peacecup wrote:The moisture content doesn't change much after the coffee is ground and left to air dry. Less than 0.5% unless there are extremes in humidty.

Is 65% humidity and 65°F extreme? It's pretty typical for the SF Bay Area.

Look at the graph that I posted. It's from the paper "Water Sorption Isotherms of Roasted Coffee...", Cepeda et al, 2001, Intl J of Food Sci and Tech. You can find a similar one in Illy. At 30% humidity the equilibrium moisture would be about 5% or so. Coffee from the roaster can be under 2% in which case you'd expect a significant change.

I placed a sample of freshly ground coffee in a tray next to the extracted grounds. It gained 5.6% in weight.

OldNuc wrote:When you air dry an extracted puck then the results will match the expected.

What do you mean, "expected"?

Result. The moisture content of the unextracted dose and the dried extracted dose should match. The climate in SF is almost unique and is not common across the country. This is going to make these determinations a bit difficult. The end result of extraction calculations are unique to the methodology being used and the local environment. They are not directly compatible to the results of others.

OldNuc wrote:Result. The moisture content of the unextracted dose and the dried extracted dose should match.

Why should they match? Extracted and unextracted grounds are different physically and compositionally.

Several years ago I noticed that dehydrated spent grounds absorbed moisture at a much faster rate than dehydrated fresh grounds. I wondered then if they would stabilize at different moisture levels. Now I have one data point: At 18°C and 65% humidity, unextracted grounds leveled off at 8.2% (which is in line with that graph) and extracted grounds at 10.6%.

Using those numbers I get an extraction of 17.7%. That's pretty far off. Dehydrated liquid and dehydrated grounds give values of 19.8% and 19.9%, respectively.

The error would have been less if I'd lived in Phoenix. But SF isn't unique; many places have humidity above 50% right now or at some time in the year. If the environment doesn't change by very much then you could obtain internal consistency by air drying. Peacecup is measuring 14% extractions and I'm wondering what they really are. If it were me I'd want to know. And the thing is, you could find out in a couple of hours whereas it takes a week to get an inaccurate number.

My point is they are all relative. You will not get the same numbers as I will as the environment and coffee are different. The values you obtain are all related to you and your process etc. but not necessarily directly related to any other persons' numbers if taken in direct comparison to each other as an absolute value. Just the variation in unextracted coffee moisture content vs. air dried or oven dried extracted coffee moisture content will impact the calculated value. These extraction calculations are an approximation of the actual extraction as the change in moisture content is inducing an uncompensated error. It appears that this error may in some cases be significant depending on the environment and roast level. I suspect that using a southern Italian roast level coffee will have a smaller error induced when using an artificial heat dried extracted sample than a lightly roasted coffee and the same or similar drying technique.

First of all, the moisture content isn't that hard to measure. For dehydrated grounds it is essential to get that number and is the reason why the extraction I calculated that way was in such good agreement with the liquid dehydration number.

The environment is irrelevant unless your process relies upon it. The way I validated my dehydration method was to mail samples of coffee across the country to someone who measured them completely differently, in a different environment, but with a validated technique. The numbers matched almost exactly.

What does change is the coffee. I'm not trying to argue that taste correlates one-to-one with extraction yield from one coffee to the next. That's obviously not the case.

What I am saying is that air drying grounds results in an inaccurate number in general. And when using that method, whether or not you have an internally consistent result is up in the air -- so to speak. Unless you dehydrate the grounds to check you really don't know.

If you do not account for the moisture content in both the fresh ground and extracted sample then you have an error. If you heat dry the extracted sample then you should have a boggy moisture value to subtract from the fresh ground or there is an obviously uncorrected error.

Further thoughts.
There would be some difficulty in determining just the moisture content of the fresh ground coffee as the light aromatic oils might be driven of with the moisture if the idea was to just dehydrate it by heating.

OldNuc wrote:If you do not account for the moisture content in both the fresh ground and extracted sample then you have an error. If you heat dry the extracted sample then you should have a boggy moisture value to subtract from the fresh ground or there is an obviously uncorrected error.

The usual method with a refractometer doesn't include measuring the moisture content. But the extraction uncertainty is smaller because it's in the denominator. For every percent of moisture in the coffee the extraction shifts by about 0.2% -- when measuring the concentration of the beverage. But when drying the puck extraction is determined by difference so each percent of moisture translates into a full percent error in the extraction. That's why it's critical to measure it in that case.

OldNuc wrote:Further thoughts.
There would be some difficulty in determining just the moisture content of the fresh ground coffee as the light aromatic oils might be driven of with the moisture if the idea was to just dehydrate it by heating.

That may be true, I don't know. There is also the question of CO2 loss. However, dehydration is the standard and it is what that graph I posted is based upon. Also, the brewed grounds undergo the same treatment, so loss of "moisture" -- whatever that is composed of -- is the same for both puck and fresh sample. The bottom line is that the mass of solids in the cup and in the puck -- minus the baked off volatiles -- add up correctly.

I can't argue that extraction isn't a fuzzy measure of brewing quality. It is. But I don't see that as a reason to make it fuzzier.

jpender wrote:... I can't argue that extraction isn't a fuzzy measure of brewing quality. It is. But I don't see that as a reason to make it fuzzier.

That is why I filed it along with puckology.