another_jim wrote:This sort of tweaking isn't the custom for brewing; but if it were, I would be somewhat shocked if it didn't do some good too.

Tweaking for brew seems useful to me, at least anecdotally. I haven't done any blind testing, but I routinely "bracket" my brew grinder setting when I try a new coffee and it's common to find that one or two steps away from "normal" tastes best for a given coffee. I don't experiment with brew dose too often but it's something that I should try more.

.

another_jim wrote:The literature reports slightly higher figures for espresso, around 25%, than the 20% for brewed. Andy says actual solubles are the same as brewing, around 20%, but that the pressure extraction adds oils and insolubles that make up the difference. I got between 17% and 23% depending on grinder, coffee, dose and basket.

I think one has to be careful to distinguish solids yield (everything removed from the dry coffee, delivered into your cup) from solubles yield (only the compounds that are dissolved in your cup). With metal filtered coffee like espresso and french press, undissolved material can easily comprise 8-12%% of the total solids yield. This extra 8-12% of "stuff" undoubtedly adds to mouthfeel, but I don't think it changes the basic flavor balance (ie, sour-sweet-bitter). So if you're interested in flavor balance, filtering out the insolubles before measuring the amount of material in solution works well, in my experience.

I don't find the 25% extraction yield quoted by Illy to be credible, no matter how it's measured. That would be a terribly bitter cup. In fact, I usually find that coffee/espresso/etc gets bitter over 20% solubles yield. I taste the most sweetness around 19%. Coffees extracted at less than 18% can be OK, but they almost always lack sweetness (to my taste buds). I find these trends pretty consistent across a bunch of different brew methods, including espresso.

But espresso is a peculiar beast. Because the isolated extraction chamber allows no agitation or mixing of grounds during the brew cycle, the extraction yield range is fairly limited. It is not nearly as easy to explore the x-y "extraction space" (from higher strength brews to lower strength brews and from higher extraction yields to lower extraction yields) as with brewed coffee.

Also, since an espresso puck tends to be extracted unevenly (the top of the puck more and the bottom less), one has to be careful not to pick up too much of the bitter, overextracted flavors from the top. Someone with very sensitive taste buds, like Scott Rao for instance, likes his espresso extracted to around 18-18.5% to avoid the bitters. Me, I usually go a percent higher.

another_jim wrote:The more I do these experiments, the more skeptical I become that the best taste occurs at 20% for all coffees and brewing methods. Most of that 20% solubles are starches and other compounds with very little impact on the taste. So the overall solids yield is not, in itself, the direct cause of the taste, but an indicator of what the flavor carrying molecules are doing. There is neither any evidence or any logical warrant to assume that the relation between total solubles and flavor solubles is the same for all coffees, roasts and brewing methods.

There certainly is a wide range of coffee to explore, from very light to very dark, high grown to low grown, wet processed to dry processed, etc. But thanks to the ExtractMojo refractometer, I think evidence is rapidly mounting that a relatively narrow extraction yield range is pretty consistent in producing the most pleasing cup.

I just ran the same sample from the OP to check to see if it has changed. The ziplock bag has been moved around the lab and spent 2 days commuting with me in the car. The mean size is <2% different and the distribution overlays nicely with the original.

That's good news. It may means there's a good chance we'll get some good and interesting information from the individual grind samples.

The main investigation will be based on the difference between the high and low dose sample from each grinder. This means that all confounding idiosyncrasies which affect both samples equally, like storage and mailing differences, as well as skills, machine or basket differences, will get cancelled out.

This is know as a repeated measures design, the design most experimenters prefer for its simplicity, and which is mandatory for small scale testing. It is more immune to noise than other designs, so it requires an order of magnitude fewer samples for a given significance level in the result. Moreover, confounding variables are cancelled by using the difference in each subject; so it saves the expense of identifying, measuring and analyzing all the possible confounding variables.

Oddly enough, these facts, which are taught in every 101 methods class, seem to have escaped or been forgotten by the self appointed experts on scientific method and experimental design opining in these threads.

The method maay work for single dose grind, but fails with a loaded hopper as an ever changing variable, bean column, is not nulled..

Loaded hopper contributors should purchase 2#, place 1# in the hopper and replace shot weight for each test.

JonR10 wrote:This is a terrific opportunity and I'm excited about the possibility!
Kendall - thanks very much for your contribution and for your offer to examine more samples.


I'd be happy to contribute if I can. I think if we can grind samples of a given coffee to produce specific flow/volume with similar baskets from similarly pressure-limited machines then the samples could come from multiple places - allowing for more grinders to be compared.

Would not everyone have to roast the exact same varietal from the exact same lot with the exact same roaster to (as close as possible) the exact same roast profile?
I admit total ignorance here, but I can say that roasting beans in a StirCrazy convection combo dries the bean out more than does roasting with my Behmor. I know this as my finished Behmor roasts weigh more (less moisture evaporation)

Would not the grinding of a drier bean produce more flakes and odd pieces.
Again...PLEASE EXCUSE my intrusion in this conversation. Just curious.
I'll stay out and just observe.

CCC is preparing a single roast of Aficionado for all who want to submit a sample