Coffee Data Science
Puck Mesh Screens for Espresso
Another look at other data
As you may know, I love data, and some new data has been collected on puck screens. A puck screen can be put on top of the coffee with the hope that it could improve coffee extraction. I have previously had issues with pucks screens because of temperature differences between the puck screen and the machine, so I stopped using them a year ago. However, not everyone pulls espresso like me, and I appreciate data wherever I see it.
This data was collected by Lance Hedrick for shots at a 2.5:1 ratio (20g of coffee in, 50g out). I usually pull shots at a 1:1 ratio (20g in, 20g out). For my own sake, I would have to retest given that I don’t know what the effects were for shorter shot ratios. In summary, the data shows a mesh on top of the puck improved extraction by 0.4% on average, and this value was statistically significant.
Definitions
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). Typically, one aims for 18% to 22% extraction or some times higher, but it is difficult to get more than 30% EY.
Intensity Radius (IR) is defined as the radius from the origin on a control chart for TDS vs EY, so IR = sqrt( TDS² + EY²). This metric helps normalize shot performance across output yield or brew ratio.
Data
Six sets of data were collected on the same coffee across 10 shots:
- Baseline
- Billet screen on top
- Paper filter on bottom (Bottom Paper)
- Paper filter on top (Top Paper)
- Dry mesh on top (Top Dry Mesh)
- Wet mesh on top (Top Wet Mesh)
While the shots were collected randomly on a Unica, I sorted them by IR to compare the best shots with the best shots. There were slight differences in output yield, so sorting by IR seemed more fair than using EY.
In the raw data, Lance used two different grind settings. I wanted to first make sure those didn’t introduce extra variables, so I compared the two settings controlling for whatever was put on top/bottom of the puck.
I also broke this down by the variants because the pattern seemed odd to me. The most notable being the Billet Screen and the Top Wet Mesh that are consistent. There isn’t enough data to really understand this, but it is an interesting piece of information in this dataset as it might reveal something else.
Now let’s compare mesh variants holding grind setting constant between the baseline and the variant.
The Top Mesh mesh seems to be winning out, so let’s remove the paper variants:
I noticed some differences in output yield, so I plotted IR to double check, and that showed the same pattern. The paper on the bottom was all over the place.
At a statistical level, the Top Dry Mesh and Top Wet Mesh had improvements in EY and IR that were statistically significant using a paired two-tailed t-test. Interestingly enough, the output yield for the Top Dry Mesh was statistically lower than the baseline.
These results were interesting. My main push is to use a lower output yield. While there were statistically significant improvements, do those warrant an extra step? I’m not sure, and for my use-case, I’m not sure I would see much improvement.
My other concern is that paper filters on the bottom generally improve performance. Is there a profile where a bottom paper filter improves over baseline and the top doesn’t? I’m not sure, and the variable space in espresso is very large so it is hard to explore all variables across all combinations.
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
