#sieving coffee

[boulder sieve, macro]

Today I’ll be cobbling together some observations from a few months’ worth of experimentation. Now that makes it sound too complicated; really, I mean I’ll be speculating based on messing about with some weird coffee brewing techniques.

First off, I tried sieving boulders. I started with a 1200 micron (1.2mm) sieve. This got rid of less than 10% of ground mass and mostly caught chaff and oblong grounds. ("Larger” oblong particles can make it through the aperture of a flat burrset if you think about the geometry of the grinding process - two discs next to each other.) I mostly used a V60 but did experiment with a Chemex and a Wave. 

The effect wasn’t terribly subtle. Sweetness increased noticeably, but I also noticed a substantial increase in a bitter, ashy, astringent quality. I also lost some of the acidity I look for. Next, I ordered a 1000 micron filter. And the results here were even worse: loss of much more acidity, and even more ashiness. I can see the 1200 being useful in some cases, but not terribly so. (Notice we’re increasing the proportion, by weight, of fines relative to our dose of ground coffee.)

Honestly, the experience made me want to try a fines sieve, but that will have to wait for another day, since I haven’t purchased one yet.

Today I first experimented with an approach similar in idea but different in implementation: sorting whole coffee by screen size. I was working with a coffee from the Guji zone of Ethiopia and eliminated both the largest (18) and smallest (<15, I think) beans. The idea here is designed to, again, increase solubility and sweetness. 

The smallest beans might tend to be “overdeveloped,” by which I mean that the process of development has been taken to excess, not that they are simply “too dark.” That is, I mean it literally - they are too soluble, they’ve lost acidity, and they’ve probably picked up too much roast aroma. This is a tendency I’ve noticed roasting really small beans, e.g. the moka variety and some Ethiopian lots. I also removed the largest beans. Here the logic is a little more obvious. These probably didn’t have a chance to develop as fully as desired - in terms of solubility and sweetness, which I consider to be essentially the same property. 

And here the result was a bit better. Sweetness increased noticeably. The coffee also took on a noticeably creamy and thick mouthfeel... really surprisingly not tea-like for an Ethiopian coffee. Overall, I found it improved, although some others didn’t agree. It’s worth noting that I care more about sweetness than many who work in specialty coffee, so you may or may not find this process useful. When I compared the brew to others, I did notice a loss of acidity - only mild, but sufficient to shift the fruits from citrus-forward to berry-forward. 

I wanted to post about these two experiments together because I think there’s a pretty interesting parallel. Namely, increasing evenness through these techniques is gaining us sweetness but costing us acidity and complexity. The concept of complexity has taken a lot of criticism from the cutting edge of the Third Wave scene. And I think this is perfectly reasonable, because it is so horribly nebulous and because it can be used to justify bad technique. But that doesn’t mean we should throw it out entirely. And if complexity really is too ill-defined, then just focus on that balance of acidity and sweetness. Yes, it’s far easier to make a beverage too bright than too sweet, but at least in the privacy of your own well-supplied coffee lab and roastery, such a concept as too even of an extraction may be possible, at least for some palates. 

Obviously, this is a really limited sample. We brewed up the coffee twice with this technique and compared to a few other brews made with the same recipe and unsorted coffee.