I've been spending the last few days measuring the solids extraction (that is, what percentage of the puck ends up in the cup) of the three grinders. Yesterday I did a set with Ambrosia, today with Red Line.

The results are interesting, to say the least, since I didn't get what I expected. When I did these experiments on the Versalab M3, lower doses in the same basket led to higher rates of solids extraction. In the LM triple basket, a 12 gram dose extracted at around 24%, while a 19.5 gram dose extracted at around 16%. Andy did a set of observations using the Mazzer Mini, and got similar results.

Here they are. Each grinder has a pair of rows, the first for the Ambrosia, the second for Red Line. There are three data columns for the doses of 12 grams, 15.7 grams, and 19.5 grams. The figure in the chart with the percent sign is the solids extraction, the figue in brackets my overall rating of the shot using the 0 to 6 WBC scale. I prefer the taste of Red Line to Ambrosia, but this is mostly subjective, since I tend to prefer livelier, lightly roasted espresso.
The big news is that these grinders do not under-extract at any dose. There is a slight drop off in the smaller Macap, but not nearly what we saw with the regular sized grinders. In regards to dose. my preferences are consistent between the two blends, I like one dose on each grinder a bit more than the others. However, it doesn't seem to have anything to do with solids extraction.

If this result remains consistent, it could explain the move to large burr grinders, or perhaps large conical ones, since one doesn't have to worry about under-extracting at higher doses. However, I'll need to collect more data before this is reasonably certain.

Hi Jim:

So far, not many comments on your post. Unfortunately, I have no comment either [yet].

I find this topic very interesting, but I'm still working on trying to understand all that went on back in May in the "Paper on Espresso Extraction" thread.... :-0

I would expect a grinder that had a less uniform grind to produce more TDS's. You will have more fines in the basket. Given their huge surface to mass ratio they should give up their solids faster and because there are more of them, the TDS percentage would be higher.

The more uniform grind would produce fewer TDS's because there are fewer fines in the basket. The tighter delta on the partial size would produce lower TDS but yield a more uniform percentage across the dose. Since the particles are more uniform they will give up a relative uniform TDS as long as the dose and extraction percentages are kept even. If you use X gram dose and Y ounces extracted at a constant time you get Z TDS's. If you reduce the dose and extraction volume by the same percentage and retain the same extraction time (say 28 seconds) your total TDS percentage will be uniform, assuming that the rate the particles give up their solids is constant. Now reduce the dose and increase the extraction time and you should get more solids in suspension since they have had longer to extract those solids. Do my ramblings make any sense?

I would also expect the grinder with the tightest delta to produce a faster extraction. The more uniform aggregate would allow water to flow through it faster because there are fewer fines to fill in the gaps around the partials. It would be like pouring water through uniformly sized rock. The water flows through relatively fast, but add a little sand to fill in the spaces around the rock and the flow is retarded.

AndyS wrote:Hi Jim:

So far, no much comment on your post. Unfortunately, I have no comment either [yet].

I find this topic very interesting, but I'm still working on trying to understand all that went on back in May in the "Paper on Espresso Extraction" thread.... :-0

The results have me completely stumped as well. I had a whole test protocol lined up based on the idea that lower doses extract more ... so much for that. Moreover, to my mind, despite the measurements, the taste on all these grinders flattens or mellows out at lower doses.

Your extractions from the Mini at the time we did these trials were similar to mine from the M3 -- higher doses extracted far less. So I have no idea why the identical methodology is getting these results from these grinders. I assume it is accurate. But I'll replicate on the M3 again (assuming it makes it), and repeat on these grinders using home roast (PIDed PIs under-extract?) just to make sure this isn't some bizarre gotcha. If the results holds up, it puts a whole new perspective on the benefits of going to larger burr size grinders. If it turns out that my paper was based on some measurement snafu, I'll look like a right idiot.

Stay tuned for TITAN GRIND OFF 2007.

Tomorrow, Marc Klein, accredited barista championship judge, joins me for a blind tasting session. Two identically PIDed and pressure regulated Silvias are doing the shot making honors. Each grinder will get tasted side by side with the venerable Mazzer Mini, sporting fresh burrs for the occasion. I'll be serving Marc Coffee Klatch's USBC blend, the one Heather Perry used to win the USBC, a lighter roasted blend. Then Marc will serve me Coffee Klatch's Bella Espresso, their regular blend, a deeper roast in the same class as Black Cat, Toscano, etc. We won't know which of the side by side cups is the Mini's, or which grinder is being served when. We'll be using the WBC espresso score sheet: 6 possible points for crema appearance, 6 possible points for crema persistence, 24 possible points (6 points times 4) for flavor, and 24 possible points (6 points times four) for body; for a total of 60 possible points.

