Espresso: high extraction yields

[MurderousCrow]

Hi,

One for the espresso experts:

In the last few years it seems a few people have been experimenting with really high espresso extraction yields using EK43 grinders. This is pushing the envelope of what espresso is IMO, but they talk about a sweet and very pleasant coffee that can be found at super-lungo volumes from relatively low initial doses. Like a 21g dose going into a 48g yield over 27 seconds.

If I understand it correctly, the reasoning is that the EK43 in particular is gifted with the ability to grind coffee (at espresso grind sizes) consistently enough to allow optimal extraction, minimising the presence of bitter end-extraction products in the cup.

My experience with higher end kit was a while back, and used pretty commonly-accepted definitions of dose to yield, e.g.

20g dose to 20-30g yield; ristretto (ratio 1:1 - 1:1.5)
20g to circa 40g yield; normale (ratio circa 1:2)
Anything over that; a lungo

But even with a really great grinder, pushing that lungo boundary too much was not practiced. The usual caveat on the trend toward disproportionately increasing bitterness, while increasing actual extraction only a little, in the final seconds of a shot.

Has anyone played around with really high yields like these? If you've done it and succeeded in making this type of coffee well, did the shot visually appear to be extracting properly throughout the process?

This might be a little out there, hope someone has some real-world experience with this method. Especially using grinders other than the EK43. I'm mostly just curious to see what's possible with the current crop of high-end grinders - the EK43 is 30 years old, and I'd be surprised if it truly were the only grinder capable of this process.

Luke

 

[ruckin]

I have read about yields but have not played with my process to actually measure or do anything about it. Someday I might start looking into it. First I want to upgrade my zillion year old mini mazzer with a proper grinder and then I need to get my hands on a shot scale so I can actually measure the shots accurately. Until then I will continue to be inefficient by using larger portafilter baskets and pulling solidly in the ristretto range.

 

[naughtilus]

Higher yield may be too much of a good thing according to this:

 

[MurderousCrow]

I have read about yields but have not played with my process to actually measure or do anything about it. Someday I might start looking into it. First I want to upgrade my zillion year old mini mazzer with a proper grinder and then I need to get my hands on a shot scale so I can actually measure the shots accurately. Until then I will continue to be inefficient by using larger portafilter baskets and pulling solidly in the ristretto range.

The basic idea of dose vs yield (disregarding high-yields altogether for now) is a really interesting one. It quantifies how the volume of a drink is far less relevant than its ratio, in terms of how someone may prefer a coffee. And it can offer guidance on how you might want to approach a particular bean; say you like a darker roast or want to accentuate caramel-type notes, you may want to aim for a ristretto yield. Whereas lighter roasts may benefit from a greater yield to allow those floral / fruity notes to emerge.

But it's not really possible to use dose vs yield with any confidence if equipment ain't up to it; I've used a scale for a few years now, but never well. It was an exercise in frustration really as I usually had to tighten my Rocky grinder as far as it would go or one step back, then adjust dose to get a reasonable pour time & volume. So day-to-day consistency in terms of shot volume and yield ratio was really poor.

What I've realised is that while over-extraction is easy to minimise as far as possible with a naked PF, I was often under-extracting massively, even when cutting the shot just before blonding, because the grinder was really bad. Hence a large amount of high-acid shots: if yield was pushed higher I'd get blonding and really bad bitterness so I knew where to stop, and accepted a half-decent pour as a win, even if it was acidic-tasting and under-extracted in flavour. Eventually I lost patience and just fell into the habit of making whatever shot I could get, usually for a long black. Don't get me wrong, not all shots were bad and some were exceptional but it was hit and miss for a multitude of reasons, mostly the grinder.

With a better grinder such as yours one can more reliably aim for specific yields as you have the latitude to adjust finer or coarser as required. I'm pretty confident you could get very nice normale and lungo yields, I've used Mazzer mini and majors before at these strengths and got very nice espressos. If the beans are evenly developed (and you've done everything possible to ensure a good distribution), you will avoid simultaneous over and under-extraction in the puck; the entire bean has the same consistency and the grinder ensures a sufficiently-tight spread of grind sizes.

(Couple of thoughts - if you experience bad pours at higher yields a fix might be to simply replace your burrs if they're old, apologies if teaching grandma to suck eggs. And re: scales I use a Triton T3 400, which can handle up to 500g at a 0.01g resolution. They're cheap and built pretty tough, although be careful around water.)

