AssafL wrote:12.5gr coffee to 31.6 grams beverage at 8.1% measured TDS comes to an EY of 21.22% (VST CoffeeTools). How did you get to 23.8%?

<image>.

NB - What this means is that, for a lesser grinder, a larger doses may "over extract" rather than "under extract" - even though the Refractometer will say otherwise - since the part of the dose that actually participated in the pull (the "not-boulders") was lower than the measured dose. The hypothesis also renders a VST plot without sifting almost useless for anything but a perfect grinder.

Edit: BTW - Testing the hypothesis that boulders do not participate in the pull should be fairly easy. Grind two doses and sieve the boulders from one. Run the same beverage weight from both. If the TDS is identical (or nearly so) means the boulders did not participate in the percolation. (unfortunately I do not own a sieve so can't run this test myself).

Sorry it caused such a confusion! Nothing wrong with your calculation - you are absolutely spot on. There must have been some error when I keyed in the figures into my phone. I should've double checked.... 21.22%(or 20.48% based on the setting on mine) is not that much higher than what I get normally, considering it was a long shot. I will do more testings this weekend, especially in regards to your boulder vs TDS question - week day mornings have been hectic, and as I was afraid, it doesn't fit well into my routine.

I did not get your comments in NB though. My understanding is the number to be entered should only be the dose you're using. So the discarded portion doesn't play into calculation.

samuellaw178 wrote:I did not get your comments in NB though. My understanding is the number to be entered should only be the dose you're using. So the discarded portion doesn't play into calculation.

Well, what you said about boulders had me thinking. What if the boulders don't participate as much as fines in the pull (as you say - they under-extract). For example, for a 1000u boulder, maybe the inside hardly wets, let alone percolates...

It stands to reason that the dose you should use for the VST calc would be the dose that is homogenously percolated. But as you say, the boulders don't extract as well (or contribute as much solubles) as fines or small particles.

If that is true, maybe the reason the EY calculation turns out low is because the dose that is used for the calculation is higher than the dose available for the pull (i.e. the dose after discarding the boulders)?

Hope this makes sense.

Edit: regarding the "discarded" statement - what I was implying is you discarded (or reground) the boulders. But had you put them back in as "non contributing" or "mildly contributing" boulders, maybe the proper dose for the VST calculation would still be the dose without the boulders; treated as if they were small Styrofoam peanuts, inconsequential to the pull.

AssafL wrote:You could try to sift it well, and reintroduce the boulders with WDT.

If EY goes down - it is indeed the boulders (and that would also explain why alignment is so critical.

If EY stays - it would be normalization.

If halfway - god's sign that we should be drinking tea.

Some more sifted shots today. First one with both <200um & >500um discarded. 13.1g in, 25.4g out in 27s. TDS 10.4. EY was 20.17%.

Second was with only <200um removed, and the sifted boulders thrown back in with a WDT. 15.8g in, 30.8g out in 38s. TDS 9.9. EY was 19.3%.

I need to predict how the grind would affect the flow and compensate accordingly. The second shot had a lower EY probably due to the choked/slow flow. So it wasn't conclusive at all.

In comparison, my 3 unsifted shots (control) were 20.5, 21.3 & 21.4 EY. Note all the TDS were measured unfiltered, so maybe it would appear higher than the filtered TDS results out there. So I am now not convinced that sifting would magically give a higher TDS/EY instantly, at least not for the Helor 101. Though sifting does seem to affect the taste in some way. I got about 4-4.5g (out of 15g) staying above 500um screen, before regrinding. Thing might've been a bit different for large conical burrs with a wider grind spread.

I ran out of the batch again unfortunately (arggh) to be able to do apple-apple comparison. It was a 150g batch done on Quest M3 so I only have around 7 shots/chance before the batch is gone.

From practical point of view, I've timed the sifting process. It took 2 minutes to sift the first 16.5g (until no grind in the top tray decreases further) and another <1 minute to sift the reground boulder. All in all, the process took 5 minutes/shot instead of the usual 1 min.

So this morning I made what i at first thought was a rather stupid mistake, I left my Monolith Flat set for espresso while grinding for my morning Clever brew. I discovered this when after sieving, I found no grounds above the top screen. and a bit over 4 grams out of 32 past the 30 screen in the bottom. I brewed it for only 1 1/2 minute or so and it drew down much faster than expected and the surprise was, it was the best cup of coffee I've had from this bag so far. So my concerns about how fine they suggest the grind for drip is may be unfounded. I did the same later for a friend and he also commented how good it was.

Ira

samuellaw178 wrote:I need to predict how the grind would affect the flow and compensate accordingly. The second shot had a lower EY probably due to the choked/slow flow. So it wasn't conclusive at all.

Hmmm. Very interesting results. I want to use your results as a way to figure out an "effective" dose weight (basically, assuming the coffee extracts evenly, at which point the difference between 1 & 2 is "dead weight", or % of dose that remains non-percolated).

So in keeping all the parameters constant and changing only dose, weight, the calculator shows that the "dead weight" in puck 2 (to get the same EY as in puck 1) is 0.7gr. It is as if your dose in puck 2 was 15.1gr rather than 15.8gr (or 0.7gr did not get percolated).

