loader image
January 30, 2017 /
The Grinder Paper: Explained

The act of grinding coffee is full of mystery and myth. A group of likeminded people recently embarked on a project to shed some light on coffee grinding, and found some interesting results. So interesting, that we decided to turn it all into a paper; subjecting our methods to peer review and the rigorous process of publication. Last week, this paper was published in a Nature sub-journal -Scientific Reports- as open access. This means anyone can read and benefit from the results. Yay for science!

Scientific papers like this must be written in a certain way that can be hard to digest. I’d like to spend some time tearing apart the paper; describing what we did, and what I think it means. Feel free to follow along with the actual paper here. http://www.nature.com/articles/srep24483

I’m going to include my personal and sometimes debatable opinions/remarks in bold like this so you don’t fall asleep.

The effect of bean origin and temperature on grinding roasted coffee.

Introduction

This is a super basic rundown of what’s up with coffee and this experiment. Not everyone peer reviewing this knows as much about coffee as most of you do, so we thought it relevant to include some pertinent details:

  • There are two main considerations with green coffee; variety and processing.
  • Roasting plays a large role in how coffee tastes.
  • Most compounds in roasted coffee are products of Maillard reactions, but there’s a lot going on.
  • We used four different coffees for the experiment:

And finally, that this experiment is primarily interested in how bean origin, processing method, roast level and temperature can affect the results of grinding.

Method

Coffee is extremely complex. It’s nearly impossible to create a computer model of how it fractures during grinding, so we have to do it experimentally.

We did two main things for the experiment:

  1. We ground coffee of different origins/roasts/processes at the same setting and measured the resulting grind samples.
  2. We changed the temperature of one coffee to four distinct levels, ground a sample while it was still at each temperature, and measured the resulting grind samples.

To measure the grinds we used…

Laser diffraction particle size analysis

We assumed that the most important metric for measuring ground coffee is the particle distribution. That is, a measurement of the size (diameter) of every single coffee grind in a sample. Yes, flavour is important, but it’s much much more difficult to measure flavour with any accuracy and precision.

Measuring particle distribution of coffee is performed with a laser diffraction particle analyser (LSA). Essentially, it sucks coffee grinds down a tube, shines a light across their path and measures the “shadows” that each individual grind casts onto a detector. It’s actually much much more complex than that, but this description will suffice. My brain hurt trying to delve further. They’re incredibly precise, sensitive machines and need to be calibrated often. Here’s a simple diagram showing the usual setup inside an LSA.

Particle Analytical ApS

 

Grinding

We used an EK43 for the study, because it holds a negligible amount of grinds in the chute after the burrs. This study relied on coffee beans being ground at specific temperatures, and not suffering from any cross-contamination. We couldn’t allow the beans to lose any energy to a hopper, grinder throat, or burrs for any appreciable amount of time before they were ground.

There were 3 EK43’s present during the experiment. One was found to be producing the most delicious coffee, and was used for the samples. It had Turkish burrs installed. EK43’s have quite a large potential for burr misalignment, so this grinder was likely the least misaligned.

We kept the grind setting precisely the same for every sample. 2.7 on the dial if you must know. We also let the grinder cool down to room temperature between every sample to rule out frictional/electrical heat as a variable.

Temperature-specific samples were kept in paper cups within different environments. Room temperature (20C), freezer (-19C), dry ice (-79C), and liquid nitrogen (-196C). They were ground within one second of retrieval and showed no water condensation.

We took 3 samples for each data set, and performed each of those twice. So, 6 data sets per temperature/coffee. The results of these data sets were also passed through an analysis of variance (ANOVA) to make sure they were similar enough to be considered accurate.

Do Differences in the Green Bean Affect the Final Grind?

Here’s where things get tricky to understand. You might be familiar with this kind of graph to communicate particle size distributions:

Along the x-axis (horizontal) is the size of the grinds in microns. 1 micron equals 1/1000th of a millimeter. This axis is on a logarithmic scale, which places 1 & 10 as far apart as 100 & 1000. This is because a sample of coffee grinds covers a huge 3 orders of magnitude (0 to ~1000 microns) and we need to fit it all in without losing too much resolution at the smaller sizes.