Higher yield may be too much of a good thing according to this:

Thank you. I'm not able to see that video ('net is really bad here at work) but most of the reading I've done on high yield espresso has been from Matt Perger, who's the founder of BH. He alludes to high yields not necessarily being good in a couple of different ways, and an important one is that not all beans will respond the same way. Is he adding further caveats there? He's also highly enamoured of the EK43 and names only that grinder as being capable of pleasant-tasting espresso at these high yields. I'm sceptical; the '43 is an amazing bit of kit but I doubt its tolerances are really that much better in the real world.

With normal espresso grinders, these people are saying their consistency for really high yields that are pleasant to drink is insufficient. If yield is pushed, one gets some degree of over and under-extraction simultaneously. Because the spread in grind sizes is still relatively wide: some particles over-extract and some under-extract, and the wide latitude in particle size is what's responsible. It's the same problem I encountered with my Rocky, only the tolerances are higher.

I'm curious though, as to why the EK43 is so much better for this style of espresso against other grinders. Mahlkonig published data showing how the '43 achieves a tighter spread of particle sizes compared with other high-end grinders (Mazzer Robur, Mahlkonig K30, Anfim Super Caimano), but to my untrained eye the differences in the graph seemed trivial. All the grinders tested showed a very similar and tight spread in the way fines were generated, which is surely the most crucial factor in avoiding over-extraction when pushing yield.

The EK43 Part Two

Understanding more about the EK43 and why it makes coffee taste better. Part 2.

baristahustle.com

 

[naughtilus]

The basic idea of dose vs yield (disregarding high-yields altogether for now) is a really interesting one. It quantifies how the volume of a drink is far less relevant than its ratio, in terms of how someone may prefer a coffee. And it can offer guidance on how you might want to approach a particular bean; say you like a darker roast or want to accentuate caramel-type notes, you may want to aim for a ristretto yield. Whereas lighter roasts may benefit from a greater yield to allow those floral / fruity notes to emerge.

But it's not really possible to use dose vs yield with any confidence if equipment ain't up to it; I've used a scale for a few years now, but never well. It was an exercise in frustration really as I usually had to tighten my Rocky grinder as far as it would go or one step back, then adjust dose to get a reasonable pour time & volume. So day-to-day consistency in terms of shot volume and yield ratio was really poor.

What I've realised is that while over-extraction is easy to minimise as far as possible with a naked PF, I was often under-extracting massively, even when cutting the shot just before blonding, because the grinder was really bad. Hence a large amount of high-acid shots: if yield was pushed higher I'd get blonding and really bad bitterness so I knew where to stop, and accepted a half-decent pour as a win, even if it was acidic-tasting and under-extracted in flavour. Eventually I lost patience and just fell into the habit of making whatever shot I could get, usually for a long black. Don't get me wrong, not all shots were bad and some were exceptional but it was hit and miss for a multitude of reasons, mostly the grinder.

With a better grinder such as yours one can more reliably aim for specific yields as you have the latitude to adjust finer or coarser as required. I'm pretty confident you could get very nice normale and lungo yields, I've used Mazzer mini and majors before at these strengths and got very nice espressos. If the beans are evenly developed (and you've done everything possible to ensure a good distribution), you will avoid simultaneous over and under-extraction in the puck; the entire bean has the same consistency and the grinder ensures a sufficiently-tight spread of grind sizes.

(Couple of thoughts - if you experience bad pours at higher yields a fix might be to simply replace your burrs if they're old, apologies if teaching grandma to suck eggs. And re: scales I use a Triton T3 400, which can handle up to 500g at a 0.01g resolution. They're cheap and built pretty tough, although be careful around water.)



Thank you. I'm not able to see that video ('net is really bad here at work) but most of the reading I've done on high yield espresso has been from Matt Perger, who's the founder of BH. He alludes to high yields not necessarily being good in a couple of different ways, and an important one is that not all beans will respond the same way. Is he adding further caveats there? He's also highly enamoured of the EK43 and names only that grinder as being capable of pleasant-tasting espresso at these high yields. I'm sceptical; the '43 is an amazing bit of kit but I doubt its tolerances are really that much better in the real world.

With normal espresso grinders, these people are saying their consistency for really high yields that are pleasant to drink is insufficient. If yield is pushed, one gets some degree of over and under-extraction simultaneously. Because the spread in grind sizes is still relatively wide: some particles over-extract and some under-extract, and the wide latitude in particle size is what's responsible. It's the same problem I encountered with my Rocky, only the tolerances are higher.

