Coffee Data Science

Refractometer Accuracy for Coffee and Espresso

An exploration in measurement

Robert McKeon Aloe
Nerd For Tech


Previous experiments have shown interesting results around 1% Total Dissolved Solids (TDS), so I wanted to test a bit more. I want to understand how different parts of the espresso shot read differently, so I pulled a salami shot and started testing.

All images by Author

My initial question: if I dilute samples from the first 1:1 of a shot vs the second, do they have similar trends in TDS reading? This gives some indication to what causes refraction changes.

For each sample, I added a known amount of water so based on the initial TDS reading, it should scale proportional. Groundtruth was based on the first sample. The original samples were brought to room temperature before testing.

The trend lines are almost proportional, but let’s zoom-in.

We can see here that the first portion of the shot has a lot of changes when diluting vs the second and third portion of the shot. We can see this is more pronounced when going to a log-log plot:

If we look at the measurement error, the percent error greatly increases around a TDS of 1.

I suspect the first portion of the shot might have issues due to oils, but I’m not sure. This test would be interesting to repeat with filters as well as comparing to other coffees. The results make me question the general accuracy of TDS for coffee when it’s less than 1. More data is required, and it is a nice reminder that the better we understand our instrumentation, the better we can make use of it to make good decisions for improving coffee.

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 Future Book

My Links

Collection of Espresso Articles

A Collection of Work and School Stories



Robert McKeon Aloe
Nerd For Tech

I’m in love with my Wife, my Kids, Espresso, Data Science, tomatoes, cooking, engineering, talking, family, Paris, and Italy, not necessarily in that order.