(BTW - CoffeeTools get me much higher EY then your calculation. I get EY=20.9% for pull 1. Do you use VST CoffeeTools or some other method to calculate?)

But given your unsifted results I am far from convinced as to the validity of this calculation. Even more, if the 2nd pull got choked, I would expect the EY to go up (as you get longer percolation, ending up in over extraction). So perhaps the difference due to removing boulders is even more pronounced (but compensated for by longer percolation).

Yet another example of how Refractometers may be a good QA tool for maintaining consistency between batches, but their ability to calculate a proper EY is based on so many assumptions around percolation - that it is probably always wrong....

AssafL wrote:Hmmm. Very interesting results. I want to use your results as a way to figure out an "effective" dose weight (basically, assuming the coffee extracts evenly, at which point the difference between 1 & 2 is "dead weight", or % of dose that remains non-percolated).

(BTW - CoffeeTools get me much higher EY then your calculation. I get EY=20.9% for pull 1. Do you use VST CoffeeTools or some other method to calculate?)

But given your unsifted results I am far from convinced as to the validity of this calculation. Even more, if the 2nd pull got choked, I would expect the EY to go up (as you get longer percolation, ending up in over extraction). So perhaps the difference due to removing boulders is even more pronounced (but compensated for by longer percolation).

I think there's a threshold for a particle size to be considered dead weight, and that number likely starts around >600 micron for a percolation time of 27s+ (based on what I see so far & some of my likely-inaccurate intuition).

I understand what you meant by longer percolation should've yielded higher extraction. But what actually happen more often is you start getting mini dead areas within the puck (sometimes not visible even with a bottomless pf). This lowers the effective dose, borrowing from your dead dose weight concept. So percolation science isn't really just about particle size & extraction, but also the packings. Maybe boulders do play an important structural role, as the Italians had alluded long ago, in preventing the finer grinds from creating dead area themselves.

I'm using Coffeetools too. My EY numbers are slightly lower because I've set the moisture and CO2 to 0%. My understanding is these settings sometimes varied among different Coffeetools version. I only take the numbers for internal checking so not a problem.

samuellaw178 wrote:I think there's a threshold for a particle size to be considered dead weight, and that number likely starts around >600 micron for a percolation time of 27s+ (based on what I see so far & some of my likely-inaccurate intuition).

Seems logical. But I have no way of verifying. Also, my guess is that 600uM specs do extract, only their centers remain under-extracted and for larger ones - dry. So it would be a percentage of their mass.

So my guess it that per coffee (and how stale it is) and per grinder (and alignment and burr sharpness) - there should be a dose correction factor.

I understand what you meant by longer percolation should've yielded higher extraction. But what actually happen more often is you start getting mini dead areas within the puck (sometimes not visible even with a bottomless pf).

Okay - but since the time of contact with the water increases - that should compensate (I'd think). The puck becomes a soup.

I can imagine a different mechanism at work - if water doesn't flow - perhaps it doesn't emulsify the oils - and that, in turn, doesn't allow some of the more stubborn solubles in. TDS then does down. I guess flow rate may be important as well and the reason stuck pours are horrible (underextracted).

This reminds me of an argument with the refractometer crowd about emulsions and how (they claim) that the refractometer should ignore the emulsion (which I found odd at best). I remember thinking that a corollary is that if you measure mayonnaise the refractometer will only measure the egg component... Very odd.

This lowers the effective dose, borrowing from your dead dose weight concept. So percolation science isn't really just about particle size & extraction, but also the packings. Maybe boulders do play an important structural role, as the Italians had alluded long ago, in preventing the finer grinds from creating dead area themselves.

That latter part is what my interpretation of the Craft and Science of Coffee book seems to suggest. That the role of normalization is not just handling fines - but how the density is controlled in the puck. A good grinding will have the fines either as agglomerates or affixed to larger particles. Doublehelix calls this "fractal" organization of grinds. I likened this to a pile of cocoa covered balls.

I'm using Coffeetools too. My EY numbers are slightly lower because I've set the moisture and CO2 to 0%. My understanding is these settings sometimes varied among different Coffeetools version. I only take the numbers for internal checking so not a problem.

Ah - good to know. As for the reasoning above - I think it the actual numbers matter little...

Still enjoying sifted coffee, I threw some of the mixed boulders and fines in chili last night for flavor. May use some as a dry rub tonight.

What's everyone finding works best for cleaning these sieves? I've been rinsing them under my kitchen spray head and putting them in a dish drying rack, but I've only been using them for making one batch of morning coffee and possibly one in the evening, I can imagine it might be less than ideal for using it multiple times and having to dry them in between.

For espresso I find it clogs quite significantly after 1 sifting season. I then use a microfiber towel for drying portafilter basket to wipe down the screen surface, before finishing with a fine bristle brush. When sifting for coarser grinds, I think it can last way longer without cleaning in between. After I'm done for the day, I too rinse it and dry them separately. Can't imagine washing and drying them in between either.