I've done some initial testing with Jonathan Gagne's Grind Size application and it seems to do the work. I now plan to test my grinders with it and map their consistency and average grind size at different settings. This way (I hope) I'll be able to find which grinder produces less fines at different grind levels and use this info when deciding which grinder to use for different brews (e.g. which is the most consistent for pour-over, cold brew). It should also help when switching between different grinders aiming to get similar extraction level.

Before I embark on this project, there are a few points that maybe others can advise on:

- Does the freshness of coffee play a role with regards to accuracy of the results? I plan to use beans that are a few months old and I have no other use for. Is this going to affect the results in any significant way compared to using (i.e. wasting) fresh beans?
- Meticulously separating the particles under a magnifying glass (improves) the test results significantly. Is there some easier way to do it? I tried using the vibration of a phone but it was too gentle to make a difference.
- Is there a way to prevent grinder retention from affecting the results? The protocol I have in mind is to grind until the grinder is clean of beans, then use a blower to blow off what's left, grind a few grams, throw them away, and finally grind a few more grams and use this coffee as the test sample. Makes sense?

The grinders that I plan to test are Baratza Sette 270Wi, Breville BCG800, Solis 166, 1Zpresso JX Pro and Hario MSS-1B.

I've done some initial testing with Jonathan Gagne's Grind Size application and it seems to do the work.

How do you know it's accurate?

The results are quite consistent between different samples with the same grind setting, as long as I spend some time to distribute the particles and as long as the image quality is high.

Specifically for image quality, I use a 90mm macro lens stopped down to f/11, 24MP at base ISO, close distance with camera sensor plane parallel to the surface where the ground coffee is placed on. I also use soft light to ensure that the particles don't cast shadows.

I've contemplated doing the same thing and will watch this thread with interest. Agree it's hard to distribute the grounds evenly. I poured the grounds through a small kitchen strainer which helped with initial distribution, but it's imperfect and I still needed to do some manual separation. Note the strainer holes are larger than the biggest particle so nothing was strained out.

I'm not sure those applications measure what we think they're measuring. A little while ago, my brother and I put some ground coffee under his microscope, and I was surprised by what we saw. At a coarse magnification:



You can see there are large structures made up of a lot of smaller particles. We put some grounds into oil to break them up a little and this is what we saw at a much finer magnification:



Here, I believe the individual particles are around 50 um in size, which is about what I would expect given the grind settings.

This makes me think the large structures in the first photo are dominated not by the grind setting, but by the small-scale clumping behaviour (not visible to the naked eye) of the individual coffee particles. It wasn't until we put the grounds in oil that we were able to separate these clumps into individual particles.

When we put grounds on a white background and take a photo with a regular camera, I think what we're seeing is dominated by these larger structures, so it's hard to see how this gives any insight into grinder characteristics. Maybe someone with more experience in coffee science can chime in?

Interesting about the use of oil; I was wondering if that could aid separation. I think the expensive particle imagers use a vacuum to suspend the particles. My own brief attempts to use the Gagné app left me frustrated as getting a decent sample was so difficult. Even the instructions that came with the app don't have a really good image.

I asked about accuracy but perhaps consistency is good enough to learn something about the grind. Will it correlate with taste?

This visual grind analysis system sounds (looks?) interesting and maybe it will prove to be practical and accurate, but it seems to me that the classic mechanical technique involving graded mesh strainers and checking the weight of the resulting sifted grounds is very simple, easy to perform, and very well standardized.

jpender wrote:I asked about accuracy but perhaps consistency is good enough to learn something about the grind. Will it correlate with taste?

Agreed. There's a part of me that really wants to have accurate measurements of an obvious physical attribute like individual particle size. But I don't know for sure that this is the best predictor for what happens "in the cup". It could be that the size of the larger "clumps" is more predictive than the size of the individual particles.

That said, the famous paper by Cameron et al.[1] used a model based on a packed bed with fine particles on the order of 50 um in diameter, and got results which I would say many in the coffee community have validated with experiment. This suggests to me that once you tamp the puck, those "clumps" merge and what you're left with is a packed bed made up of the smaller particles surrounded by natural oils. If this is the case, then it's hard to see how measuring the size of the larger "clumps" would be helpful.

[1] https://www.cell.com/action/showPdf?pii ... %2930410-2

It's a really difficult problem. Consider that any individual particle might be a rod, a sphere or a disk, while not an accurate representation, you can see that measuring with a camera, a sphere always measures the same, a disk is somewhere between a sphere and a rod on it's side and a rod is between being a rod and being a point. I think in the end, the measurement over time from one grinder with one set of burrs will keep you apprised of the condition of the burrs. Using it for anything other than for gross determination of particle spread of gross adjustment of burrs for picking between vastly different grind sizes is likely still futile.

But please, prove me wrong.

As far as I can tell, out in the world where people measure grind size constantly as part of the packaging process, the optical measurement of grind size and distribution is mainly beneficial due to the dramatically increased speed of measurement and the ability to measure almost in real time.

Ira

That was a fascinating read and an excellent article. While it did a good job of explaining how espresso extraction varies as coffee grinding/dosing/prep and brewing parameters are changed it seems the analysis of the end product (brewed coffee) focuses more on its physical characteristics that are easily measured with refractometers and other lab equipment than what I am interested in (how the stuff appeals to my senses). I know that our senses respond to the physical attributes of the stimuli that trigger sensory impressions, but this is the part of the article that is less convincing for me. The authors make it clear that a major goal of their research is improving the efficiency of the coffee making industry.