The variations in the Mini's scores will give us some idea of our shot pulling margin of error. So there won't only be a winning score; but we'll also know if the victory is clear or disputed.

Here's the latest Vegas line:

-- Robur 3:1
-- Macap MXKR (the big one), 8:1
-- Macap MXK (the little one), 9:1
-- and a late surprise entry from Spain, my very own Compak WBC, 13:2
-- the heavy underdog Mini, 100:1
-- or the heavy favorite, "Within the Margin of Error," 3:2

cannonfodder wrote:I would expect a grinder that had a less uniform grind to produce more TDS's. You will have more fines in the basket. Given their huge surface to mass ratio they should give up their solids faster and because there are more of them, the TDS percentage would be higher.

Or more likely the opposite. The grinder that produced a less uniform grind (and therefore, more fines) forces you to coarsen the grind in order to maintain reasonable shot times. Consequently, you end up with fewer TES (Total Extracted Solids).

But this is mere theory. The interesting part is when someone actually takes the time, pulls the shots and directly measures the TES . As Jim is doing....

I'm using the silly term TES because in espresso, the high energy extraction strips out a significant quantity of solids that are not soluble. TES is the sum of Total Dissolved Solids and Total Dispersed Solids. :-0

another_jim wrote:Your extractions from the Mini at the time we did these trials were similar to mine from the M3 -- higher doses extracted far less. So I have no idea why the identical methodology is getting these results from these grinders. I assume it is accurate. But I'll replicate on the M3 again (assuming it makes it)

I will try and do a decent number of trials comparing the M3 to the three-phase Robur (83 mm burrs).

another_jim wrote:If it turns out that my paper was based on some measurement snafu, I'll look like a right idiot.

HIGHLY unlikely. But as always, your openness is as commendable as it is unusual.

This is getting very interesting again. Sorry I don't have time for more comment at the moment.

Just anecdotally, are you also finding the higher dosed shots from Ueber-Robur mellow?

The grinders were waiting ...



Left to right: Robur, Small Macap, Large Macap, Compak


The arena was set ...




Marc arrived and we were off.


The first round, in which I served Marc paired rounds from the Titans and the Mini, of Heather's winning USBC blend, went as expected. The Mini was a distant last, and the Titans were all bunched together. In the second round, in which Marc served me paired rounds of Coffee Klatsh's Bella espresso blend, there was a surprise -- the god shot came from the Mini, and brought it up to second place. The Robur was first by a mile. In the third round, we switched from the paired Silvia's to the Elektra and served only one drink at a time, one from each grinder. Marc served me an impromptu blend I roasted, in which the lightly roasted aged Sumatra had not yet settled. Basically, all the shots except the Robur's sucked.

The results:
Margin of error (standard error of Mini's scores): 2.4

The scoring was WBC standard, 6 for crema persistence, 6 for crema appearance, 24 for flavor and 24 for body. The margin of error is computed from the standard errors of the mini's scores in the first two rounds.

The conclusion: the Robur wins clearly and decisively, the rest of the grinders are in a pack with no clear order. Surprisingly the Mini is not a clear loser, but in the pack with the rest. It was the clear loser in a light roasted blend that was perfectly ready to go. But for the medium roast, it fared better than the smaller conicals (paralleling the observations on the Major versus the Kony), while for the wonky home roast, it sucked along all the rest except the Robur.

A note: the Compak was delivered with burrs slightly out of true, so it may perform better when up to snuff, although it doesn't say much for either Compak's QC or UPS shipping. {EDIT: I'm not certain what's going on, but the mounting screw on all the conical burrs in the test show some wobble, more than is possible if the burr is wobbling the same amount -- It could be an artifact from the way the screws thread, rather than a defect}

another_jim wrote:The conclusion: the Robur wins clearly and decisively, the rest of the grinders are in a pack with no clear order.

Great work, Jim. I'm glad I wasn't the only one to be impressed with the Robur.

RapidCoffee wrote:The next day I took the Mazzer Robur out for a test drive. Allow me to preface this by stating that I did not want to like this grinder. It's far too big and pricey for home use.
... <snip snip snip> ...
Sigh. I like this grinder.

RapidCoffee wrote:Titan Grinder Impressions
...
The Robur produces a balanced taste profile, characterized by richness and body, and a flavor profile somewhere between the transparency of the MXK and the muted blending of flavors in the Super Jolly.

I've been greatly impressed by the Robur from the very beginning, and it rapidly became my grinder of choice. Do I want one? Hell yes! Do I want one in my kitchen? Hell no! It's far too big, heavy, and costly to be a strong contender in the home barista market. That's too bad, for this is truly a wonderful espresso grinder.