Sieve Coffee Grounds for Espresso, Eliminate High-End Grinders?

Grinders are one of the keys to exceptional espresso. Discuss them here.

Postby coffeekid » Jul 12, 2016, 5:59 pm

[Long-time reader, first-time poster here. Be gentle.]

Short version of question: Can sieving ground espresso eliminate the need for a high-end grinder, particularly for home use?
Said otherwise: What, if anything, prevents us from achieving an ideal particle distribution with a cheaper grinder and a pair of sieves?

Before elaborating, I realize there are a few known issues and inconveniences about using sieves: [1] using sieves slows workflow, which might make sieving unrealistic for professional settings; [2] even when saving "fines" or "boulders" for other brewing purposes, sieving espresso will create some waste; [3] because of issues like static and clogging of holes, sieves don't eliminate all of the particle sizes they are intended to separate, which means making a science of sieving is not perfectly precise. For example, using a 200-micron sieve will not likely remove many or most 150-199 micron particles. [4] Even with the introduction of the Rafino—which I've pre-ordered—sieves are an added expense that negate some of the cost savings from not investing in a better grinder.

These caveats aside, here's the reasoning behind my question:

The long-standing consensus has been that the mark of a good grinder is a narrow particle size distribution. At the same time, notable figures such as Matt Perger have analyzed the best performing grinders (notably the EK43) and discovered that while they have a relatively narrow particle consistency at the target grind size, these grinders also produce a large number of "fines", or particles well closer to turkish than espresso. This discovery would seem to betray the belief that narrow particle distribution makes for the best coffee in general and espresso in particular. Perger's conclusion (, however, is that though fines CAN contribute to poorer extraction when particle distribution is very disperse (as with bad grinders), fines do not INHERENTLY contribute to poorer extraction. More accurately, the presence of fines in tight particle distribution (as with EK43s) allow us to reach higher extractions (>25%) in a reasonable amount of time; though fines have been said to instantly overextract, that overextraction is what contributes to the EK43 getting very high and delicious extraction percentages. Said otherwise, the problem is not with fines, but rather with how we extract them in the brewing process. In the end, Perger's conclusion supports the superiority of the EK43, which would seem to thwart my cheap agenda of finding a sub-$2k alternative for making excellent espresso.

Yet what doesn't make sense about Perger's reasoning is this: Even with the EK43's amazing particle range (fantastic particle consistency, yet relatively more fines), it's unclear why this would be better than a sieved grind with a more consistent particle size range, but just adjusted to a slightly smaller than the EK43's mean grind size. That is, if the benefits of the EK43 is that is produces more fines and therefore increases extraction rates, why not simply sieve a grind that is somewhat closer to turkish?

Perger's own data on the typical range, mode, and mean size of the EK43's grind distribution is this:

EK43: Range: 10 - 800+ microns; Mode: 295.5 microns; Mean: 237.3 microns

With my $200 Baratza Virtuouso, apparently I could never hit those numbers. BUT, with a few sieves added to my setup, I may be able to hit something like this:

Sieved: Range: 150 - 300 microns; Mode: TBD; Mean: 210 - 240 microns

The average/mean particle sizes in these two scenarios are similar; both are capable of achieving >25% extraction. But whereas the EK43 grind achieves this quickly because of the presence of fines, the sieved grind may indeed require more time to fully extract. So, the EK43 succeeds in [1] extraction and [2] brew time, and also does far better than any other un-sieved grind with [3] evenness of extraction. Yet it my hypothetical sieved setup, I would appear to actually beat the EK43 in [1] total extraction percentage and [3] evenness of extraction, though perhaps being slightly slower and therefore losing on [2] brew time. More still, if brew time were a real issue, one could simply reduce the mean particle size with smaller micron sieves. That is, the sieves could match the EK43 in [1] extraction and [2] brew time while also BEATING the EK43 in [3] evenness of extraction, because it would have a tighter particle distribution.

In short: It would take more time and definitely some experimentation, but even with a low quality grinder, it seems like one should be able to artificially create an ideal particle distribution with sieves.

All that said, there is one other concern I've heard floating around, which I have no way to verify or falsify: Espresso brewing NEEDS a wide range of particle sizes in order to maintain the structure of the puck in the brewing process. That is, one could perfectly make every particle in one's dose exactly one's ideal size (e.g. - 200 microns), but the uniformity of the particle distribution would block the flow of water somehow. This seems to be why some have argued that sieving is appropriate for all types of brewing EXCEPT espresso. Though I have no way of disputing this claim, this seems to be a convenient fallback in support of high-end grinders.