The y-axis (vertical) is the volume% of the grinds. This one’s easy: the higher the peak, the more of that size particle there are.
Eg. trace vertically above 400um to the brown line. That particle size makes up 8.5% of the sample by volume (not weight!).

We took this style of data presentation a few steps further for this particular experiment.

Firstly, let’s look at a count of the particles. Instead of showing volume, we simply graph how many of each size particle there are (blue). The first thing that’s rather obvious; there’s an INCREDIBLE number of tiny grinds in each sample. 99% of the particles are below 70 microns (0.07mm) in diameter. This means that for every single grind with a diameter above 100 microns, there’s one hundred million with a diameter below 100 microns. That’s an excellent piece of trivia with which to sound smart at dinner parties.

When analysing coffee grinds, a very important factor is their surface area. The more surface area they have, the faster and more easily water can extract its flavours. To turn the data above into something resembling surface area, we firstly assume that every particle is a sphere (this is fairly common practice, and for gnarly, uneven coffee it’s a “conservative estimate”). Then, we take the size of every particle and calculate what the surface area would be if it were a sphere.

This gives us some really cool data!

The solid lines below are the ‘counts’. The same as the blue line above

The dotted lines below are the ‘relative surface area contribution’. That is to say, the proportion of the total surface area that is provided by each particle size. Once again, something is quite obvious. The smaller grinds contribute the overwhelming majority of the total surface area. ~70% to put a number on it.

This graph also shows the particle distributions of the various coffees used in the experiment. Straight away, you will notice that they’re all incredibly similar. The Tanzanian, Ethiopian, El Salvadorian and Guatemalan profiles are shown in black, purple, red and blue, respectively.

Turns out that origin/processing/roast has much less of an effect on PSD than I had ever previously thought. This is arguably a good thing: we have less variables to worry about!

The next thing to wrap your thoughts around is that fines contribute 70% of the total surface area. Yes, water moves inside the grinds to extract solubles, but it takes exponentially longer for the water to get inside, do the work, and move back out into the brew. Fines are our friends!

 

Do Differences in the Roasted Bean Grind Temperature Affect the Final Grind?

This is where things get really interesting.

Coffee is amorphous – it’s made up of thousands of different molecules jammed together within an irregular plant structure.

The diamond in an engagement ring is crystalline – it’s a perfectly repetitious pattern of carbon atoms.

When you change the temperature of amorphous things, they sometimes undergo a ‘glass transition’. That is, they change from a soft and rubbery material to a hard and glassy one very quickly. Some materials also undergo a shattering transition, where they tend to break into more and smaller particles. This is all pretty key when discussing how coffee breaks apart in a grinder.

We found a very pronounced difference in the particle size distribution of one coffee ground at different temperatures. Firstly, make yourself familiar with each colour and the temperature it represents. You’ll notice that the mode (peak) of the PSD gets smaller as the temperature dips (31% drop over the 4 samples). As the coffee gets cooler it also becomes more brittle, throwing off many more tiny particles in the grinder. It also fails to escape the burrs at the larger sizes of the warmer samples (ie. shatters more easily).

The largest difference is between 20C and -19C. The coffee likely undergoes some kind of shattering/glass transition between these two.

We also confirmed that this transition is reversible, so you don’t need to worry about a coffee getting too hot and not being able to go back to tasty town.

The mode is the most-occurring particle, easily identified as the highest “peak” of each line. This gets larger as the temperature increases.

Skewness is merely a numerical representation of asymmetry in a data set. The colder samples were less skewed because they contained more small particles and fewer large ones, reducing area under the line to the right of the peak.

Mean is the average. You’ll see the freezer has a higher mean than the room temperature sample. This is because the room temperature sample had many more particles in the 3-5um range, but far fewer in the 8-30um range.

This transition temperature will likely be a very important discovery for all future grinding technology.

Concluding Remarks and Thoughts

Origin and Processing

Anyone who blends coffee will be happy to note that the surface area of different coffees is pretty much the same at a fixed grind size. This means that the major consideration for blending is to ensure that each blend component is equally soluble. In other words, that they all reach the same level of extraction within the same brewing time. Roasters, dust off your refractometers.

