AssafL wrote:Yes. Cellulose doesn't dissolve in water. I use it though as it binds with water forming a colloid. The colloid will obviously allow more particulate stay dispersed in the suspension.

So? Cellulose doesn't get bitter if you pull too much or sour if you pull too little. It isn't a part of the extraction process no matter how much it's a part of the coffee.

AssafL wrote: In online searches n' (refractive index( is related linearly to the ratios of the dispersed to continuous phases. https://www.google.com/amp/s/www.resear ... rsions/amp

That would be one source of error that can be evaluated.

Just remove it and don't have that source of error. Then you can measure dissolved solids with perfect accuracy and know exactly how much of the dissolvable stuff was removed from the coffee.

CwD wrote:It is also accuracy, though. The filters or centrifuging don't just help to make measurements more repeatable, but to make them more accurately reflect the real tds. You can't inaccurately measure something you're not measuring in the first place. Refractometers do not measure suspended solids or lipids in the first place. If you only feed it the thing it's actually supposed to measure, it gives you a very accurate number for what that is. And we can see by the comparison to the centrifuge not only that centrifuging is equally effective, but also that the VST filters do not remove a substantial amount of what is actually supposed to be measured. They work as a check on each other to prove that filtering in no way makes a reading less accurate to reality when you realize that reality applies to what it's actually measuring, real total dissolved solids, and not total "stuff" in drink, which it makes no attempt to measure.

Yes, we seem to be in violent agreement.

In the metrics sense, filtering (or centrifuging - which I can't attest for - as I haven't personally tried it) we are talking accuracy in taking refractive measurements.

Since this is a foods discussion, it all too often gets interpreted that accuracy is somehow accounting for all that is in the cup, and the wheels come off the bus for the rest of discussion, so to speak.

We are measuring a specific set of things in order to calculate a specific term, E.Y. and we're using the value as a process control.

Undissolved materials in the sample only serve to interfere, unpredictably, with measurements of refractive index. Interfere a little, maybe not at all, or a lot, just sort of depending on the luck of the draw.

CwD wrote:Just remove it and don't have that source of error. Then you can measure dissolved solids with perfect accuracy and know exactly how much of the dissolvable stuff was removed from the coffee.

I did not know that a colloid could be centrifuged or filtered into a solution.

Polysaccharides are hard to push through filters (I use a super bag) but what happens? Do they break down into monomers?

I guess one learns something new every day....

nuketopia wrote:... we are talking accuracy in taking refractive measurements.
....
We are measuring a specific set of things in order to calculate a specific term, E.Y. and we're using the value as a process control.
.

Violent agreement. The only question I am still at odds is whether the EY is the same EY that one gets using a dehydration oven.

Excellent tool for process control. Even better would be to measure in nD or RI. Then there is no doubt about proper process. And to each their own (as Atago seem to imply).

Depends.

If you're dehydrating filtered coffee samples for TDS and using the extraction equation, yes.

If you're dehydrating the puck? No. That is not an extraction yield measurement, but a total yield measurement that includes loss of mass through means other than extraction as well as extraction. Oils and sediment are not a part of extraction, and any method including them is not measuring extraction, full stop.

If you were willing to spend a few hours of measurement time per shot, you could get both extraction yield and total yield to see how much soluble stuff came out of your coffee AND how much non soluble stuff came out. But I'm not sure what you'd do with that information, since you can't precisely control the non soluble part anyway. And only the amount in drink matters for it, not the percent of possible to extract like with soluble stuff since there's no gradient process.

CwD wrote:Depends.

If you're dehydrating filtered coffee samples for TDS and using the extraction equation, yes.

If you're dehydrating the puck? No.

Here is where I don't follow the math...

It is my understanding that if I weight the dose and measure the humidity in the coffee and subtract it from the dose (to get dry mass)

That^^^

Should equal:

Dehydrated puck weight + deydrated shot weight.

Even that isn't accurate as some heavy carbon and oxygen atoms escape as CO2. And maybe some volatile flavonoids.

What you are saying is that the two terms do not add up to the dry mass - that one is only extracted weight while for the puck is the TS. Why?

This conversation has been enlightening for me.

I've always thought of extraction as a noun for the process of pulling a shot. "The Extraction has a nice cone formation." That would sum up my personal feeling for the word...

Where I went wrong was in assuming that calculating extraction yield would be calculating the yield of the extraction. I say that plainly, and not in an effort to be contrary. This conversation has helped me understand that the term extraction yield, as applied to coffee is a measure only of dissolved solids. This is a critical sticking point.

I own an optical refractometer. I notice that suspended solids and oil droplets (amazing post showing this, Sam!) lead to a blurry line. My measurement error of Brewster's angle is bounded by the transition area from dark to light within the blurred region. Filtering the sample gives you a sharp transition and lowers the measurement error such that you can measure the dissolved solids of the sample with much less uncertainty. An electronic meter does not have the capability to make a determination between a blurry line and a sharp line, as you simply get a number. Something tells me that the imaging algorithm within the electronic meters is using either a light to dark transition or dark to light to spit out a number. One would think the instrument could make both calculations and then spit out the average of the two, along with the relative uncertainty, but that would just be too convenient...

Thanks to the latest interchange here, regarding dehydrated pucks, (which gives you the total yield of stuff that made it to the cup) and dehydrated filtered samples (which gives you the total yield of dissolved solids that made it to the cup), I am seeing the profound difference between the extraction and the extraction yield.

An even extraction process leads to more complete extraction yield, as calculated by measured TDS. I think it is important to recognize that EY%, while a good indicator of the completeness of the extraction, is not inclusive. I agree that if it isn't dissolved solids, it isn't a part of our current definition of extraction yield. However, I disagree with the notion that if it isn't dissolved solids, it wasn't extracted. Extraction is "the action of removing something, especially by use of effort or force". Thus when we think of an extraction, it necessarily includes all the "stuff" that making espresso strips out of the puck. When we talk about extraction yield, whether or not we are filtering the sample, we are only talking about dissolved solids.

Cheers!

- Jake

I think the fact we use the word extraction for multiple different things really doesn't help. Extraction including oils and sediment is by no means wrong just by the meaning of extraction in a vacuum, but the context is important.

There's extraction from the puck, that I've taken to calling total yield, but there's also extraction from the coffee itself, which is what extraction yield is based on. Moving coffee from the puck to the cup is a dictionary correct extraction, but it's not extraction as it refers to extracting parts of the coffee from the coffee by dissolving.

Okay - I've located my mistake. EY is defined as the dissolved coffee / dose. (Note to self: must remember EY is solubles; like soylent green is people)...

So the refractometer (with various corrections for colloid... etc) should bring us close to TDS.

So what number do drying ovens give us? Even the VST manual states drying ovens are the most accurate at giving us TDS and Nestle (according to the Brita Folmer book) use drying ovens.... But they will also measure the colloid and the lipid (as well as the pesky boulders). How does the "accurate" drying oven account for these when calculating TDS?