This is much less scientific, but in the end if feels like there is a lot of dubious and motivated reasoning in support of the particle distribution of high-end grinders: "You want your grind to be consistent, but not TOO consistent. What you really want is the particle distribution that happens to be produced only by grinders with 98mm flat burrs." Sieves may have their issues, but it seems that the ability to manipulate particle distribution in whatever way you want with a sieve would allow you to achieve an optimal grind for espresso.

Has my reasoning gone awry? Have I misunderstood something? Seriously, I'm not trying to build some grinding manifesto or to take down Matt Perger. I'm mostly curious if there isn't a way cheaper option for home baristas to achieve excellent espresso grinds.


Postby Marcelnl » Jul 12, 2016, 6:14 pm

One question, did you consider how much ground coffee will go to waste if you start sieving and the targeted bandwith particle size is not present in a large portion of your grinders output already? Sounds awfully wasteful to me, calculating that waste over many years might well buy you that 2k grinder
LMWDP #483

Supporter ♡

Postby pizzaman383 » Jul 12, 2016, 6:16 pm

That is a very interesting idea.
LMWDP #551


Postby coffeekid » Jul 12, 2016, 6:26 pm

Great question: Yes, I've considered it, though in truth it's hard to estimate exactly how much waste there will be without having a sieve set yet. There's no question that a better grinder will have less waste, as its initial particle distribution would more naturally fall into the ideal range. So, one might not exactly want to try this method with a blade grinder. Yet my plan would be to regrind "boulders" to bring them into the target size range, and to save fines for brewing turkish coffee and for other cooking/baking purposes. Again, I realize this would be some added time and effort, but if it has excellent results and can save me $1800+, it would unquestionably be worth it to me.

User avatar
Supporter ♡

Postby Peppersass » Jul 12, 2016, 6:30 pm

It might work, and I would encourage you to try, but I see some issues with your reasoning:

1. While you may be able to use sieves to obtain a narrow particle distribution, if the grinder doesn't produce the majority of its output in that range the method may not be practical. For example, if only 25% of the particles fall in the ideal range, then you'd be wasting 75% of your coffee (less fines).

2. What if your grinder doesn't produce enough fines to match the ratio of fines to other particles produced by a high-end grinder? What if it produces too many fines? Do you add some of them back after sieving?

3. I'm not so sure there's a broad consensus that narrow particle distribution is the mark of a superior espresso grinder. Yes, there's been plenty of buzz about that in recent years, but don't forget that long before the buzz began there were highly-regarded "high-end" grinders that don't produce as narrow particle distribution as the EK-43.

4. For me, the true mark of a high-end grinder is consistency. I think that's more important than particle distribution because you need consistency in order to properly dial in espresso. As long as you can dial in, you have a good chance of producing a good cup, even if the particle distribution isn't ideal. True, an EK-43 might let you squeeze more extraction out of a light roast, but there are other ways to get there (e.g., grind finer, run longer, use flow profiling, etc.) But if the grinder's particle distribution varies from one grind to the next, you can never dial in. That's why cheap, small-burr grinders tend to fail at producing good espresso. Your Baratza Virtuoso's grind probably varies, which means each time you sieve you would get different size piles of particles. I suppose that's OK, as long as you get enough of each size to match the desired particle distribution.

5. The latest trend in unimodal grinders is mainly focused on extracting more from today's light roasts, some of which are notoriously difficult to extract. But more extraction isn't always the way to get the best out of other roasts. Baristas were pulling great-tasting cups from darker roasts for decades before the light-roasting wave came along. The high-end grinders of that era (e.g., the Robur), don't produce unimodal grinds, but they sure are consistent.


Postby coffeekid » Jul 12, 2016, 7:01 pm

Thanks for your analysis, Peppersass. It's seriously helpful. I'd love to try this out and report back, if I can swing some investments soon.

1. You're definitely right on this one. Even in home use, the matter of waste and inefficiency shouldn't be trivialized. I imagine that this will become a matter of marginal costs and benefits on a person by person basis. One could adjust or broaden the range to waste fewer grams, but potentially sacrifice brew quality. In the end, people would have to decide how much waste they can tolerate for the sake of a better shot.