Temperature

Remember all those mornings when you dialled in a coffee perfectly, only to have it thrown out the window after you made ~20 coffees? The culprit isn’t the grinder expanding from the heat; it’s the beans soaking up the heat from inside the grinder before they’re ground. That heat energy makes them less brittle, creating a coarser grind even though you didn’t change the grind setting.

Here’s some more questions with pretty much the same culprit. Ever wonder why…

  • coffee doesn’t taste as good on a hot day?
  • your grinder can seem so inconsistent in quiet periods?
  • the shots run faster during a rush?
  • this all doesn’t happen with an EK43?
  • After grinding finer to achieve the same shot time, that the shots don’t taste the same as they did in the early morning?

The culprit for all of these is: the beans heating up inside the grinder and grinding differently.

The less time the beans spend in the grinder, the less they’ll be affected by its heat. It’s extremely difficult and costly to create a grinder throat that evenly heats or cools the beans before grinding. With that, I strongly believe that the only way forward is to use grinders that don’t contain any coffee between doses.

I’ve never really been a fan of the results with the Mythos grinder’s heating feature, and this experiment is an excellent explanation as to why. Heating the beans to achieve consistency pushes the beans above the shatter transition temperature and significantly reduces the total surface area (read: less and less-even extraction).

More cold = finer particles = more surface area = higher extraction. Lower temperatures could also mean less evaporation/sublimation of aromatic compounds (aroma loss).

Keep your pre-weighed doses in the freezer for higher, tastier extractions (though make sure they’re sealed without too much moisture or any oxygen).

Fines

Wow. Back in 2012 I won the World Brewers Cup with a routine centred around removing fines. As it turns out, sifting doesn’t really get rid of all the small particles. There’s still millions of them stuck to the larger grinds. When coffee is torn apart in the grinder it leaves pockets of positive and negative charges all over the grinds that attract the fines. Conclusion: sifting is pretty useless for particle segregation and testing.

Now that we know just how much of the brew is made up of fines extractions, it becomes increasingly obvious that fines aren’t the villain; otherwise every coffee ever made would be horribly over-extracted. Here’s how to think about it: The upper limit of tasty extraction is decided by the most-extracted particle. This is always the smallest particle. So it’s up to you to make sure no portion of the grinds ever get over-extracted.

It’s also up to you to reduce the amount of coffee that’s under-extracted (ie. the inside layers of the largest grounds). The simplest way to do this is -as I’ve always said- to use a grinder that produces an even particle distribution. Nothing new here.

Once I figure out how to brew fines properly I’ll be back at the WBrC with an apology routine!

A good days work. Looking forward to the comments, suggestions and corrections!

If you have found this useful and want to enjoy delicious coffee with the rest of the community – register for our monthly Superlatives coffee subscription. Or if you just want to keep up with every thing Barista Hustle – sign up to the Newsletter. 

Thanks to the team:

Christopher Hendon (Author of Water for Coffee and Post-Doctorate at MIT) came up with the whole thing and managed every aspect. He’s a human embodiment of science itself.
Maxwell and Lesley Colonna-Dashwood (Colonna Coffee, Bath) played integral roles designing, facilitating, and performing the experiments.
Erol Uman and Brian Miller (Meritics Ltd.) graciously supplied the laser particle analyser.
Stephen Leighton (HasBean Coffee) supplied delicious roasted coffee.
Christian Klatt (Mahlkonig) contributed grinders and experimental design.
Keith T. Butler (Department of Chemistry, University of Bath), Brent C. Melot (Department of Chemistry, University of Southern California), and Rory W. Speirs (School of Physics, The University of Melbourne) provided invaluable advice towards experimental design and execution.
Matt Perger contributed to the experimental design and wrote this explainer!

Coaching Calendar

Find a course with a BH Certified Coach

March 2020

Mon Tue Wed Thu Fri Sat Sun
1
2
3
4
5
6
7
  • MY, Emery School of Coffee, Kuala Lumpur, Percolation with Cadence Sim
8
9
10
  • FR: BARISTA ONE
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

News & Updates

Sign-up, Take part and keep in touch!

84
Leave a Reply

Please Login to comment
avatar
  Subscribe  
newest oldest most voted
Notify of
Barry Williams
Guest
Barry Williams

Hey Matt,
I just came across this interesting reading but I cannot find a answer to questions asked regarding the difference between producing fewer fines and grinding finer. Also I would be interested in continuing this experiment to the other extreme ala James Hoffman’s heating the beans to 60C to grind.

