Coffee Data Science

More Thermal Pulsing Variants for Coffee Roasting

Diving deeper

4 min readJun 14, 2024

After stumbling onto thermal pulsing for coffee roasting, I thought I should better know a roast. I previously explored the number of pulses and the temperature offset. Now I wanted to explore when pulsing really matters. To do this, I did some partial pulsing, and then I looked at raising or lowering the pulsing temperatures.

Thermal Pulsing Variants

Here is the baseline Bean Temperature/Inlet Temperature (BT/IT) profile and thermal pulsing.

I dropped the last two pulses, but it looks like I just dropped one; it just went flat.

This dropped the first two pulses, but it looks like there is an extra pulse because of the second half starting on a drop in temperature.

I moved the max peak to be anchored on the baseline. This didn’t quite work out as planned, but it was close.

I did the same, but moved the average of the pulses to be at the baseline. The original thermal pulsing was made without much exact settings to be centered around the baseline.

Lastly, I put the minimums of the peaks at the baseline, but the roast went so fast that it didn’t quite emulate what I was hoping for.

Roasts Compared

Dropping the first 2 pulses matched well to thermal pulsing (TP), and dropping the last two matched closer to the baseline BT/IT profile.

The Rate of Rise (RoR) matched as expected for TP Drop the Last 2 and Thermal Pulsing. There was a slight offset to the BT/IT baseline with Drop the First 2.

They had similar crack patterns except for drop the first two. I suspect the First two pulses are not as important to getting the results of thermal pulsing.

Moving on to where we anchor the pulsing, anchoring at the min causes a faster roast, and anchoring at the max causes a much slower roast.

The same can be seen in the RoR. The TP Average at Baseline matches well to the original Thermal Pulsing.

All had similar crack patterns just offset in time.

Roast Metrics

Percent of weight loss didn’t seem to be differentiating, but the TP Max and Min had similar weight loses offset from the TP Average.

Moisture wasn’t so helpful either.

All of these had similar coffee color, so I have to wonder how much the taste would be affected.

Density also didn’t say much.

While the roasts themselves were very different, the post-roast metrics were not helpful. Tasting will have to be done to see how these roast variants affected the inside of the beans.

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.