Measuring espresso extraction yield by weight

Beginner and pro baristas share tips and tricks.

Postby cunim » Mar 30, 2019, 1:08 pm

I am interested in my extraction yields for different conditions of preinfusion. I recall reading that you can simply weigh the puck before and after so I dump the grinds into a small bowl, distribute them, and leave them on the warming tray for a day. They seem thoroughly dry when I weigh them. I get yields in the 21-22% range from medium roast beans. This is from grinding fine, performing drip preinfusion (just boiler pressure for about 20 sec), and then brewing for about 33 sec 22g in, 54 g out. These shots tend to be smooth and exhibit less acid than shots ground a bit coarser and without preinfusion. Clearly, drip preinfusion has a repeatable effect so I am encouraged to try some lighter roasts.

Does anyone know of a standard method for calculating extraction yield by weight? I suspect I am doing something wrong.

User avatar
Team HB

Postby another_jim » Mar 30, 2019, 2:00 pm

Assuming that you are collecting the entire spent puck, including everything stuck on the shower screen, the formula

1 - dried_spent_puck_weight/dose_weight

will be a slight underestimate of the EY. The reason is that the puck is wet with coffee, not water, so that the dissolved solids from the coffee redeposit on the puck as it dries. This adds some weight to the spent puck and underestimates the EY.

If you are simply comparing EYs from two different techniques, there is no need to bother with a correction, since both dried pucks will be slightly heavy. However, if you want an accurate estimate, you will need to compensate. You'll need to weigh the shot and the wet puck. The sum is the weight of water plus the weight of the extracted coffee. This means that ...

(wet_puck_weight - dry_puck_weight) / (shot_weight + wet_puck_weight - dry_puck weight)

Is the proportion of extracted solids redesposited back unto the dried puck, call it red_pro. Therefore,

dry_puck-weight = dose_weight - extracted_weight + red_pro*extracted_weight

solving this equation for extracted_weight ...

extracted_weight = (dose_weight - dry_puck_weight) / (1 - red_pro)

Ta da!
Jim Schulman
★ Helpful


Postby jpender » Mar 30, 2019, 3:33 pm

That's an interesting take on the notion of extraction yield in percolation brewing, that one should include dissolved solids left behind. This is at odds with the way one determines extraction with a refractometer. It also makes an assumption as to the concentration of those dissolved solids in the puck liquid. How do you know that it's the same as the %TDS of the shot?

But setting aside the question of whether that fudge factor has merit, it seems to me the biggest source of error will come from the moisture contents of the dose and "dry" puck. What seems dry may in fact contain some percentage of moisture. And when determining extraction by weighing the puck each percentage error translates directly into an error in the extraction yield. It's highly unlikely that the dried puck and the dose have the same moisture content.

Any undissolved solids in the shot will introduce an additional error.

I wouldn't put much stock in the method for determining a portable number. However, if you do everything the same way with the same coffee you stand a good chance of obtaining relative numbers that might be useful. But it could be argued that it makes more sense to just see if it tastes better.


Postby cunim » Mar 30, 2019, 3:52 pm

What useful answers! Thank you.

At the moment, I just need to know I am in the ball park. From Jim's explanation, I am probably underestimating extraction a bit using a simple post/pre dose measurement. Granted, this value is not portable but if I am seeing EY of 21-22% it's all good. Actually, I am a bit uncertain about refractometry. It is a correlate of yield as opposed to a direct measurement, so it would be portable within a rather narrow use case as well. Please correct me if I am wrong.

The point about taste is fair. These shots do taste well extracted. For example, one of the coffees is a third wave medium roast with a distinct juicy red wine taste. Very unusual. At the roastery, this was bright and complex without being sour. My own first pulls were bright verging on tart (at various ginds and ratios) and that tartness dominated other flavors. I suspected under extraction (aided by poor barista skills). Sadly, my tank-fed Synchronica lacks true PI, but it does drip onto the puck at a pretty fair rate. I can saturate 22g in <30 sec. I usually just go to about 20 sec and then pump.

The drip PI shots are similar to the roaster's sample. Red wine is there and muted acidity, so I have solved the dominant tartness. The shot is pleasant, but less clear and complex than I remember. I believe that the longer I drip, the more mellow and less complex the shot becomes. Ideally, I would like to saturate the puck more quickly. I see a flow control mod for the Synchronica will be available shortly, so that is encouraging.

Now that I know my extractions are OK, I will muck about to try to improve things.