Prufrocker: The Blog

The Kaminsky Report

Progress in action report from Ben Kaminsky’s Coffee Shots work shop.

by  |   |  Coffee Science, Equipment

Ben Kaminsky in action

This is progress in action. And we can give you up-to-the-minute reports. This is report one:

We went to Ben Kaminsky’s Coffee Shots work shop last night and are at the moment totally convinced that the espresso machine is able to make filter-like coffee amazingly well. An inspiring moment saw us (a group of 10 fee paying trainees) with a Coffee Shot dialled in to 22% extraction yield and Ben says, “Sweetness only goes up from here”…and it did. Right up to 24% it was getting richer and richer and the aroma from coffees we know well was overpoweringly intense and fantastic.

If you don’t know what a Coffee Shot is, check this out. WBC 2013 Finals: Matthew Perger, 2nd Place. The name was coined by the fantastic Barismo in Cambridge, Massachusetts some years ago it is more or less a Lungo.

We learned what grinder is best for making Coffee Shots and now we won’t stop until we’ve got one. I can tell you as of 10pm on the 8th of August, Coffee Hit has the only one on sale in the country for the next five weeks. The EK43 by Mahlkonig. We’ve even done a Keep Cup mosaic that spells EK43. Evidence is pointing to the EK43 as having the most modal grind distribution of all grinders. The grinds are all mostly the same size.

All we really knew about fines before Ben Kaminsky came along was that they were cell wall particles and a coffee cell is about 50microns (a micron is 1000th of a mm). We knew all this from Scott Rao’s book Then the fantastic Matt Perger entered the frey and used a 250 micro sieve to win the 2012 World Brewer’s Cup and we really started thinking about fines. But it’s not just fines you need to worry about. There’s the other side of the particle size distribution curve where there are the over-sized grinds. Scott Rao uses the term boulders (which has always seemed a bit overstated to me) to describe the big bits and it has become clear we need another sieve. Ben was suggesting we should all start using a 1000micron sieve (1mm) to remove boulders. Boulders are bigger than the space between the blades because they slip through side on. They get planed, but not properly ground to the standard deviation.

Prufrock learned a lot from Isa the manager at Talk House when she won the 2013 UK Aeropress competition earlier this year. She used a Tanzania and didn’t thwack the thwacker at the top after grinding so as to leave out the chaff that always falls out at the end. This was something she did that no one else did and it must have made a difference. Now Kaminsky has handed us a very quick easy solution; a big sieve. Chaff on its own tastes like pop corn, grassy and grainy and papery. And a lot of it doesn’t get broken up by the blades.

He wasn’t really endorsing the 250micron sieve at commercial brew bars as you need to have a damn good go at it to remove absolutely everything under 250micron in less than 5min of hard shaking. If you remove some but not all of the fines, the amount you remove for each brew may vary and make subsequent brews too hit and miss. You would probably be more consistent if you left everything as it was. Really he was saying, get a grinder where you don’t need to sieve to achieve modal grind distribution and failing that, remove the boulders and what chaff you can every time in just a couple of seconds just by dropping it through your 1mm big sieve.

So a big (aperture) sieve and a small one, or an EK43…or all three. The EK still produces some fines btw, just much less. The best solution of all would be to buy a proper laboratory test sieve set and shaker. Let’s not rule this out. Might be a bit noisy.

Refractometry cynics have always maintained TDS readings are an average of different sized grinds extracted to different extents so what’s the point? Ben had data on average extraction yields for different particle sizes. If I remember correctly, the particles under 250microns constitute about 25% of the total mass of grinds for your average filter grind and they extract to a very high level, like 27% and taste bitter. (James Hoffman said the other night in a talk at Look Mum No Hands, “The most you can extract coffee to is about 30%, the other 70% is basically wood.”)

Boulders extract to 13% on average and constitute about 10% of the total ground mass and make the coffee taste sour. And because we are zoning in on 19-20% extraction yields at Prufrock and one quarter of the coffee is extracting extremely high, we are tending to prefer brews where the modal group of particles (the biggest group) are under extracted to 16 or 17%. The average gives us better cups in the gold cup range 18-22% but we are always drinking over and under extracted tastes too.

No coffee professionals of note really dispute the merits of the Gold Cup standard, 18-22% extraction yield. That is until Kaminsky came along, and his early adopter buddies Colin Harmon, Matt Perger and rumour has it Tim Wendelboe is getting an EK, or has one. Colin is totally going for it. The 3FE staff are loving it and they are pre-weighing 350 doses before work each day.

Ed Gooding the legendary Mahlkonig go-to guy in the UK assures me the retention is zero, if you don’t sweep out the collar each time you grind. There are a couple of cavities in the throat that fill up with coffee on the first dose and there after are filled and retention is negligible. 3FE are dosing to within the weight of a single bean. I can happily confirm one bean doesn’t weigh more than 0.2g. I just weighed 18g of Honduran Pacas and there were 109 beans. That makes the average bean weight 0.165g and I weighed out 18g of Kenyan SL28 and SL34 Peaberry and there were 119 beans with an average bean weight of 0.151g. These are high density light roasted Square Mile filter roasts so they are heavy beans. Colin is very fond of Pacamara so he might be in trouble but this is an exceptional margin of error.

Kaminsky worked a lot on the Marzocco Strada. He seems to have concluded that pressure profiling makes only trivial difference to espresso extraction with ordinary espresso grinders. His experiments were done with a Robur E. He considers it might not be perfect for making espresso but it is very well suited to making Coffee Shots. He mentioned he favours 8bars of pressure as a standard.

The real important thing about coffee shots seems to be that the flow is not restricted by the grinds but rather by the flow restrictors, pump pressure and basket hole sizes. (Bigger baskets have bigger holes, and bigger holes means less flow ristriction.) So the surface area of the grinds is the variable that affects extraction the most. Just like with filter coffee. But unlike with filter coffee made with a pouring kettle, the espresso machine makes temperatures really stable. And contact time becomes a given because the grinds aren’t fine enough to alter contact time with the EK. Also the EK produces so few fines they don’t create any significant blockages. So it’s all back to grind size.

Also a tamped coffee bed in a portafilter he describes as a static flat bed. This is another advantage of using the machine to help keep things stable because turbulence isn’t a factor. Just the erosion of the outside of grinds.

Actually, Kaminsky gives Dave Walsh from Marco a lot of credit in his talk for introducing him to sieved coffee several years back and the joys of 24% extraction. We are sieving our little hearts out on the brew bar for our coffees to share, we have started removing the boulders and what chaff we can with a cheap tea strainer but we need to get something more critical. And we are more than ready to start serving Coffee Shots, from the Tanzania. Tanzania is a modern state of the art grinder that doesn’t look like E.T nor was it developed in the 80s. I can’t help feeling some chagrin that the world’s most modal grinder reemerges from the 80s and we haven’t come up with something better. Also we are needing to establish how much the E.K is going to be able to withstand high usage without heat recalibrating the grind aperture. Colin can tell us. I’ve left him a message on his answering machine.

Our Petra is leaving us at the end of the month to open a cafe in Czech Republic and she’s excited about being able to use one grinder for her espresso blend, her decaf, her guest coffee(s) and her brew bar. This is potentially a huge saving for her even if the EK retails at around £1500-2000.

I think Ben’s message is to experiment and is as much about the merits of the espresso machine as it is about the EK. The espresso machine is an extremely stable flat bed brewer with the most temperature-stable water in our building. And the Tanzania isn’t so bad itself for modal grind distribution so expect coffee shots on the menu this week and an EK in no more than 5.

/ /