ExtractMoJo for exploring the extraction space

Want to talk espresso but not sure which forum? If so, this is the right one.
User avatar
AndyS

#1: Post by AndyS »

another_jim wrote:This is less pro versus am issue, and more science versus commerce one. The refractometer is a very normal case of the strengths and weaknesses of commercial R&D. Andy did the basic research on using refractometers for coffee TDS about four to five years ago.


Hi Jim:

I appreciate that you're giving me props for experimenting with the refractometer back when you and I were working on our espresso extraction project. But I think it's a gross exaggeration (and profoundly unfair to Vince Fedele) when you say, "Andy did the basic research on using refractometers for coffee TDS."

Among many other things, Vince did the truly massive amount of testing to correlate refractive index and temperature to coffee strength....he correctly identified that total dissolved solids, not total brew solids, was the correct variable to focus on....he tediously negotiated with several manufacturers to develop instruments that offered a reasonable compromise between accuracy and cost...he developed the software that crunched numbers, displayed data in an easily-understood graphical format, and made it easy to share with others...he developed instruments and techniques to measure both brewed coffee and espresso...etc, etc. And this is not to mention the fact that Vince had already been working on refractometers for coffee TDS while you and I were thinking about espresso extraction.

another_jim wrote:If you want all your brewed coffee to be precisely at the golden cup standard, this refractometer will do the job much more easily than the previous methods.
If you want your coffee to be at the official "golden cup" standard or your own personal extraction standard, the refractometer will allow you to do it with consistency.
another_jim wrote:But it fails in fully exploring or controlling the extraction space. It was impossible to predict or predetermine the TDS, total solids and extraction yield from a given grind setting, coffee/water ratio, brew temperature and steep time. The conversions required for such a determination were not stable enough to be useful. So the refractometer needs extra data tables to fulfill its promise as a complete extraction controller.
I strongly disagree with your statement here because:
(1) IMO the refractometer is a great success in allowing one to "fully explore" the extraction space. Using it, a skilled barista (home or pro) can navigate the extraction space with unprecedented ease, and with confidence that s/he knows where s/he is after each measurement.
(2) Accurately predicting or predetermining the TDS and extraction yield WILL ALWAYS be near-impossible, no matter how sophisticated the system. There are too many uncontrolled variables above and beyond the ones you list: coffee type, coffee age, roast variation, burr condition, grinder temperature variation, brewing agitation, water quality, and many others. I am amazed that you would criticize the VST refractometer and software for not living up to a standard such as the one you propose here. As I said previously, IMO this isn't a fair standard, it's more along the lines of a fantasy.

[disclosure: I have done and continue to do beta testing of VST refractometer/software systems. I have received free instruments, software, and coffee in order to do this testing. I have no financial interest in the company or its products]

[additional disclosure: text edited 1x to remove a phrase which made me sound "ever more like a shill" according to a very prominent HB person. I didn't think I was a shill, but perhaps I'm not a very good judge of whether I am or not.]


[additional additional disclosure: text edited 1x more to remove gratuitous sarcasm]


...split from Is there a new disconnect between coffee hobbyists and professionals? by moderator at Andy's request...
-AndyS
VST refractometer/filter basket beta tester, no financial interest in the company

User avatar
another_jim
Team HB

#2: Post by another_jim »

If everything that can be done is sufficient; and everything that cannot be done is a fantasy; why talk? Just drink the coffee.

