Grind Sifting for Brewed Coffee

Coffee preparation techniques besides espresso like pourover.
User avatar
shadowfax
Posts: 3545
Joined: 19 years ago

#1: Post by shadowfax »

Has anybody out there messed with sifting/sieving the fines from dry coffee grinds before brewing? I'm thinking mainly for cupping/French press/vac pot/any method that doesn't rely on coffee for flow-restriction to control steep time, though I guess it could apply to those as well, with more complexity as to how to set the grind (probably much finer overall) and the probable resultant change in optimum brew time. At any rate...

I've been vaguely interested in sieving coffee grinds for awhile now, with some roots of that in the TGP and related discussions, as well as the look at the Baratza Vario grinder. I've always been rather bugged by the silt in my French press coffee, and while there are various methods to ameliorate this, it strikes me that though these methods make a marked improvement, in theory don't solve the problem of 'extraction differential' from steeping bimodally distributed coffee grinds entirely. Can your cup get cleaner? Would this allow you more flexibility with brew ratios/brew strength?

My first real interest was sparked by David Walsh's post about brew ratios, the commonplace occurrence of 'up-dosing' in traditional coffee brewing among specialty coffee professionals, and fines produced by most burr grinders, particularly consumer grade ones. Be sure to read the comments from James Hoffman/Nick Cho and David's responses; they clarify his article pretty well. This was David's superficial head-turner for me:
David Walsh wrote:I've heard it mentioned that the fines add complexity, I would argue brew a cup of 100% fines and see how complex you find it. A turd in a swimming pool probably makes the swimming pool more complex. It seems natural enough therefore that those of us sensitive to these types of flavours , identifying them as undesirable will perhaps instinctively move to coarser grinds that produce fewer fines. When we find the mouthfeel and body lacking we increase the dose.
David's article mentions the Brunopasso MPS-50 (MPS=Micro-Particle Sifter, I believe), a cool Japanese contraption for easily sieving grinds. I believe you can accomplish something similar (albeit dramatically less usable) with a tea infuser if you want an on-the-cheap, very soon solution; I have no idea where else you will find the MPS-50 other than Avenue 18, and I believe they are out of stock till mid-January.

In any case, after months of having this idea on the back-burner (technically, it still kind of is), I got one for myself earlier this month, and it arrived late last week. I've only just begun to play with the thing, but for starters, when sieving grind samples from the Vario in the range between about 2 macro notches finer than 'filter' all the way up to the coarsest settings, I see a loss by weight of about 10-30% (higher at finer settings). The 'micro-powder' is truly fine, rather finer than an espresso grind to my fingers.

I haven't taken David's advice and brewed with 100% fines yet, but I have been messing around, just a very little bit, with siphon brews and cupping of the larger particles after sieving. The siphon brews are interesting; even at pretty fine grinds, the draw-down time is as fast as you could imagine, not unlike a plain-water draw-down. I haven't been able to do any side-by side comparisons of those brews, but I did a little cupping with a fellow coffee enthusiast on Sunday, just brewed 2 pairs of samples of Black Cat (the only coffee I had on hand that wasn't on reserve for the holidays) at the same grind setting, with one pair sifted. No blinding or anything like that, though I don't think it would have been necessary--VERY different cups of coffee. The sifted samples were dramatically cleaner, maybe a bit in mouthfeel, but mainly in taste. Going back and forth, it really amped the sense of grainy bitterness from the unsifted samples. Yet, it wasn't a total wash that one was better than the other. The unsifted samples came off badly at first, but as the samples cooled the chocolate/aromatic wood flavors emerged from that sample, an element that, while it was rather muddied in the unsifted samples, was comparatively absent from the brighter, clearer sifted samples. I'm guessing they were under-extracted, but I have no idea.