I'm curious though, as to why the EK43 is so much better for this style of espresso against other grinders. Mahlkonig published data showing how the '43 achieves a tighter spread of particle sizes compared with other high-end grinders (Mazzer Robur, Mahlkonig K30, Anfim Super Caimano), but to my untrained eye the differences in the graph seemed trivial. All the grinders tested showed a very similar and tight spread in the way fines were generated, which is surely the most crucial factor in avoiding over-extraction when pushing yield.

The EK43 Part Two

Understanding more about the EK43 and why it makes coffee taste better. Part 2.

baristahustle.com

The key claim of the presenter is that even extraction is crucial for bringing out nuances and flavor character in coffee. High yield leads to muddy overpowering complexity on the palate, so every coffee defaults from flavors to intensity. It begins tasting like an espresso in generic sense but stronger (more TDS) .

Even aka unimodal grindind brings transparency to flavor profile so specialty coffee can be experienced in its full individuality. If that EK43 grinder does more uniform grinding rather than just finer, you should in theory accent the flavors without overextraction and raising intensity.

 

[MurderousCrow]

The key claim of the presenter is that even extraction is crucial for bringing out nuances and flavor character in coffee. High yield leads to muddy overpowering complexity on the palate, so every coffee defaults from flavors to intensity. It begins tasting like an espresso in generic sense but stronger (more TDS) .

Even aka unimodal grindind brings transparency to flavor profile so specialty coffee can be experienced in its full individuality. If that EK43 grinder does more uniform grinding rather than just finer, you should in theory accent the flavors without overextraction and raising intensity.

Thanks for outlining the gist of the video for me, that's much appreciated Yes that is my understanding too - the more uniform the grind, the better for evenness of extraction. I know what you mean when talking about higher yields muddying the taste, although I was unaware of how that worked when I experienced this; I just thought it was odd-tasting espresso

Unimodal however is a bit of a mythic ideal though really, at least from what I can see. The data in the Mahlkonig graph linked shows clearly that the top-end grinders used, all produced a significant quantity of fines.

The current dogma seems to be that big, flat-burr grinders are better at producing unimodal grinds, whereas conical grinders are 'inherently bimodal' (source: coffee forum ). But to me, in terms of consistency of output, the conical-burr Robur looks very impressive. It produced the smallest amount of fines, as well as the lowest ratio of fines. I acknowledge it had the widest spread of the tested machines (the graph is not linear on the x-axis so eyeballing curve widths to compare is grossly inaccurate), but it was far from bi-modal in that it probably had the most consistent output of any of the grinders tested (smooth curve with a minimum of peaks).

What I'm curious about is whether anyone has tried to hit high extractions and made palatable espressos with any high-end grinder. I acknowledge the EK43 has features which physically set it apart from other grinders (a 98mm flat burr set for instance!) but I do find it hard to believe it's the only grinder capable of making this enjoyable, clearly-flavoured high yield stuff.

 

[naughtilus]

Unimodal however is a bit of a mythic ideal though really, at least from what I can see. The data in the Mahlkonig graph linked shows clearly that the top-end grinders used, all produced a significant quantity of fines.

These small output countertop grinders can't do true unimodal regardless of price. I've used Macap/Fiorenzato, Mazzer, Rossi etc. and they don't come close to the uniformity of some manufacturers preground coffee. You have to sift if you want consistency with small grinders. Grind coarse, sift, grind finer, sift.

Size of the burrs is obviously the advantage. There's nothing sophisticated about an EK43. Size matters.

(not my photo)

 

[MurderousCrow]

These small output countertop grinders can't do true unimodal regardless of price. I've used Macap/Fiorenzato, Mazzer, Rossi etc. and they don't come close to the uniformity of some manufacturers preground coffee.

Your personal experience there is really interesting! In one way I'm quite surprised that these high-end counter top grinders we're talking about really don't perform that well. In the same vein, the Mahlkonig test graph was eye-opening, as it gave some insight into what passes for 'unimodal' even at the pinnacle of 3rd-wave coffee development.

You have to sift if you want consistency with small grinders. Grind coarse, sift, grind finer, sift.

Really useful, thanks. But it's sending me down a rabbit hole of sifting for espresso. I can see Kruve filters in my future, it's uncanny

Size of the burrs is obviously the advantage. There's nothing sophisticated about an EK43. Size matters.

