores were subjective, of course, but they were calibrated to my tastes and helped me improve my shots. There is some variation in the scores. My aim was to be consistent for each metric, but some times the granularity was difficult.
Total Dissolved Solids (TDS) is measured using a refractometer, and this number combined with the output weight of the shot and the input weight of the coffee is used to determine the percentage of coffee extracted into the cup, called Extraction Yield (EY).
Shots
I pulled 4 shot sets per coffee bean type to provide some insight. The first shot baseline for the Ethiopian bean seems to be an outlier.
Reorienting the shot data by the roast, we can look for a trend in taste. These results aren’t large enough to be conclusive, and I didn’t see a trend. For the Ethiopian bean, all three roasts seemed equal. For the Indian bean, Fast C and Slow C were both better. I was hopeful that one would be worse in taste and one would be better.
EY didn’t yield much information.
These tests did not point to evidence that modifying the fan speed in the last minute of the roast had any effect on the taste. It is possible I should try different beans or explore modifying drum RPM as well. I was hoping for a variable that could make the roast better or worse, and these tests came back as showing neither.
If you like, follow me on Twitter, YouTube, and Instagram where I post videos of espresso shots on different machines and espresso related stuff. You can also find me on LinkedIn. You can also follow me on Medium and Subscribe.
Further readings of mine:
My Second Book: Advanced Espresso
My First Book: Engineering Better Espresso