As I alluded to in this thread, looking some more at this thing is something I hope to do while visiting family over the holidays, with my captive audience of a few extra coffee drinkers. In the mean time, I am curious if there's some more wisdom on this topic that I've missed out on, or if any other HBers sieve their brews before adding water.
Nicholas Lundgaard

User avatar
RapidCoffee
Team HB
Posts: 4994
Joined: 18 years ago

#2: Post by RapidCoffee »

Great stuff, Nicholas. I'm sorry I missed that article, if only for the plot showing how particle size distribution changes with grind setting:


One observation: if you integrate under the peaks, the percentage of fines by weight is far lower than the ~20% you are sifting out. I assume the horizontal axis is particle diameter in microns, and the vertical axis is number of particles (expressed as a percentage)*. The fines peak is composed of particles whose diameter is roughly one tenth that of the main peak, so particle volume (and mass) is one thousandth as large. Combine that with the smaller peak area and, well, it's hard to believe there's a significant impact on taste from fines extraction.

The big guns in the linked article are focussing on fines extraction, but I'm guessing that other factors come into play. Are you using a glass filter in your vac pot? Fines contribute heavily to mouthfeel and muddiness in presspot (and glass-filtered vac pot) coffee. Fines not only help regulate flow in espresso, but can also impact the flow in filter brew (and cloth filtered vac pot) by clogging the filter pores, which will alter the extraction of the larger particles.

Any idea of the mesh size in the MPS device? It's a cool toy, no doubt.

EDIT: It's possible that the vertical axis represents particle volume (rather than number of particles). In that case, there would be a lot more fines... but still much less than 20% by weight. The area under the fines peak would give an estimate of the volume (or weight) percentage of fines.
John

Advertisement
User avatar
cannonfodder
Team HB
Posts: 10497
Joined: 19 years ago

#3: Post by cannonfodder »

Interesting. I only have espresso grinders (Cimbali Max's) which produce fines. They make super espresso but French press can be tricky. I thought about sifting the grinds in a fine mesh strainer to drop out the fines before steeping my coffee but never got around to messing with it. I may have to try it just to see what happens, and see how much drops through the strainer.
Dave Stephens

User avatar
shadowfax (original poster)
Posts: 3545
Joined: 19 years ago

#4: Post by shadowfax (original poster) »

RapidCoffee wrote:One observation: if you integrate under the peaks, the percentage of fines by weight is far far lower than the ~20% you are sifting out. I assume the horizontal axis is particle diameter in microns, and the vertical axis is number of particles (expressed as a percentage). The fines peak is composed of particles whose diameter is roughly one tenth that of the main peak, so particle volume (and mass) is one thousandth as large. Combine that with the smaller peak area and, well, it's hard to believe there's a significant impact on taste from fines extraction.
Does integrating a histogram of diameters get you an accurate histogram of weight? That seems a little off, but I am super-rusty. That said, I agree my numbers are quite high just compared to David's numbers, and I suspect that the mesh size on the MPS-50 is such that it's taking out a decent bit of the smaller particles in the bigger peak as well as all the fines. This matches really well with my experience so far with respect to the feel of the sifted-out particles in my hand: it's finer than espresso, but it's a far cry from confectioner's sugar which is what I'd expect if it were just getting the tiny peak. This seems further evidenced by the spread I see in terms of loss from French press grind (low; ~10-15%) vs. a few notches north of espresso (high; ~30%)--to get such a differential you'd have to be taking a chunk out of the low part of the peaks.

On the other hand... David also wrote an interesting review of the Vario, and one of the things he delved into was that the Vario is (to him) a far cry from the best shop grinders (Mahlkönig Guatemala/Tanzania). The relevant graph:


The Vario produces about double the fines of grinders that are about 4x its price; It also exhibits less consistency in its larger particles (wider, shorter peak).