Specifically:
  • The calibration standard for coffee extraction is dry weight. One evaporates the liquids and weighs the solids. The dry weight of solids in the cup is the TDS calibration standard (a misnomer, granted, it should be "TDSS"). The difference of dry weight of the fresh and brewed powder is the extraction percentage.
  • To say 1) that Vince "correctly identified" that this lab standard isn't as good a guide to taste as total solids; and to simultaneously say 2) that predictively controlling extraction is a fantasy; is to be caught in a self contradiction. To ascertain the one, you must be able to do the other.
  • But there is no calibration definition of dissolved solids other than the refractometer readings themselves. I saw no methods for centrifuging or bleaching the suspended compounds out to differentiate the two measures or to convert between them
  • I say no large body of tasting data regressed against both dried solids and refractometer readings, so that one could positively state that the refractometer is a better taste correlate than dry weight.
  • And as you say, controlling the extraction enough to produce a differentiating experiment is not possible
My attempt to do some of these things failed. This may be due to not trying hard enough or not knowing enough. I have had much less trouble getting predictable and controllable extractions in terms of dry weight. However, taking dry weight measurements is tedious and smelly. Moreover, my correlations of taste to dry weight extraction variables are no more convincing or repeatable now than they were four years ago. I was looking forward to getting an improvement in this research using the refractometer. I did not; so perhaps my disappointment shows through.

I recall you once also thought that proving basic claims about coffee taste was not a fantasy. I'm sorry that you have lost that belief; but perhaps it is the sign of a realism I stubbornly refuse to share.
Jim Schulman

Urnex: 100% dedicated focus on coffee and tea cleaning
Sponsored by Urnex
User avatar
AndyS (original poster)

#3: Post by AndyS (original poster) »

another_jim wrote: 1. The calibration standard for coffee extraction is dry weight. One evaporates the liquids and weighs the solids. The dry weight of solids in the cup is the TDS calibration standard (a misnomer, granted, it should be "TDSS"). The difference of dry weight of the fresh and brewed powder is the extraction percentage.
2. To say 1) that Vince "correctly identified" that this lab standard isn't as good a guide to taste as total solids; and to simultaneously say 2) that predictively controlling extraction is a fantasy; is to be caught in a self contradiction. To ascertain the one, you must be able to do the other.
3. But there is no calibration definition of dissolved solids other than the refractometer readings themselves. I saw no methods for centrifuging or bleaching the suspended compounds out to differentiate the two measures or to convert between them
4. I say no large body of tasting data regressed against both dried solids and refractometer readings, so that one could positively state that the refractometer is a better taste correlate than dry weight.
5. And as you say, controlling the extraction enough to produce a differentiating experiment is not possible
1. I'm under the impression that all the classic work done on extraction percentages was on paper filtered coffee. So it's really an open question what the "standard" should be for metal-filtered or unfiltered coffees. BTW, your last sentence ("the difference of dry weight...") is incorrect.
2. No idea what you're talking about.
3. Vince says he repeatedly filtered out suspended solids, then dried and weighed them. They were typically well under 10% of total brew solids. Lets assume that he really did this and he isn't lying. Now, a little example: let's assume a 20g dose, 40g beverage, TDS (Total Dissolved Solids) measures 20%, TBS (Total Brew Solids) measures 22%. That means dissolved solids actually weigh 4g and total brew solids weight 4.4g, so the insoluble brew solids weight 0.4g. Of that 0.4g, the theoretical maximum soluble components are 30% (70% of coffee is insoluble). 30% of 0.4g = 0.12g. So there could be a maximum of 0.12g of potential solubles lurking in the insolubles. 0.12g/20g = 0.6% possible error. I don't think I can reliably taste a 0.6% change in extraction. Can you? In fact, the error is undoubtedly a lot less because most of the solubles in that 0.4g of suspended solids have already gone into solution and are measured by the refract. Now this little example isn't real data, just approximate numbers, but it appears likely that (unless Vince is a liar) the difference between TDS and TBS is simply not that important.
4. Do it! Knock yourself out!
5. Agreed.
another_jim wrote:I recall you once also thought that proving basic claims about coffee taste was not a fantasy. I'm sorry that you have lost that belief; but perhaps it is the sign of a realism I stubbornly refuse to share.
I don't get your point. Instead, I continue to believe that the use of the refractometer to measure extraction yields for both brewed coffee and espresso will improve our coffee and help advance the state of the art. And I believe that DESPITE THE FACT that I don't make money on it.
another_jim wrote:why talk? Just drink the coffee.
I'm just about at that point.
-AndyS
VST refractometer/filter basket beta tester, no financial interest in the company

User avatar
another_jim
Team HB

#4: Post by another_jim »