Solution - everyone clearly needs 98mm+ flat burr grinders

 

[ruckin]

Oh great! Now I need to change my lobbying efforts from telling SWMBO that we need a large conical burr grinder to one with monster flat burrs.

 

[MurderousCrow]

Oh great! Now I need to change my lobbying efforts from telling SWMBO that we need a large conical burr grinder to one with monster flat burrs.

Hahaha! Show her this thread, I'm sure it will help her understand the requirement for such

I'm in a sweet place right now despite spending a fair bit of late; my partner was actually impressed by my new machine's output... And it was helpful that she's a frequent drinker of take-out coffee. So when I showed her that a home setup can save money, she was muy feliz:

Invest in home espresso and save over $1k per year.

While the initial sticker price can seem daunting, home espresso is a terrific long-term value. See how much money you'll save with our online calculator.

clivecoffee.com

 

[MurderousCrow]

'By lowering the dry coffee mass and grinding to maximize EY, the operator may notice that they are able to push their extractions much higher than before, while achieving highly reproducible espresso.'

An update on increasing / optimising espresso extraction yields (EY) for anyone interested, with a cool bit of research published in the last few days. Essentially some researchers physics-ed the heck out of espresso extraction. They outline their findings in detail here. Heavy on the maths, frankly most of it is way over my head, but in essence they're saying we currently grind far too fine, which has negative effects both in shot reproducibility and waste. And the recipe they arrived at was interesting: 15g ground coffee, 7-15sec pour time to produce 40g espresso. I could do that one.

Background: using computer simulations to model water flow under 6bar of pressure through a standard puck of espresso-ground coffee, the researchers found that espresso extraction yield is adversely affected by grinding at industry-standard settings; the simulations and experimental evidence clearly suggested that this is far too fine. They concluded that the presence of overly fine grinds causes a reduction in EY, by disrupting flow within the coffee bed. Non-homogenous flow is in general a bad thing, as it leads to simultaneous over-extraction of some parts of the puck, and under-extraction of others. Counterintuitively, they showed that one can reliably get the same EY from a 15g dose run over <15s and ground quite coarsely, as from ~20g dose ground at more traditional grind settings, with far improved reproducibility.

A downside of this method may be reduced flavour complexity in comparison with a traditionally-brewed espresso at the same EY, but from what I can see that's a simple consequence of reducing the amount of over-acidic and over-bitter flavours.

Personally I'm good with sweet-tasting espresso, and always aim for that anyway - whatever gets me there is absolutely fine, I'm not bothered about dogma. So I tried their recipe a few times using slightly different methods. First time I used my Kruve sieves to sift out the really small fines and big boulders, using a wide range of 400-1000nm. That's way bigger than traditional espresso, visibly so. I used 10g of sieved grounds, poured over about 11sec, into 30g espresso - a pretty lungo ratio but not worlds away from what was suggested. In the pour it started nicely with a 3-4sec interval to first drops, and symmetrical output from the bottom of the basket (using a naked PF). It quickly blonded out, and began gushing just as I cut the shot. Crema was present, but not thick. This did not bode well. But in the end the taste is all that matters, and it was surprisingly good.

For the rest of the shots I've tried since, I simply dialled my grinder up to a setting more suited to aeropress and haven't bothered sifting. The taste is consistent: slightly more bitter than the sifted shot, but still really pleasant, as sweet as the better conventional shots I've made of late. For some shots I've pushed extraction higher, like 15g grounds to 50g espresso - and the taste is pretty good. But for most I've stuck around what the researchers outlined.

No, this recipe is not a dense Italian ristretto and tastes nothing like it. And the method is certainly much better suited to lighter-roasted coffees. But IMO, and certainly if you've read this far - it's worth a shot!

 

[ruckin]

I might have to play around with this a bit. Next week I replace the pump in the GS3 so why not introduce multiple variables so I really don't know what is going on

 

[MurderousCrow]

Haha go for it... Do what I did, play with it just for the hell of it, you might be happy with the results!

 

[ruckin]

lol - while being happy with my coffee is important... keeping SWMBO happy with the coffee is waaay more important. I roasted a couple kg today so I will have some fresh coffee to play with.

 

[MurderousCrow]

I know EXACTLY what you mean. My lass' disgruntled face when I'm messing around before she gets a brew? I can be a pretty quick barista when needs be