BHLearn
BHLearn

Hey Barry, You might find out white paper called the Grinder Heating experiment interesting reading. We follow up on Hoffmann’s 60°C idea there.

Christopher H. Hendon
Guest
Christopher H. Hendon

Killer overview Matt. Thanks for the hard work.

Daryl Grunau
Guest
Daryl Grunau

So instead of heating it up to have consistent temps like the Mythos it seems we instead need to find a system more like the Mazzer Kold to keep beans cool. Maybe liquid cool the whole chamber and borrow from computer and server systems to keep it chill.
This would explain to me why it’s easier to keep tasty shots consistent in winter months here in Canada as well.

Geoff K
Guest
Geoff K

Watch for condensation though.

Geoff K
Guest
Geoff K

Again thinking back to the Sette’s design – its not terribly difficult to extend passive fins from the outer ring, and still drive it with a gear arrangement. You can then actively air cool the fins on a thermostat.

Finally looking at the Hopper arrangement – simply add some vertical distance between the grind chamber and the hopper. Maybe even do a reverse doser arrangement (you dose the unground beans into the feed tube to grind).

All of this is probably why the Versalab is such a great grinder too – because you’re effectively single dosing with their arrangement.

Arnhem Coffee Roastery
Guest
Arnhem Coffee Roastery

Brilliant summary and makes the complicated maths and science understandable!

Thinking further, for a warmer climate (I am in the tropics), keeping my beans in the fridge rather than just on the bench would make sense – to get them below 20 deg.

Geoff K
Guest
Geoff K

re: Grinding solutions. I don’t think that a busy Cafe punching 600-700+ coffees per day is going to practically Single Dose to keep unground beans cool. Extending the distance between the Grind chamber and the Hopper will go some distance – but then you have to stop the popcorning. So a grind burr arrangement like the new Sette which actually sucks the beans in during the process is probably the go. This all presumes that particle distribution size of the Sette is bang on the money – something people have been really quiet about so far (or better accuracy can… Read more »

Matt Perger
Guest
Matt Perger

Yeah, you don’t need much water. Best to keep it low, and also controlled rather than dictated by random humidity and temperature and resulting surface condensation.

Matt Perger
Guest
Matt Perger

Can you please expand on your concerns re rigour? I’m curious.

Geoff K
Guest
Geoff K

The window for reducing static but keeping a great grind/shot is pretty small.

You are far more likely to push into the “ruined coffee” territory. Fairly certain there is a well-respected article out there that whilst coffee itself doesn’t mind being frozen/reheated – it assumes that you have eliminated condensation.

Arnhem Coffee Roastery
Guest
Arnhem Coffee Roastery

True, although that helps reduce static, so not all bad either!

Matt Perger
Guest
Matt Perger

Volumetric can be +/- a bean or two. An augur or feed mechanism could deliver bean-by-bean onto a load cell. Stops when weight is reached. Dumps into grinder.

Luke Pchr
Guest
Luke Pchr

Thanks, such a great article.

Geoff K
Guest
Geoff K

whats the typical variance if you volumetrically dose? I know as i adjust grind size at home (time based grinder), I can move out of my target dose quite easily. The benefit i see of a feed tube releasing an amount of beans – means that whilst it might not be accurate (you’ll gain that from weighing as its ground into the PF though) – its fast, and can still be a robust setup. Weighing the beans in a 2 stage chamber is probably viable as well – but this assumes zero/near-zero grind retention if we’re still chasing god-tier quality/control… Read more »

Matt Perger
Guest
Matt Perger

All of this is quite difficult and in some cases still results in beans sitting inside grinder parts. The easiest solution is weighing or volumetrically dosing the beans before grinding.

Matt Perger
Guest
Matt Perger

I don’t believe a busy cafe will pre-dose (although we have at St Ali. It sucked).

What needs to happen is grinders that weigh the beans before grinding and dispense it all on demand. It’s really the only way.

