Coffee Data Science
A Critique of Radial Uniformity Experiments in Espresso
Looking back at some data and exploring images
Back in 2020 and 2021, Stéphane Ribes did two radial uniformity experiments, which found that the edges of the pucks were under-extracted. From these results, he suggested an expansion of the hole area for the baskets he was using. His data did not match my data on where coffee extract, and I figured it was due to the differences of shot styles.
However, I have decided to take another look as I was reviewing all of his work recently made public by Decent. I think his experimental design missed a key finding in his own data: the very side edge is full extracted. I believe he missed seeing the evidence in the puck itself.
Experimental Design
He used two pastry circles to cut spent pucks to isolate the center, middle, and outer edge.
Afterwards, he added water to each and measured the TDS. From this, he calculated how much was extracted from each part and plotted them. He assumed the average extraction of the coffee output was the baseline from which to subtract the measured TDS left over in the grounds. I think this reasoning is sound with the assumption that solubles don’t move laterally in the puck. A shot splitter with the same radial cuts would help resolve this issue.
He measured the remaining TDS by adding a known volume of hot water and measuring the brew after it has settled.
My Previous Data
I used a different protocol called grounds TDS (gTDS) that measured the TDS of coffee grounds with the addition of some water directly on the refractometer. I looked at one puck on a square grid and found the gTDS fell off on the edge indicating less solubles were there.
So these results are different than Stéphane’s results, but I used a different machine and different equipment.
However, I did have some espresso pucks of very fine coffee where I could see when there is channeling and coffee solubles are stuck, the coffee grounds stay dark even up the side of the filter basket.
Bottom of My Espresso Shots
I take a lot of images when I’m pulling shots, and I have quite a few of the bottom of the puck. So I will start with some of these to help give context for what I’m seeing in Stéphane’s photos.
This shot was on a Kim Express single basket which has a tapered edge. There was channeling through the center, and a dark ring was left outside of the primary hole area. Outside of the dark ring is a light ring indicating that the coffee at the very edge was extracted similar to the center.
This next shot was on a Wafo Spirit basket, and there was a darker ring around the edge that would be consistent with what Stéphane found, but at the very edge, there is a light ring. This is lighter than any other area of the puck, and it indicates side channeling by the filter wall.
Now, let’s talk about dark spots. When I have done staccato shots with finer coffee coffee on the bottom and coarser on top, I have seen some streaks of solubles in the puck. This image is a good example. They aren’t wide or deep, but it is clear that if some of the side of the basket isn’t extracting, dark spots are left over and go up the sides.
For a regular shot, I have seen this phenomena occasionally. This shot showed a great example of darker spots going up the side.
Stéphane’s Data
With this in mind, let’s look at the images from his work in 2021.
First, the puck isn’t coming out so easily, but you can see there is a non-continuous dark ring, but the edge is not dark.
Another image is harder to see because it didn’t come out of the basket easily.
In another clean example, the pattern is similar. This is particularly true up the edge of the basket.
Analysis
I think this experiment could be better in three ways:
- Cut the very edge because the solubles left are confusing the very edge of the basket and the ring where there aren’t holes.
- Cut the puck vertically in halve to better confirm what is being extracted and where vertically there are still solubles.
- Examine performance at multiple brew ratios (1:1, 2:1, and 3:1).
When I have cut the puck, I found the top is mostly extracted, and the majority of solubles are stuck in the bottom.
I suspect more samples radially for these pucks would look like the blue dotted line in this modified figure.
These experiments are repeatable with such observations to verify the evidence I presented. Before I went down that path of experimentation, I wanted to have this discussion using Stéphane’s data. I’m glad he performed this experiment, and I hope others will continue to experiment because there is always more to learn.
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.