So, make of that what you will. I'll try to get some good pictures of the Brunopasso filter to help get some sense of the mesh size, though of course that's pretty murky waters. I'll look into contacting the manufacturer for that data if I have a chance--a number would be nice.
RapidCoffee wrote:The big guns in the linked article are focusing on fines extraction, but I'm guessing that other factors come into play. Are you using a glass filter in your vac pot? Fines contribute heavily to mouthfeel and muddiness in presspot (and glass-filtered vac pot) coffee. Fines not only help regulate flow in espresso, but can also impact the flow in filter brew (and cloth filtered vac pot) by clogging the filter pores, which will alter the extraction of the larger particles.
I have been using the glass rod now that I have the sifter, actually. I'd never been able to use it before, as I found the draw-down times inconsistent and upsetting; they ranged from 1:30 up to around 3:00 for me; I think that the glass rod I have (an antique Cory rod, though it appeared to never have been used when I got it) is unsuited to meeting the glass of my 3-cup Yama. I never quite found a spot on the Vario that would get the draw-downs lower than this, and it annoyed me to no end--I'm not a fan of worrying that my glass brewer is going to implode. So I gave up and went with cloth, which is really superb--I don't see much variance on draw-down time there. I also got the Hario mesh filter and holder I mentioned in your tea thread, and that produces a cup that's similar to glass-rod cups, minus the finicky problems with draw-down (owing to the comparatively massive filter area). As I mentioned in your thread, the mesh filter produces a cup similar to French press, with the heavier body and silty cup. It's certainly not as crisp or light as the cloth-filtered brew.

Just in case it's not clear--none of the above comments have to do with sifted grinds--I haven't even done a sifted cloth-brew yet, only glass-rod brews. As you say, the fines obviously effect flow through the filter, and it's really dramatic with that method. Draw-down time goes from 1:30+ to about 10-15 seconds. These brews still let some silt through, oddly enough, but it's very strange: it's so ultra-fine on my fingers that I can barely feel it, and drinking the silty bottom, while unpleasant, is a lot more palatable than drinking the silt from a brew of unsifted grinds. I am thinking this silt is from extremely, extremely tiny particles that are statically attached to the big particles, where the static charge is stronger than the force of the agitation of the sifter, given their tiny mass. I can 'feel' these particles on my tongue when I cup the sifted coffee as well. My impression of them is that they have a comparatively diminished effect on cup flavor. I am sure they'll disappear in the cloth brew, but it's just like with not sifting the fines--if they make it into the water, no matter what you do (break and clean, french pull, paper filter, etc.) they still extract into your coffee. I do think it's a non-issue, but I guess a purist could whip out the feces-in-the pool argument on me and I'd be pretty speechless. :wink:
Nicholas Lundgaard

User avatar
RapidCoffee
Team HB
Posts: 4994
Joined: 18 years ago

#5: Post by RapidCoffee »

shadowfax wrote:Does integrating a histogram of diameters get you an accurate histogram of weight?
Having played around on a laser diffractometer during the TGP, I know a little bit about how they work. As the name of the instrument implies, raw measurements are associated with the diffraction of a beam of coherent light. Assumptions are made about the shape and reflectivity of the particles (and the refractive index of the medium for liquid analyses), and the measurements are translated into particle counts, volume, and surface area for different particle diameters. If you accept the results, then integrating under the histogram peaks should give you a reasonable estimate of the number (or volume or surface area) of particles.* It's kinda like sieving with progressively coarser screens, one per histogram bin. The weight ratio won't be perfectly accurate, because the particles are not spherical, but we're talking about large differences here. At least, that's my take on it.
shadowfax wrote:I have been using the glass rod now that I have the sifter, actually. I'd never been able to use it before, as I found the draw-down times inconsistent and upsetting; they ranged from 1:30 up to around 3:00 for me...
Those draw-down times seem really long to me. I typically brew for 90-120 seconds, and adjust the grind (on a Mazzer Major) to produce draw-down times of under 30 seconds. I'm using a brew ratio of about 1:15 (30g coffee to 450ml water), preheating the water and allowing it to travel "north" before stirring in the grinds. The resulting vac pot brew is nothing like French press. It's smoother, flavorful, yet completely devoid of muddiness.