Arnhem Coffee Roastery
Guest
Arnhem Coffee Roastery

The assumptions made about ‘shattering’ and their impact. They may well be correct but there wasn’t quantitive analysis of the speculation about shattering in terms of microscopic observation. Its simply too subjective in my opinion. Please understand I am not throwing the baby out with the bath water, its a fascinating paper and a discussion paper it will hopefully generate lots of debate and further speculation – leading to an improved understanding of the coffee making process. Its important that people realise that a published paper with a lot of science in it does not have the same standards applied… Read more »

Geoff K
Guest
Geoff K

Its an interesting problem. And the reality is you need to prep shots in under 10-15 seconds (including weighing) in a busy cafe scenario.

I presume the usual suspects are thinking hard about it now.

Sean Martin
Guest
Sean Martin

I think there are other aspects about grinding that are critical to grinder and bean shattering behavior were simply ignored, so I think the tests aren’t that useful if the results don’t account for them. Such as grind speed. The speed at which a bean is struck as well as the mass of the striking object, will determine how that bean reacts, and at different temps. Think of smashing taffy with a hammer. The hammer needs to travel at different rates to smash the taffy at different temperatures. Anyway it’s not the only consideration but an obvious one that comes… Read more »

Matt Perger
Guest
Matt Perger

I don’t see what your actual problem is though.

Coffee is amorphous. At a temperature, it’s more likely to shatter. We measured that with LSA. How is that speculation? Where did we overstep the bounds? Can you pinpoint it for me?

Mark Burness
Guest
Mark Burness

Hi Matt, Is there a chart missing from the post, the one discussed as showing 8.5% volume at 400um particle size (the visible charts only seem to show count & surface area)? You say you know how much of the brew is made up of “fines extractions”, can you elaborate on this, for example, how much have you determined they contribute to a typical extraction? Also, what do you anticipate as being the relative mass of particles below 70um? How have you determined that the most over-extracted particles (or over-extracted flavours) are specifically due to the fines, as opposed to… Read more »

Jordy V
Guest
Jordy V

Matt, as your article shows fines represent an enormous amount of surface area, so getting rid of 10% fines (e.g. 2g for 20g grind) can have a huge impact on the balance of the extraction (by changing the particle distribution). I know you tend to aim for consistency and uniform particle size (EK43 stuff) but I believe some coffees tend to taste better with a different distribution. I guess it’s like a blend of different particle sizes producing a balanced multi-dimentional unity. A nice example of this is the recent win by Todd Goldsworthy at the US Brewers Cup. He… Read more »

Job
Guest
Job

Hi Matt, awesome to see we’re finally walking into the domain of science! Reading your paper, I was wondering about a couple small things: – the sample size seems rather small considering the vast amout of different coffees there are. It would seem to me that more work needs to be done, and with different grinders (those EK’s that didn’t taste so yummy, you could’ve finally checked if the grinding had anything to do with that, and if so, in what way?). – That being said, I think you can’t (definitely) say that origin of process doesn’t matter, because there… Read more »

Matt Perger
Guest
Matt Perger

Speed is easy. Higher rpm = smaller mean and mode. Very linearly.

Seiving is useless. Look at seived grinds under a microscope. There’s millions of fines attached to the boulders. Seeing as 99% of the particles are fines, that’s pretty key.

Nick
Guest
Nick

Not only is the argument about grinding heat made from extrapolated and poorly interpreted data, it essentially demonstrates a negligible change in grind size with respect to temperature. If the mean size difference between -20 to 20 degrees is a single micron, and we can see that the distributions are similarly shaped, we are talking about an absurdly small change in the actual coffee particles of the sample. Then this gets extended on pure faith to the difference between 20 degrees and a hotter bean temp? Talk about surface area all you want, but the amount of actual coffee flavor… Read more »

Job
Guest
Job

Ah I see, excuse me for jumping the gun there. And I know that you’re probably right, but let’s hope others pick up from this start to get some grip on the things we work with.

I’m very curious though on what you make of Colin’s theory that hot beans extract faster (haven’t seen the scientific data that he’s referring to, but would love to) because some of the solubles turn liquid when hot.

Matt Perger
Guest
Matt Perger

Hi Job

Check out the extra details of the paper online and you’ll find the graphs from the misaligned grinders.

Re the coffees. I agree there could’ve been more, but if you think about the level of difference we’re normally concerned about (5 days vs 8 or this lot vs that lot from the same farm) these coffees -which are very different- are incredibly similar!

