Thanks Kran for the suggestions. I'm not sure I want glycerol in my grinder. I'm having a hard enough time accepting water (basic RDT). Speaking of drop sizes, I tested a spritzer bottle and the amount per spritz was between 0.10 to 0.15 gram. In comparison, I measured drops of water from a pipette to be around 0.03 gram. I think the latter would make a lot more sense when grinding 2 gram samples, and even that might be too much.

I also bought a Kruve Brewler and find it useful to validate the results from the software. Here's an example:



In this sample the particles in the marked area measured at 890 micron on average, with a range of 650-1130 micron (based on scatter of 0.24mm). I think it looks close enough to what one could gather from the Brewler (possibly the software underestimates the size a bit). Confirmation bias on my part?

The glycerol is for after the beans have been ground. It disperses the grinds and also helps with the imaging.

boren wrote:In this sample the particles in the marked area measured at 890 micron on average, with a range of 650-1130 micron (based on scatter of 0.24mm). I think it looks close enough to what one could gather from the Brewler (possibly the software underestimates the size a bit). Confirmation bias on my part?

I don't doubt that the software, coupled with a properly set up camera, can measure sizes accurately. That's the easy part. The question is whether you can get a large enough representative sample to average out orientation effects and that also doesn't have many small particles that are stuck together.

With respect to using RDT, is it actually working to separate the tiny particles? It's not possible to tell from the image you just posted. What if the coffee particles stick together *because* of moisture? Or what if lipids or other forces are responsible?

Kran wrote:The glycerol is for after the beans have been ground. It disperses the grinds and also helps with the imaging.

Are you suggesting that I spread the glycerol on the measuring surface (to which I sprinkle the ground particles)? If so, wouldn't this make them stick and prevent distributing them apart by bumping the surface?

As for imaging, I think the amount of detail is fine as it is. See the full size image here (21MB).

jpender wrote:I don't doubt that the software, coupled with a properly set up camera, can measure sizes accurately. That's the easy part. The question is whether you can get a large enough representative sample to average out orientation effects and that also doesn't have many small particles that are stuck together.

That's a good point. Maybe I should measure different parts of the image to ensure consistency. If I get significantly different results for different sections it's probably a red flag.

With respect to using RDT, is it actually working to separate the tiny particles? It's not possible to tell from the image you just posted. What if the coffee particles stick together *because* of moisture? Or what if lipids or other forces are responsible?

In the above image I didn't actually use RDT. I should experiment with and without RDT to see which flow provides more useful samples. Without RDT I'll obviously get more ground coffee stuck in the grinder, but that's anyway coffee that gets cleaned and doesn't end up in the brew.

boren wrote:Are you suggesting that I spread the glycerol on the measuring surface (to which I sprinkle the ground particles)? If so, wouldn't this make them stick and prevent distributing them apart by bumping the surface?

As for imaging, I think the amount of detail is fine as it is. See the full size image here (21MB).

Yes. What I did was to combine glycerol with grinds, mix, add to a slide, and then put a coverslip (thin glass) on top. This spreads everything out and maintains a consistent layer of particles to image. Not sure if your setup would allow for this but on mine it works.

Keep in mind I'm looking at espresso grinds. In the pictures I posted a few posts back the particles are in the 30-50micron range. It looks like you're more in the 600-1000micron range. So what i need to do to disperse clumps and the level of detail i need is different.

In terms of quality of analysis, how accurate is the thresholding. Does the red overlay match the outline of the particle appropriately? I can't tell from the image you posted with the sieve.