* EDIT: I went ahead and did this for the Vario peak on the plot linked above. The vertical axis label (Verteilungsdichte) appears to translate as probability density function. Adding the data points below 100um resulted in just under 2% for the Vario curve. I'm not sure why David used 200um as the delimiter, or how he came up with 15% in the linked article.
John

SL28ave
Posts: 127
Joined: 18 years ago

#6: Post by SL28ave »

I've done it for cupping and drip of a few light roasts. I remember doing a small updose to compensate. Sifting made for a smoother flavor. If I'm preparing coffee for a girl I'm interested in, grind sifting is now standard protocol.

While I had screens, I didn't use them for this. I'd just put the grounds in a teflon coated paper cup, bounce the grounds up and down, and the fines would migrate and stick to the bottom.
"Few, but ripe." -Carl Friedrich Gauss

User avatar
hbuchtel
Posts: 755
Joined: 19 years ago

#7: Post by hbuchtel »

SL28ave wrote:If I'm preparing coffee for a girl I'm interested in, grind sifting is now standard protocol.
:lol: It is also a good introduction to coffee for people who think "coffee is bitter".

I used a hardware store sifter for a while before buying a 'MPS'. I would compare a sifted French Press cup to an americano - they both share that very 'clean' aspect. I slightly prefer the complexity :roll: of the non-sifted FP, but I've never had a guest agree with me.



Sifters like this can be used to roughly compare fines production from different grinders- adjust each grinder to produce approximately the same maximum size particles, then shake the sifter a certain number of times for each sample and weigh the results. I was surprised to find that for a FP grind my expensive espresso grinder produced as many fines as my cheapo "Little Flying Horse" (ubiquitous mainland China grinder- same design as the Kitchenaid Proline).

BTW, does anybody know of an 'industrial size' equivalent of the MPS?

Regards, Henry
LMWDP #53

Advertisement
User avatar
hbuchtel
Posts: 755
Joined: 19 years ago

#8: Post by hbuchtel »

More info about 'bimodal' grinding -

http://www.ambexroasters.com/pages/arti ... art_2.html

(near the end of the article)
LMWDP #53

User avatar
dsc
Posts: 1166
Joined: 17 years ago

#9: Post by dsc »

Hi guys,

I've tried sifting a while ago, few other people on TMC tried it as well after I've asked the same question as you. I agree it creates a different cup, I would even say better cup, fines tend to screw everything up and increase bitterness, slow down flow etc. Sifting is not a bad idea, but time consuming and that's the main reason I switched to a Guatemala which isn't perfect size/noise-wise, but great for non-espresso brewing.

Regards,
dsc.

User avatar
luca
Team HB
Posts: 1135
Joined: 19 years ago

#10: Post by luca »

About five or six months ago, I was sent a Kyocera ceramic burred hand grinder to test out. I noticed that it seemed to produce a lot more fines at coarser grind settings than other grinders that I had had, so I started using the finest sieve that I could find and it certainly improved the cup for french press. I also have a little sunbeam grinder that I use for brewed coffee and I sieve the grinds from that. I bought a really fine tamis to make killer mashed potato and that has been quite good. My impression, which I have certainly not tested, is that the particle size spectrum gets larger as you grind more coarsely.

The sifted cups certainly taste cleaner and less bitter, however they do seem to lack a bit of complexity. Seeing as cupping seems to do a great job of bringing out complexity, one brewing method that I have been playing around with is steeping in a pyrex jug for a while, breaking the crust and skimming, then pouring the liquid off through filter paper. Using a coarse grind that is well sieved, I have been able to steep for a fairly long time; four minutes is the standard time that I have tried, but I suspect that you could probably leave it for ten minutes comfortably. This is probably the most foolproof brewing method that I have stumbled across; if you can't make coffee taste decent this way, you should probably throw it out.

Interestingly, even after sifting, if I brew pourover filter I still seem to get fines on the filter paper. Is it possible that fines are produced by the very act of brewing?

Cheers,
Luca
LMWDP #034 | 2011: Q Exam, WBrC #3, Aus Cup Tasting #1 | Insta: @lucacoffeenotes

Post Reply