TGP II: Particle Distribution Analysis of Grinder Adjustments - Interim Results - Page 3

Grinders are one of the keys to exceptional espresso. Discuss them here.
User avatar
peacecup
Posts: 3649
Joined: 19 years ago

#21: Post by peacecup »

What Ken meant to say was that the variation inherent in pulling (temp, tamp, bean freshness, dose, pressure, etc. etc. ...) and subsequently tasting (I won't even begin to list things that influence the subjective quality of taste) even on the best of days is so great, that it will inevitably mask the effects of the minutia of differences in grind particle distributions.

Or, in plain English, the effects of the differences between decent grinders amount to far less than do all the other factors that contribute to what an espresso tastes like.

There are a lot of reasons why someone would chose one top-end grinder from another. Discernible differences in the taste of the espresso they are capable of producing is probably a reason reserved for the very few people with many many years of tasting experience.

PC
LMWDP #049
Hand-ground, hand-pulled: "hands down.."

User avatar
TrlstanC
Posts: 505
Joined: 16 years ago

#22: Post by TrlstanC »

Ken Fox wrote:Do you, or does anyone else here know whether the above "explanation" for the lack of results is due to this limitation of the equipment used and how it would relate to any "results" from this study?
Well, just looking at the 4 data points representing the tests and retests of identical samples, I can't think of any other conclusion except that this reveals a certain amount of measurement error. Unfortunately we know that the machine assumes that the particles are spherical, and we know coffee grinds aren't, we were hoping that this wouldn't lead to too much measurement error, but it appears that there's enough margin of error to make exacting comparisons difficult, at least in the range that we're interested in.

If you or anyone else have any other ideas, I'd love to hear them, that would certainly be a useful contribution to the discussion. And I'm not being sarcastic here, there's a great resource here in Kendall's interest to run samples and having access to time on such sophisticated equipment, if anyone has any ideas that could explain the variance in those 4 data points (besides the obvious), it would be great to hear it.

Advertisement
User avatar
cafeIKE
Posts: 4703
Joined: 18 years ago

#23: Post by cafeIKE »

If I wanted to measure
peacecup wrote:the minutia of differences in grind particle distributions
, I'd probably have suggested doing all the samples at one location on one coffee on shots pulled on one machine by one fist. Oh yeah, I did. That's why I didn't participate.

User avatar
peacecup
Posts: 3649
Joined: 19 years ago

#24: Post by peacecup »

In fact, the relatively low variation in results suggest that they are robust. If the effect of particle shape was important, there would probably be a lot more variation in the figures that what we see.

So, there are some very useful results here: 1. All the grinders are practically the same, and 2) changing the grind setting does not change the distribution of fines. BUT, we know that changing the grind setting does change the flow rate. SO, how about an alternative hypothesis:

When you add 3g to a 15-g dose (i.e. increase it by 20%), using the same basket, you increase the depth of the puck. This changes the rate of flow, due to hydrodynamic resistance (or whatever you call it). This 20% difference in dose has far more influence on flow than does the 1% difference between grinders, or between doses on a grinder.

PC
LMWDP #049
Hand-ground, hand-pulled: "hands down.."

User avatar
TrlstanC
Posts: 505
Joined: 16 years ago

#25: Post by TrlstanC »

cafeIKE wrote:If I wanted to measure , I'd probably have suggested doing all the samples at one location on one coffee on shots pulled on one machine by one fist. Oh yeah, I did. That's why I didn't participate.
Ike, here's your chance to add something to the discussion: Lampert did two sets of samples, at one location, on one coffee, pulled on one machine (although there's a chance he used two hands instead of one fist). Those samples were then tested, and retested (again, by one person at one location, on one machine) and came up with 4 distributions that were fairly widely spread apart. Since this is exactly what you wanted in a test, would you like to participate now, and toss out any useful ideas?

User avatar
tekomino
Posts: 1105
Joined: 14 years ago

#26: Post by tekomino »

The conclusion I draw from this, given very close results for all samples, is that as far as laser is concerned, i.e. the things it can measure the samples are very close. Which means that difference is outside of the measurement that can be performed with this device.

I would though run one more experiment where I would keep the dose the same, say 18 grams, but vary run-time and output (sample #1 18g output in 40 second, sample #2 36g output in 25 second) which in essence just plays with the grind fineness. Probably Kendall could do this alone.

