Nick wrote:I asked Don about using the deffractometer to analyze grinds from different espresso grinders, and he immediately replied that the wavelength of the laser was not going to work very well for that grinds of that (small) size.

The laser diffractometer I used several years ago during the TGP was capable of accurately sizing particles between 10um and 1000um, covering the appropriate size range for espresso grinds. This was verified with standards comprised of small glass spheres.

The basic problem is not size per se, but shape and material composition. If coffee grinds would diffract light in the same way as glass spheres, laser diffractometers would give reasonably accurate size distributions. It may yet be possible to tweak some of the parameters to improve upon the results. If not, we may have to resort to sieving the grinds.

Sieves are worse. The laser results were consistent at least, they just didn't have enough resolution. The sieves I've tried never gave me the same thing twice in a row.

Wouldn't that be expected?
As the coffee is manipulated, wouldn't it be 'grinding'?
Think of a particle like an old anti-ship mine. As it bumps and grinds about in the sieve, the spikes become fines.

That too.

If you ever screen larger stuff (like coffee beans), you also see that beans that pass through a screen tip first, don't do so when on their side. When screening coffee, you can check the beans that haven't passed the screen. But that's not going to work with ground coffee. You can get a fairly reliable sieving of the fines versus coarse; but figuring out how the coarse sizes distribute -- not a chance.

John is right, this is not so much a problem with the actual measurement technique, but with using the right statistics and repeated measures to handle very irregularly sized particles.

Kendall is going to use a far more precise technique, but the lab with the gear has been commandeered for a higher priority (aka a more senior professor's) project.

Hi everyone, the laser scattering machine I use uses a blue and red laser to have a broad range of size (10nm-3mm) but all that will never let you overcome the fact that the data has noise superimposed by the shape of the particles. The uncertainty is related to the aspect ratio of the particles. It can "see" the same particle a number of ways (edge on, face on, in between) leading to fuzziness. Even an optical analyzer will do the same thing. The difference in grind for dose size is less than the variance of the noise; it seems we can never resolve the true difference using any method that looks at projections of irregular shapes (microscopy, optical analyzer, laser scattering, even sieves of arbitrarily fine size). I will get to the BET as soon as I can which is a direct computation of surface area, not related to edge on projection.