There are two possibilities here:

1. If there is only a ten percent difference between total and dissolved solids, then over the entire range of coffee to water ratios from weak brews to ristretto espresso, they are too correlated to make a difference as taste predictors. I did no tests on coffee concentrations so I cannot really comment on this.
2. I am interested in extraction. If that ten percent difference also exists in the 18% to 22% solids extraction range, which it does I think, and this range is what counts for judgments of over and under extraction, then the refractometer readings for extraction might only be weakly correlated to evaporated weight. This is what I found. I've found passable but not great correlations of taste features with extraction measured by dry weight. I abandoned trying to correlate the refractormeter levels of extraction to taste features after a week of trying, and getting nothing remotely useful.

As I said, this is an excellent tool for doing QC on brewing, something I believe is hugely important, as my other posts on this thread underline. But to claim that it is a major step forward in solving the problems of correlating extraction to taste is in my opinion premature.

But I'm listening ... The experiments from the 1950s showed that roughly 20% solids extraction is the desired zone for coffee. My much more humble experiments indicated, but didn't conclusively prove, that overextracted coffee tends to first get a little sweeter, then duller, and finally slightly cooked. Underextracted cofffee can first get slightly clearer, then become either or both more cutting bitter and sour. This means, if the results hold up, that small variations in extraction are appropriate for different coffees ... I'd be delighted to hear that studies using the refractometer have added to this store of knowledge.
Jim Schulman

Ken Fox

#5: Post by Ken Fox »

I tend to approach "matters of taste" in a fairly simple way since I find that the more specific and precise I try to be, the more I go off in the weeds. Just my take.

I don't have a refractometer, and probably will not buy one. I do like simple things like $8 Chinese 0.1g scales.

I have found the exercise of weighing coffee destined to dose a PF to be worth doing, virtually every time. This is the single most important thing that a home barista can do, if he or she wants to achieve consistency. Similarly, I have found that weighing shots as they are produced, on the same $8, 0.1g scales, to be a useful exercise worth doing at least once per day, often more. I am convinced that using my usual parameters, of shots that pull about 1.5 ounce in ~25 to 35 seconds, using 14-16g of coffee for a double shot, that I can tell the difference between a shot that weighs 12g and one that weighs 18g.

And I prefer the 18g shot, while as recently as a few months ago I thought that my shots were better when they weighed 12g. For that I thank Andy S., who opened my eyes to that idea.

But the rest of this is just beyond my pay grade. Sorry.

ken
What, me worry?

Alfred E. Neuman, 1955

User avatar
AndyS (original poster)

#6: Post by AndyS (original poster) »

another_jim wrote:There are two possibilities here:

1. If there is only a ten percent difference between total and dissolved solids, then over the entire range of coffee to water ratios from weak brews to ristretto espresso, they are too correlated to make a difference as taste predictors.
2. I am interested in extraction. If that ten percent difference also exists in the 18% to 22% solids extraction range, which it does I think, and this range is what counts for judgments of over and under extraction, then the refractometer readings for extraction might only be weakly correlated to evaporated weight. This is what I found. I've found passable but not great correlations of taste features with extraction measured by dry weight.
I'm not sure what your expectations are. I can't make fine distinctions in extraction yield. I would say that in general, espresso shots that measure in the 19-20% range of SOLUBLE extraction yield taste sweeter to me than 18% shots. I prefer them. Because I tend to use lighter-roasted coffees, it's difficult to get shots around 22%+. But when I try them they are watery and sometimes a bitterness starts to show through.

