Curiosity got the best of me today. I set up a few very crude, unscientific experiments, that I'd like to work on some more going forward. I brewed two different coffees and looked for patterns of preference in samples with the "fines" and "boulders" reduced, versus just the "boulders" reduced hence the silly quotes to differentiate it from an objective thing. Note off the bat, I'm not declaring either "removed" just reduced. And I'm not strictly defining either entity to any strict level. This was all a rather crude test.

My first run thru wasn't blind, and my bias wasn't eliminated. My findings with my first test brew (using a Brazen brewer with the same grind setting, brew temp, 16:1 ratio throughout all tests) showed a significantly better cup in my personal opinion when the "fines" were reduced.

My second attempt, seeing how the first one was so strongly in favor of my personal bias, I labeled the bottoms of two identical containers, and using a lazy suzan, spun the samples enough with my back turned, to randomize in a crude fashion, which was which. Both looked identical prior to randomizing them, and I made the effort to not look at the containers when loading the brew basket. They were set down and out of sight afterwards, so I couldn't tell which was which.

First unblinded test was using my own roast of this months Roast And Learn Together- Kenyan. The sieves are 1000micron for the coarsest, 475 micron for the middle, and 250 microns for the finest. I did not quantify the amount of "boulders" or "fines" isolated from either sample.







The first test of my Kenyan showed more sweet, sparkling and cleaner floral notes and a vastly improved finish that lacked astringency with the brew that used only the middle cut, forgoing the "boulders" and "fines".

The second brew of the same coffee, this time including the "fines" yielded a cup that had a momentarily more intense hit to the palate, that quickly became more noticeably sour, sharp, less defined and finished less pleasantly. This cup without the "fines" reduced had a fatter, heavier feel to the cup.

Moving on to the Verve Colombia Diyer Orozco, a caturra, I decided to do my best, albeit crude attempt to eliminate some of my bias by attempting to blind the test.

Same parameters as the previous coffee, same approach. The brew with the "fines" and "boulders" reduced had the following generalized observations:
-Acidity: bready lactic, sparkling.
-Aromatics: clean florals, cherry compote
-Flavors: sweet cherry pie filling
-Finish:clean, sweet, no astringency

The coffee with only the "boulders" reduced, leaving the "fines" had these observations:
-Acidity: sourdough bread, lactic, twangy
-Aromatics" Intense, dark cherry
-Flavors: Dark cherry
-Finish: less sweet, goes towards astringent.


None of these are highly scientific, nor very reliable tests, but I'd rather start with a blurry picture than no picture. I should have done a third brew of the same coffee, same parameters, with the only difference being no sieving at all, but I didn't. That will have to wait for another day.

Prematurely, this reinforces my previous opinion that I subjectively find that reducing "fines" from a brew leads to a sweeter, less astringent, better beverage.

I'm not sure if you skimmed over where I clearly stated things like "subjectively", my "own personal bias" and "crudely unscientific", "preferences" "opinion" and what not. I did not know which version I was using in either of my second set of tests. It's still an A to B comparison, and full of pitfalls in assumptions. It would take a fair amount of work (and coffee) to do something like an A-A-A-B-A group and conversely, a B-B-A-B-B group, and blind both and verify if I can detect the differences. But that's something I want to work on.

And for the bias thing, it's difficult to remove in a one person controlled test. I'd rather repeat the same thing over and over with other coffees, and other people doing the grunt work, and just assessing the results, then reversing the roles.

I've seen the same bias and influence you mention in things like Penn & Teller's Bull$hit series, using things like "organic" bananas vs regular bananas, when in fact it's the same damn banana.

Tom,

Critical readers don't need your qualifiers. Or Jim's post either. It's clear that what you're doing is just the very beginning of an investigation, is fairly loose in its approach and that the results so far not very meaningful.

But the world is dominated by non-critical readers. So Jim's post has a lot of value.

That said, I'm curious about what you think you perceive.

Please enlighten me on what I need to do right. I think you're way over reading into what I've stated.

