How to Roast Coffee at Home, Chapter 4: Why do my temperature readings jump around during a roast?


If you read Chapter 1 and decided to retrofit your own home coffee roaster with a probe to measure the temperature of the green coffee beans during the roast, you may have noticed the temperature jumping around a bit. 

What's going on, and does it matter?

The Short Answer

What's going on?

During any given roast, the probe is measuring more than just the roasting coffee beans. As green coffee beans tumble in the roaster, at any given moment the probe could be touching:

  • One or more coffee beans
  • A temporary air pocket between coffee beans
  • The drum itself

So depending on what the probe is touching, don't be surpised to see the temperature reading jump up or down a few degrees.

Does it matter? 

Not really.

As long as you pay attention to the smell, color and sound of beans and don't overreact to a single, unexpected reading, you'll roast some delicous coffee.

The Long Answer

What's going on?

Second-to-second variability in temperature readings happens in all coffee roasters, both home and commercial. The root cause is the same. At any given instant, the probe could be touching none, one or several coffee beans.

In addition, each bean within the batch will be at a slightly different temperature at the same point during a roast. This could be due to a particular bean's size, shape, moisture content. It could also be a function of where exactly the bean was in the bean mass in the few moments before it touched the tip of the probe. 

Was the bean hugging the drum wall? If so, that bean will be a little hotter. Was that bean shielded from the heating element by it batch mates? In that case, it would be a little cooler. 

A similar thing happens on Thanksgiving when you check the turkey with a meat thermometer. Barely insert the probe? Higher temperature. Push a little deeper, lower temperature. Poke through the rib cage and into the stuffing? A very different temperature.  

Commercial coffee roasters are equipped with software that averages out these small variations, producing a nice, smooth curve. (Being the geeks we are, we just capture the raw data in a CSV file then analyze the roast in spreadsheets.)

Does it matter?

No...provided that you are paying attention to the overall trend and react accordingly. For example, If you see a reading of 450 F, but you're still in the middle of first crack, don't stop the roast prematurely. Almost certainly, the temperature will drop in a few seconds. 

In other words, think of the temperature as a compass, not GPS map. It will tell you the general direction you're headed, but not when the next turn is.