I’ve done hot beans (not part of this paper) and its definitely a continuation of the effect seen here. It never gets better.

Thanks for chatting!

Matt Perger
Guest
Matt Perger

Hi Jordy. What you’re talking about is usually referred to as a ‘melange’. I firmly believe that that only way this mix of extractions tastes good is when the coffee is somehow flawed and tasting a nice even extraction of it isn’t delicious.

Matt Perger
Guest
Matt Perger

I agree that hotter grinds extract faster. Just as hotter water does. He doesn’t talk about hotter beans getting ground though! 🙂

Jordy V
Guest
Jordy V

Could be… or this could be the typical single malt vs blended whisky discussion… 🙂 I can see why many pro’s like the ‘nice even extractions’ because it’s easier to spot the tasting notes and describe the coffees (something we coffee geeks love). On a different note: the fact that any grinder produces different size ranges implies that every brew is a mix of extractions… my only point was that seiving can be a way of to change that mix. And a last note: any non-immersion brew will result in a melange… the particles at the bottom of let’s say… Read more »

Isaac Loh
Guest
Isaac Loh

Hey Matt, would a dose of strictly 500 micron grounds with millions of fines attached to it, technically has a smaller particle distribution to a dose of 400-800 micron grounds with millions of fines attached to them? Sieving may not be perfect but like you said if 99% of the particles are fines, then wouldn’t it meant that the 1% of the particles play an even more significant role in how our brew are tasting? Has it not been true, it would also mean that essentially what ever grind size we chose on the grinder, all other parameters kept the… Read more »

AndyS
Guest
AndyS

So we no longer have to bother changing out the worn burrs on our grinders. Since “fines are fine,” a set of smooth old burrs will make lots of fines and give us better coffee than a new set. Thank you, “Science.”

Runei Matsumoto
Guest
Runei Matsumoto

Hi Matt, thank you for an interesting experiment. It would be wonderful if you could run some throughly sieved particles into the LSA to see how much fines are attracted to larger particles.

Sieve test graph vs LSA graph would also be interesting to see as well.

Since many of us do rely on test results derived from sieving for simple grinder comparison experiments, having a numerical understanding of its margin of error would help a lot.

Alex
Guest
Alex

Can we brew with frozen grounds or will they cool off the brew water too much?

RM
Guest
RM

Hey @jordy_v:disqus , can you tell me how you tested this?

“And a last note: any non-immersion brew will result in a melange… the particles at the bottom of let’s say a v60 will have had more water contact than the ones at the top.”

I’ve split post-brew v60 coffee grinds into several horizontal sections, extracted individual samples and taken tds readings to measure remaining solubles per sample. I’ve used a special device to control low agitation water flow. I’ve been trying to find a way of measuring this more accurately so any advice would be appreciated.

Daily Dose – April 25, 2016 - The Best Damn Coffee
Guest
Daily Dose – April 25, 2016 - The Best Damn Coffee

[…] 01   Matt Perger breaks down some of the science of grinding coffee that was recently published in Scientific Reports, a sub-journal of Nature. The paper covers some of the variables that can affect the final outcome of a brewed coffee, such origin or the temperature of the coffee beans when grinding. Really interesting stuff! […]

Gregory Levine
Guest
Gregory Levine

Re higher speed of grinding generating smaller particles: do you have any speculation as to the reason for this? More shattering? What about frictional heat?

I also have a question about this experiment specifically. Why isn’t laser particle size distribution affected by static and fines sticking together?

Ian Miller
Guest
Ian Miller

Why hasn’t a grinder like this been developed? We do it on a larger scale with bagging machines, so the technology is there.

Robert Cowles
Guest
Robert Cowles

Hi Matt,

I remember reading somewhere (Schulman?) that particles under 50 micron largely constitute cell walls, and aren’t strongly correlated with soluble material. The assertion was that fines don’t tend to ruin extractions due to their higher surface area and increased relative extraction, but more because they choke the brew and cause larger (non cell wall) particles to be extracted more.

Responses to that in light of your findings?

Matt Perger
Guest
Matt Perger

Thank you!!

And you’re right – the outer surfaces of the largest particles extract at the same rate as the fines. That’s why it’s so important to make every particle as small as possible: so there’s less of a difference.

