This Changes Grinder Testing Forever... For Free!

[@homecafecharlie]

Okay so a lot of you messaged and commented with your results and it looks like there are a lot of repeatability and accuracy issues with the grind size analyser. The problem seems to be that camera fidelity is not high enough to detect particles smaller than about 450 microns, so this test could not be reliable for espresso. Repeatability is also an issue and while in my testing I got very similar results with the same grinders (repeated 3 times with each) this is highly dependent on distributing the grinds evenly on the paper with a brush or some distribution method that reduces clumping significantly. I will see if I can get a DiFluid Omni particle analyser and compare, but for now I think the unspeciality analyser needs a bit more work to ensure reliability. I'll message the team and see what they plan to do to upgrade this tool, and will update the pinned comment again when I hear more. Thanks everyone for testing this out for yourselves! I love that there are so many curious and scientifically minded people in this community.

[@PCH2024]

I compared this result to results from 2 other apps and got similar numbers so I think this result is accurate. I got a reply to my email to unspecialty today: This is Ukjin from and I'm a co-founder of this company. I've input the image you attached and confirmed it only recognized the 2 particles which means almost every particles were recognized as fines. Our algorithm excludes the fines when calculating the average particle size. This is because the grind size analyzer is optimized for brewing(hand drip/filter) coffee and not for the espresso yet. We will mention this on the web page to minimize the confusion caused by inaccurate results in espresso coffee powder. We will keep continuing the improvement and analyzing the espresso coffee powder accurately is one of our top-priority. Thanks for your thoughtful feedback and sorry for the inconvenience you experienced due to the inaccurate result.

[@kenal1532]

Im saving this and share it on my local group with your link, thanks

[@erikngomez]

Since you are talking to the author of the tool, it would be great if they could say what margin and printer settings to use. If you use zero margins the four squares will be smaller in size. DiFluid has a similar printout and they provide a 104mm ruler on the paper so when you print it, you can take a real ruler and know with 100% certainty you have the same measurements others are using.

[@homecafecharlie]

That's smart! I'll ask Ukjin about that

[@Phil_OG]

Very important! Otherwise it's completely useless...

[@MsKingoDisco]

From what I see in the document, it has to be 12 cm between boxes, from outside to outside

[@FatNorthernBigot]

What a great idea. Nice way to use our existing technology.

[@homecafecharlie]

I think the cheapest specially made one is about $700 so this is great!

[@nathanpiccio9127]

can't wait to try this with my spice grinder!

[@dubios42]

Wow, great job, I'm going to try that out right away :D I just wonder how consistent the results will be.

[@ziptiefighter]

Color me intrigued. I bought a Eureka Zero a few months back and was considering getting a set of Kruve sieves. I like this better. It would seem to be, at the least, a good reference point. P.S. Why'd the audio cut out @5:25?

[@homecafecharlie]

Oh I don't know what happened with the audio! I will check the export settings 😅

[@PCH2024]

OK. I very carefully distributed grinds from my Mazzer Major that I use daily for espresso with a Rancilio Sivia. I picked up the grinds with a fluffed out paintbrush and lightly tapped the handle 2 feet above the printed test pattern to make an even distribution of particles with almost no overlaps. The results I got were misleading. It says "Average of Particle Size 466.70 µm", even though the graph shows it at around 700 µm. "The particle size of 466.70 µm is suitable for Moka pot. It may be a bit too fine for hand drip." The COFFEE GRIND SIZE GUIDE shows ranges of particle sizes for Moka Pot - Brewing - French Press/Cold Brew. Not espresso where grind size is most crucial. 466.70 µm is at the high end of the Moka Pot range. Yet I make consistent shots of espresso at 2 fl oz using 15 gm of coffee with PID controlled temperature day in day out. This test is useless for espresso grinders.

[@homecafecharlie]

Thank you for sharing your testing with your setup! The particle size seems to be inconsistent and while my tests could be repeated if I met the right conditions, it is a bit too sensitive to the equipment and distribution being perfect. Tools like the Omni use a light tray and vibration to settle the grinds and take several images of the same grinds when moved to make a more accurate approximation. I suspect phone cameras simply lack the fidelity to detect particles that are under about 450 microns. I'll add a pinned comment with context 😉

[@PCH2024]

@@homecafecharlie Thanks for your reply. The image is 3024×4032 and was taken using the flash. That's high enough resolution for a tool like this to work I think. Several images would be preferable though. I will run this by the authors of this tool.

[@alisonw3483]

That is awesome!

[@homecafecharlie]

It really is!

[@nathanpiccio9127]

Just wondering whether they have a specific smartphone camera that they'd recommend. I'd imagine the sharpening on some smartphones could alter the results. Interesting concept though! Nice to see further democratisation of coffee equipment.

[@homecafecharlie]

Good point. Sharpening filters may change the results. I'll get more details from Ukjin and add to the comments if there is a recommended setting.

[@gregorytortissier5518]

Nice one Charlie, definitely gonna give it a try. Hope being back to UK is all what you expected. Missing Japan ^^?

[@homecafecharlie]

I miss Japan sometimes, but lots more opportunities in coffee in the UK for me. I will be going back to visit Japan this year though, and working on a cool project there IO'll talk about nearer the time ;)

[@markholm7050]

Need to have someone with a dedicated particle size analyzer check this method versus their machine. I suspect there may be sources of error that need to be investigated. James Hoffman?

[@homecafecharlie]

I'm thinking to just buy one and test it out properly. I thought this was a cool experiment but the comments have me wondering if it is as accurate as it needs to be

[@stimhack]

I tried it with the same grinder, same beans, same grind setting, and got two *dramatically* different results. It was also extremely finnicky to get to work. Cool idea, poor implementation at this time.

[@homecafecharlie]

Thanks for sharing! I had a few results differ slightly, but the overall distribution was reasonably accurate. I'm going to buy a professional tool with which to compare it so I'll share those results in a short or at least the pinned post.

[@d4nte857]

this doesnt work trivially, you get vastly different results by rearanging grounds on the paper and retaking the photo (with the same grounds...) nice idea, not useful

[@Phil_OG]

You results look way off. Seems like you measured espresso grounds? No way you are in the 700 micron range for espresso. This probably works better for pour over grounds....

[@homecafecharlie]

You're right it's better for measuring distribution. I'll check with Unspecialty why the range was so high.