What would this prove? Well if you get same result as in initial study it would prove that difference between grind fineness cannot be measured by laser device.

User avatar
another_jim (original poster)
Team HB
Posts: 13871
Joined: 19 years ago

#27: Post by another_jim (original poster) »

Ken Fox wrote:Given the circumstances, the outcome could have been easily predicted before the "study" was done.
Captain hindsight strikes again.

Your insulting prediction was that we were too scattered, stupid, and disorganized to produce systematic results, and that the samples would be all over the map. This turned out to be wrong.

The outcome was that pairs of grinds that were set at different grind fineness, and that required different doses to flow correctly, i.e. that were clearly different in terms of espresso making terms, showed up as identical in 15 out of 20 pairs on the laser sizer, since the laser sizer isn't as accurate as purported to be in the published articles.

But you were right that the study, so far, has yielded no insight. That this makes you happy reflects on your character, not our wisdom.
Jim Schulman

Advertisement
Ken Fox
Posts: 2447
Joined: 18 years ago

#28: Post by Ken Fox »

Here is an alternate theory for why a Titan level grinder produces better results than a small planar (or other lesser) grinder:

There is more than one important variable at play, and "well-behaved grinds" are well-behaved for different reasons, e.g. due to more than one measurable characteristic such as size.

What if, for example, it was some nuances in the shape of the grinds, combined with "something else?" Maybe there is more than one "something else" factor that works with shape. If shape is an important factor, then the way to find this out would be to look at the grind shapes, perhaps with scanning EM (like John did) or some other technology. Some descriptive observations could be made, although they probably would not be conclusive. And as I said above, I personally think it is more likely that we are dealing with a combination of factors rather than just one.

I am certain that "better" grinders produce better grinds. I have experienced that for myself in my own house and I strongly believe it to be true. Of course, I could be wrong, but so many other people have made the same observation that I'm pretty sure we are onto something. I do not think, however, that there is some very simple explanation for this that will be easily provable in the sort of test that us home baristas could pull off.

Another take-home message from this whole topic is that just because a well known person (Dr. Illy in this case) has written something in a book, it is not necessarily true. The internet and this board should be a place for the free exchange of ideas, and hopefully we can all learn something in the process.

Just my opinion.

ken
What, me worry?

Alfred E. Neuman, 1955

Ken Fox
Posts: 2447
Joined: 18 years ago

#29: Post by Ken Fox »

another_jim wrote:Captain hindsight strikes again.

Your insulting prediction was that we were too scattered, stupid, and disorganized to produce systematic results, and that the samples would be all over the map. This turned out to be wrong.

The outcome was that pairs of grinds that were set at different grind fineness, and that required different doses to flow correctly, i.e. that were clearly different in terms of espresso making terms, showed up as identical in 15 out of 20 pairs on the laser sizer, since the laser sizer isn't as accurate as purported to be in the published articles.

But you were right that the study, so far, has yielded no insight. That this makes you happy reflects on your character, not our wisdom.
I am not happy whenever potentially useful resources are squandered when they could have been used more productively. That having been said, if this forum is really here for the free exchange of ideas, then the goal should be to further our collective appreciation and understanding of coffee. To the extent that individuals allow themselves to be distracted with comments directed at individuals rather than the topics being discussed, this forum suffers and nothing is gained.

ken
What, me worry?

Alfred E. Neuman, 1955

User avatar
cafeIKE
Posts: 4703
Joined: 18 years ago

#30: Post by cafeIKE »

TrlstanC wrote:Ike, here's your chance to add something to the discussion: Lampert did two sets of samples, at one location, on one coffee, pulled on one machine ... and toss out any useful ideas?
The Robur-E retains so much coffee and has such a wide sweet spot that I remain sceptical. It could take several pounds of coffee, particularly with a hopper and varying bean load, to obtain data.

IIRC, the MC4 has a 1mm [1000µm] burr carrier pitch. Normal grind range is about 1/5 rotation from zero, about 200µm. Depending on the coffee and the day, the change from 15g to 18 is about 3 to 5 teeth or 3-5% of a rotation or ~630 to 10µm. Since the gap is several times the fines size, little change on the fines is expected. Similarly, I'd not expect the 630 to 1050µm change have a significant effect on 500µm particles.

EDIT : calc error - 3 to 5% is on 1 rotation, 1000µm not 1/5 :oops: :oops: :oops: