In a possible diversion from the rudeness and ensuing hurt feelings that so often prevail on H-B, people might want to look at the newly-posted research results from Marco (the wizards of Dublin). They have been looking at a better design for non-espresso grinding, with mono-polar distribution of particle sizes as their goal: http://lamarzoccousa.wordpress.com/2010/10/27/uber-grinder-what-it-is-what-it-isnt/.


...split from TGP II: Particle Distribution Analysis of Grinder Adjustments - Interim Results by moderator...

Marshall wrote:In a possible diversion from the rudeness and ensuing hurt feelings that so often prevail on H-B, people might want to look at the newly-posted research results from Marco (the wizards of Dublin). They have been looking at a better design for non-espresso grinding, with mono-polar distribution of particle sizes as their goal: http://lamarzoccousa.wordpress.com/2010/10/27/uber-grinder-what-it-is-what-it-isnt/.

What is on the vertical axis of the graphs in that article? The label is so fuzzy I cannot read it. (It also looks like it's in German or something, but that I could handle.) They seem to have far fewer fines with their reference grinder as well as the "uber grinder".

PSD graphs generally have the percentage of total particles at a given size (indicated on X axis) on the vertical axis. Its not really important what it is, however. The shape is more important, primarily the relative heights of the peaks in a multimodal distribution. The total area under the curve represents all the particles measured.

In TGP I as I recall every tested grinder had vastly more fine particles than coarse particles, in terms of number. But in terms of weight or total volume, which was also measured there and is what is measured here in this thread, the coarse particles were generally greater. So the shape depends on what is being measured.

And I believe there was never a graph in TGP with the shape of the "uber grinder" graphs, with the fine particle peak so insignificant compared to the coarse particle peak.

Was there any note on the Marco site about how they were measuring the particle distribution? The method could certainly have a big effect.

Another chart I remembered from the original TGP was the solubles yield comparison, where the flat grinders had a steeper slope (high extraction in low doses, to lower extraction in high doses), the big conicals had a mild slope, and the Robur was basically flat (with the Lux also having a mostly flat slope).

Comparing that data to the particle size profiles we have here (even taking in to account the amount of noise present) it looks like the grinders that produce a relatively high volume of coarse to fine particles are the ones with the steeper extraction slopes, while the ones with similar volumes of coarse to fines have the flatter slopes.

I'm sure I could come up with lots of theories to try and explain what could cause the difference in solubles yield, but it's just as likely that there's some other factor that's determining both (like particle shape), and I don't want to confuse correlation with causation.

artichoke wrote:What is on the vertical axis of the graphs in that article?

The y-axis is labeled "Verteilungsdichte (distribution density) q3" and is a percentage distribution, apparently of particle volume. The x-axis is the particle size in microns.

I believe there was never a graph in TGP with the shape of the "uber grinder" graphs, with the fine particle peak so insignificant compared to the coarse particle peak.

The Marco "uber grinder" graphs are for bulk grinders, which produce less fines by design, compared to the TGP espresso grinders. You can see this clearly in another Marco plot that compares a Vario to a pair of notable bulk grinders.

TrlstanC wrote:Was there any note on the Marco site about how they were measuring the particle distribution?

The data was developed on with a Sympatec QICPIC instrument, which is an automated image analysis particle sizer. It is similar to a laser diffraction particle sizer, but purports to be more accurate for non-spherical particles. It photographs the particles and produces particle counts as well as calculated particle volumes based on the particle shape e.g. Feret diameter, aspect ratio. What is on the graph appears to be the calculated particle volumes.
.

The "Uber Grinder" certainly bears a striking resemblance to a Tanzania. It's no wonder it compares so favorably on those graphs.

Über grinder indeed is Tanzania, with different burrs.

How do they differ? The description certainly sounds like a Tanzania/805 burr.