Matt Perger
Guest
Matt Perger

You can. Use slightly hotter water. Coffee has a far lower specific heat capacity than water so the effect isn’t as great as you might think.

Matt Perger
Guest
Matt Perger

I think soluble material is all pretty evenly distributed throughout the grind sample. Yes, they’re small cells, but isn’t the entire coffee bean made up of cells?

Matt Perger
Guest
Matt Perger

Hi Nick. You seem a little aggressive. I’ll respond a few notches down. I hope you’ll follow. You’re only referring to the mean particle size here. Your arguments all rely on the mean size being similar, but that’s only part of the story. The rest of those curves (especially *at* the mean) are quite different. Separately to this paper, I made espressos with beans ground at varying temperature points. The extractions ranged from 21.5% with -20C beans to 19% with 40C. Between these temperatures the change was quite linear and predictable. You should try it yourself, it’s quite remarkable just… Read more »

Mark Burness
Guest
Mark Burness

Hi Jordy, No one is extracting coffee to 100%, there’s probably only 35% that is even extractable, less from certain coffees, either way, I’m only aiming ~21% extraction yield & this is achievable with whatever number of pours, hot brews & over a wide range of grind sizes. Never tried cold drip, even with hot drip you might well hit limits at either end of useable grind size for manual brewing. Nevertheless, given a typical hot drip scenario, number of pours & their timing only changes the extraction yield if you do not change the grind to compensate.

Jordy V
Guest
Jordy V

Mark, it depends on the extraction efficiency of your method. When that is close to 100% I can imagine that the difference is no longer measurable. This could be the case for fine grinds and/or very hot water. When you use an extraction technique that is far from 100% than multiple extractions will give a higher yield. It’s basic chemistry education… see e.g. here http://chemwiki.ucdavis.edu/Core/Analytical_Chemistry/Analytical_Chemistry_2.0/07%3A_Collecting_and_Preparing_Samples/7.7%3A_Liquid%E2%80%93Liquid_Extractions and have a look at example 7.15 Like I said, the higher the initial extraction yield, the harder to notice the difference… but e.g. with a very coarse grind or low termperature (cold brew?) there… Read more »

Mark Burness
Guest
Mark Burness

Hi Jordy, one extraction of 200ml will reach the same extraction yield as another of 2x100ml, or one of 4x50ml, if you normalise brew time via grind adjustment (for the same coffee), assuming you have efficient wetting at the start of the brew. I generally keep my V60 grind the same & adjust pour regime to account for very soluble coffee (one pour after bloom & stir) & less soluble coffee (multiple pulses after bloom & stir) to hit a similar EY. Randy Pope discussed this. Keep records of your brew times & EYs and you’ll see this is true,… Read more »

Jordy V
Guest
Jordy V

Hi, I didn’t do any specific measurements, it’s just physics…it’s easiest to understand for a V60…in theorie the particle at the very bottom will see all the water, the ones at the very top won’t. I guess you could calculate the amount of water that passes through a certain area, but in practise it’s much more complicated because of blooming effect and agitation (on purpase, by the pooring, by the bloom, by thermal gradients in the brew). And what I’m personally very interested in is the effect of equilibrium situations. What I mean is that fresh water will extract more… Read more »

Alex Bitsios-Esposito
Guest
Alex Bitsios-Esposito

I’m feeling a frozen-vacuum-dose-weighing EK43 hopper in the making. Who’s gonna make it? Mr Klatt?

David Baillie
Guest
David Baillie

Thanks Matt. Great exegesis of the paper. ☺

AndyS
Guest
AndyS

[Note: I posted this reply on Matt’s Facebook link by mistake, but I’m copying it over here because this is where the main discussion is. Thanks.] Hi Matt: I have one really dumb question and one not-quite-as-dumb question about your post. 😉 1. Really dumb question: You graph shows the colder grind samples with smaller particles than the warmer grind samples. Doesn’t it seem like one could (for instance) come very close to the -19C particle size distribution using 20C beans by simply adjusting the grinder a bit finer? And isn’t that what baristas have been doing for like, 50… Read more »

Copyright © 2019 Barista Hustle, All Rights Reserved!

You have Successfully Subscribed!