BTW, if you read Vince's article in Barista Magazine (link in another thread) he says that the undissolved portion of total brew solids typically measures 4-6% when the basket it working properly. If you have a 0.01g scale it wouldn't be that hard to get your own measurement of this.
another_jim wrote:I abandoned trying to correlate the refractormeter levels of extraction to taste features after a week of trying, and getting nothing remotely useful.
I'm surprised to hear that since you have previously reported correlations of taste to TBS (total brew solids, dissolved and undissolved) as measured by drying pucks. And aside from the small offset, I think the two measurements are comparable.
another_jim wrote:As I said, this is an excellent tool for doing QC on brewing, something I believe is hugely important, as my other posts on this thread underline. But to claim that it is a major step forward in solving the problems of correlating extraction to taste is in my opinion premature.
It is still fairly early in the game. But I'd say that of serious people who have worked with the technology, your opinion is in the minority.
another_jim wrote:But I'm listening ... The experiments from the 1950s showed that roughly 20% solids extraction is the desired zone for coffee. My much more humble experiments indicated, but didn't conclusively prove, that overextracted coffee tends to first get a little sweeter, then duller, and finally slightly cooked. Underextracted cofffee can first get slightly clearer, then become either or both more cutting bitter and sour. This means, if the results hold up, that small variations in extraction are appropriate for different coffees ... I'd be delighted to hear that studies using the refractometer have added to this store of knowledge.
The experiments from the 50's showed that roughly 20% DISSOLVED solids extraction was the center of their desired zone. Unfortunately (according to Vince), the poor suckers were measuring their water input by volume and forgot to compensate for the change of water density with temperature. If you correct that error, then apparently the center of the preferred range ends up at 19%.

This is yet another example of how foolish it is to specify coffee measurements by volume rather than by mass. As you know, I've been campaigning for years for people to report their shot volume in grams rather than some vague volumetric measure. Numerous people have switched over, much to the benefit of our mutual coffee communication. Other people -- you, for example -- refuse to do so and continue to use ounces. Whether they mean fluid ounces or ounces of weight, one never really knows. It's kind of like a religious thing, I guess, once people latch onto something, they just don't want to give it up.

I pretty much agree with your taste observations, except I probably prefer espresso in the "little bit sweeter" range that to you is slightly overextracted. No harm in that, just personal preferences.
-AndyS
VST refractometer/filter basket beta tester, no financial interest in the company

CoffeeOwl

#7: Post by CoffeeOwl »

Since you started talking refractometer and so on, I'd love to see a study on what substances extract from coffee depending on brewing method. I do perceive different effects of drinking coffee brewed different ways and also precise information (and health benefits) is known for tea. Why leave coffee behind?
'a a ha sha sa ma!


LMWDP #199

Espresso Vision: the perfect cup of coffee starts with understanding your roast
Sponsored by Espresso Vision
Nick

#8: Post by Nick »

AndyS wrote:This is yet another example of how foolish it is to specify coffee measurements by volume rather than by mass.
A little semi-side-topic: I think we'll agree that measuring by mass is better than by volume. However, it's only a little better.

I've come full-circle on all this volume-hating. To be sure, volume is problematic for espresso, which varies significantly in density, and is therefore a lot less useful or helpful for determining quantity of extracted product.

But mass is only slightly better than volume when brewing filter coffee, even less so for manual brewing. With a brew that extends over a period of 3 minutes and all that happens during that time, the precision that a gram scale affords is mitigated by the other factors at work. Basically, when you factor for significant digits, volume turns out to be indeed good enough.

I read once on a sportscar forum, something like "If you can't drive fast with 100 hp, 300 hp won't help you." Similarly, if you can't brew well measuring brewed beverage by volume, scales won't help you. 8)
Nick
wreckingballcoffee.com
nickcho.com

User avatar
AndyS (original poster)

#9: Post by AndyS (original poster) »

Nick wrote:mass is only slightly better than volume when brewing filter coffee, even less so for manual brewing. With a brew that extends over a period of 3 minutes and all that happens during that time, the precision that a gram scale affords is mitigated by the other factors at work. Basically, when you factor for significant digits, volume turns out to be indeed good enough.
If you're doing a pourover, and it's foamy and sloppy and half the liquid is up top and half the liquid has already come through, how do you use volume to know how much more water to add?
-AndyS
VST refractometer/filter basket beta tester, no financial interest in the company

Nick

#10: Post by Nick »

Easy: You don't look there. You look at the liquid volume below.

Like this:
Nick
wreckingballcoffee.com
nickcho.com