Edge extraction, "volcano" effect and better baskets

The turbo paper introduced the concept of the “volcano” effect. This is essentially a statement about the trade-off between finer grinds causing an increase in extraction yield but only to a certain extent. Going too fine causes channeling that then reduces bulk extraction yield due to uneven extraction. The paper has some really nice ideas about modeling extraction, and the discrepancy between model predictions and experimental results were attributed to channeling. The experiments in the paper used a VST basket without any bottom paper filters. For a while now, I have been wondering if that volcano effect is really a manifestation of channeling as is usually thought of (general lateral unevenness) or if it is really a manifestation of edge underextraction stemming from the blank areas without holes on espresso basket edges.

I got the chance to try out the WAFO espresso baskets recently (courtesy of Brian Quan) and thought this would be a great way to test out this hypothesis. Specifically, I used the WAFO SOE Spirit basket that has holes that go very close to the basket edge as well as being slit shaped instead of the standard circular holes.

WAFO basket

WAFO make 4 baskets. Two of them are labeled SOE for Single Origin Espresso and the other two are labeled Blend. For this post, I was mainly interested in looking at the SOE baskets since they have the edge to edge holes promising to mitigate edge underextraction. After some initial testing, the SOE Spirit basket looked to have the most even puck bottoms and the SOE Classic basket had a very tiny bit of edge discoloration of spent puck bottoms. For the rest of this post, I will only discuss the SOE Spirit basket.

Extraction measurements

The experiment here is simple. I ground and prepped pucks with different grind settings on my P100 with the Pre-2015 EK43 burr geometry manufactured by SSP. All shots were pulled on a Decent Espresso DE1 with a flat 6 bar profile. The coffee I used for this experiment was the Worka Site from SEY coffee. For all the results, I plotted burr gap instead of the grind setting using the conversion factor of 75 microns per large tick on the P100. Shots were pulled using the WAFO SOE Spirit basket and with a Pullman 17-19g basket for comparison.

Puck prep

The puck prep process I used was as follows:

1. Grind a set dose of coffee.
2. Dose into a basket, weigh the ground coffee, no bottom paper filters.
3. Deep WDT with the 4z level method as described in this video followed very light vertical tap.
4. Tamp with a Normcore V3 tamper that I modded with a 25 lb internal spring for a level and calibrated tamp.
5. Add a top Aeropress filter for cleanliness (this was not my machine). I didn’t use a Flair 58 screen since there have been some reports of the Flair screen affecting edge extraction and didn’t want to add that additional variable to this experiment.

Results

Here is the main result. The plot below shows the extraction yield as a function of grind burr gap:

We clearly see the “volcano” effect here for both baskets. Starting on the right with the coarsest grind, extraction is limited by just the exposed surface area and so the effect of the basket is not significant. As we go finer (left in the plot), more surface area is exposed causing extraction to increase. After a certain point, extraction starts dropping again because of unevenness.

Comparing the two baskets, the WAFO enables much higher extractions. The WAFO peak is almost 2.5% higher EY than the Pullman indicating that the Pullman is limited by edge underextraction. This is reflected in puck bottom images as well.

At the volcano peak (right half in the images above), the WAFO puck bottom is perfectly even but the Pullman puck has dark edges indicating edge underextraction. This edge underextraction is why the Pullman has a lower bulk extraction yield compared to the WAFO basket. At the finest grind setting (left half in the images above), both baskets show signs of unevenness in the center of the puck but the Pullman has edge underextraction here as well.

Interestingly, both baskets have the extraction peak at the same grind setting. This indicates that though the Pullman has edge underextraction, it isn’t sufficient to reduce bulk extraction.

In terms of shot time, the WAFO basket has slightly higher flow resulting in slightly lower shot times for the same target yield:

Some other thoughts

• Can pick out WAFO vs Pullman in blind triangle taste testing
  • WAFO has less harshness, better fruit quality to the acidity, more balanced and generally more enjoyable
• WAFO with vs without bottom filter, cannot pick out in blind triangle taste testing
  • Bottom filter not required with WAFO
  • Surprising that WAFO without bottom filters doesn’t have noticeably more fines in the cup presumably due to smaller holes on the WAFO
• WAFO vs Pullman with bottom filters, very very close in taste
  • Could pick out Pullman with bottom filter, but probably because the Decent stopped a few grams short with the Pullman

Conclusion

The aim of this experiment was to look at the role of edge underextraction in the “volcano” effect. The results suggest that edge underextraction is indeed the first EY limiter when going from coarse to fine, but it isn’t sufficient to cause the volcano effect itself. The reduction in EY at very fine grind stems from more broad unevenness or channeling. The other more practical conclusion is that the WAFO SOE Spirit basket does indeed do a very good job at mitigating edge underextraction. Practically, this means that one can forgo bottom paper filters without worrying about that slight harshness that comes from uneven extraction.