I didn't have a cup labeled anything in the second group (pointless in the first, knowing what was what). I didn't know which grind pattern I was brewing. Like I said, it's crude. And what I did can be reproduced, cheaply. It doesn't mean it can be defended as anything other than thin, poorly supported observations. I'm not implying that it's anything more.

I'd need a Ro-Tap and a dozen sieves and damn near lab standards to get anything objective. My thread title clearly states "preferences". I'm not sticking a flag in anything.

Who the hell is talking about even extractions? I'm not. I didn't declare anything objective about the grinds I used, only that they were more consistently within a median range.

Thanks Tom. I have been doing similar tests of fines and have come to a similar opinion. I don't have the resources to do double and triple blind tests to support my opinion, YMMV

I'm slightly confused about your process.

When you say you removed boulders, just larger than 1000 microns? When you removed fines, just <250 microns? Because you have a bunch of possibilities given the sieves shown (e.g. you could've sieved smaller than 475 microns as "fines," etc.)

I'm also curious if you've tried adjusting your parameters to compensate for sieving. In my experience with sieving boulders, I get nearly the same extracted solubles yield after sieving with less ground coffee (similar solubles from less mass of ground coffee, so higher %EY), so I compensate by grinding coarser to gain additional clarity and prevent astringency from creeping in. (Conversely, you'd probably want to compensate by grinding finer if you're sieving fines, since the tastiest EY should increase.)

My past experiments with fines sieving were limited by the equipment I bought, so I'll have to reserve that for the future. But I'm assuming you're happy with 250μ? The other possibility would be 212μ.

NoStream wrote:I'm slightly confused about your process.

You're not alone.

NoStream wrote:When you say you removed boulders, just larger than 1000 microns? When you removed fines, just <250 microns? Because you have a bunch of possibilities given the sieves shown (e.g. you could've sieved smaller than 475 microns as "fines," etc.)

Yes, hence not focusing on trying to define specifics, just a more median particle size. I don't think half of the coffee nerd world can agree on what a "fine" is, and this wasn't an attempt to define them.

NoStream wrote:I'm also curious if you've tried adjusting your parameters to compensate for sieving. In my experience with sieving boulders, I get nearly the same extracted solubles yield after sieving with less ground coffee (similar solubles from less mass of ground coffee, so higher %EY), so I compensate by grinding coarser to gain additional clarity and prevent astringency from creeping in. (Conversely, you'd probably want to compensate by grinding finer if you're sieving fines, since the tastiest EY should increase.)

No measurements of EY. It wasn't even on my radar. I'm using my palate only. I've long been a proponent of coarser grinding and sieving out "fines" however, for the exact same reasons you've stated. I've stated as much, in the past.

NoStream wrote:My past experiments with fines sieving were limited by the equipment I bought, so I'll have to reserve that for the future. But I'm assuming you're happy with 250μ? The other possibility would be 212μ.

These are nothing more than observations, It would take an inordinate amount of time and effort to fine a unique grind size, (using sieves to narrow that bell curve even greater) for each specific roast of each unique coffee.

SCAA drip grind suggests that anything under ~600um ends up in the pan & may be considered smaller than ideal with 2/3 of the ground mass falling between ~600 & ~1200um, ECBC less than ~400um in the pan, 2/3 of ground weight falling between ~400 & 850um.

For multiple sieve tests there is a long standing protocol for mesh sizes, e.g. for a 4 sieve test, the sieves increase by a factor of 1.41.

There's no lower cut off for SCAA cupping grind, in immersion brews it's probably less critical to eliminate the smallest particles by sieving (as long as you are not getting excess in the cup), as there's no flow through the grinds to exacerbate over-extraction & you're probably relying heavily on the smaller particles to get the job done. I sieve the boulders for some unfiltered immersions, also grinding marginally coarser to control the smallest suspended solids.

Randy Pope of Bunn did an interesting test at the NBC, his whole talk is worth hearing, but for the results of the blind cup tasting of coffee brewed with isolated grind sizes by sieving, FF to 01:08:50