2. I've played around with the idea of reintroducing some fines back into the dose for precisely the reason you mention. Yet my more central concern is that Perger's reasoning about fines seems flawed. It totally makes sense that fines increase the extraction rate, but it's hard for me to see how they don't also make for an uneven extraction. So, I'm inclined to say "leave out the fines." If the extraction rate goes down as a result, it makes more sense to me to simply adjust the range further and further down until a similar extraction rate is achieved.

3. Yes, I have no disagreement here.

4. I also agree on the importance of consistency, particularly for developing a brew recipe. I do wonder if the sieving couldn't help manage the inconsistencies slightly, however. That said, I also realize that simply limiting grinds to 150 - 300 microns could still permit inconsistencies within that range, particularly with a cheap grinder like my Virtuouso.

5. That's a helpful insight on the utility of unimodal grinders in particular for light roasts. I'd be curious if there are any good resources on ideal extraction rates for various roast types. For example, light roasts are ideally extracted at 24 - 25%, whereas darker are better at 22 - 24%.

Again, thanks for thinking through this method with me.

User avatar
Supporter ♡

Postby Peppersass » Jul 12, 2016, 8:12 pm

coffeekid wrote:5. That's a helpful insight on the utility of unimodal grinders in particular for light roasts. I'd be curious if there are any good resources on ideal extraction rates for various roast types. For example, light roasts are ideally extracted at 24 - 25%, whereas darker are better at 22 - 24%.

Again, thanks for thinking through this method with me.

As far as I know, a reference on ideal extraction yields for various roast types doesn't exist. I don't think such a correlation is possible.

That's because the ideal extraction yield varies with each specific coffee and roast, and may also vary with other factors such as grind, dose, brew method, temperature, pressure, etc. I think the best we can do is to identify the extraction yield that produces the best balance (taste) for a specific coffee/roast using specific brew parameters. This is something that's done when dialing in a particular batch of coffee and may not apply to a different coffee roasted exactly the same way or pulled differently.

Some background on extraction yield: Modern brew charts are based on research done decades ago that showed most people prefer coffee brewed to an extraction yield of between 18%-22%. For many coffees and roast styles, this is the range in which the sour and bitter flavors are in balance with each other. Less extraction results in a sour cup, more extraction results in a bitter cup. Balance is good because when the cup is too sour or too bitter, it's usually not possible to taste the more subtle varietal flavors of the coffee.

Bear in mind that the brew chart research was done a long time ago with coffees and roasts that were popular at the time, and may not apply to all of the specialty coffees and roast styles we have access to now. Also, preferences in coffee flavor may be quite different today than they were decades ago. So the effective range for today's coffees and preferences could be considerable wider than 18%-22%. We've seen some evidence of that with reports of 25% extraction yields producing great cups with light roasted coffee.

If it isn't clear by now, I should point out that higher extraction yield doesn't always produce the best cup. That seems to be true with today's ultra-light roasts, probably because the flavors are harder to extract. But it doesn't necessarily apply to medium or dark roasts. If you pull too much out of roasts like that, you'll probably get a bitter cup. That's why Ristrettos were popular in 3rd Wave cafes before the present era of light roasts.


Postby Wacobipe » Jul 12, 2016, 8:46 pm

This is an intriguing idea as you could toss a lot of coffee before you get to the point of having paid for the grinder. One alternative to consider is the HG-1. While manual, it is a considerably less expensive high end grinder option...but nowhere near as low cost as the sieves.


Postby Dpablo » Jul 12, 2016, 8:47 pm

The sieve method seems to me like it won't work. Sieves allow one type of particle size through, and unless it is in the same range as the grind setting, the sieve will effectively be changing the grind setting. Moreover, one is limited in grind setting to whatever the sieve will allow. I suppose you could mess around with dosing to get around this, but it will make things much harder.

User avatar
Team HB

Postby TomC » Jul 12, 2016, 8:55 pm

I've tried this before and it just makes a mess and simply doesn't work. At espresso fineness, there's a lot of static clinging and cohesive forces that make it difficult to separate. A good portion of fines are practically magnetized to other smaller particles. You'd have to set up a Ro-Tap, most likely, to get a decent yield, while still wasting a ton of coffee, making a bigger mess, etc.