Espresso Distribution Tool Debate Update #2

(Matthew Perger) #1

Hey team!

Here’s another video update for y’all.

Remember: only contributors can comment in here. If you’re reading this and would like to throw your hat in the ring or help us along, you can contribute here.

One thing I missed in the video: The annulus tool will have to stop before the edge of the basket (~54mm sacrificing the very edge) because the edges off the basket aren’t vertical. We can’t know what the volume of the area would be without maths that’s beyond the scope of this experiment!


Espresso Distribution Tool Debate Update #1
(Matthew Perger) #2

Just a quick update for y’all!

OHAUS is sponsoring an analytical lab scale accurate to 0.0001g so we’re good on that front :wink:

This makes me think I could make a tool that does both pizza and annulus sections, and we’d have enough resolution to see what’s going on in all sections at once! Like this:

(Simon Ho) #3

It will be super fun and meaningful to see how it goes!

(Brett Brooking ) #4

With regard to a control method for distribution, has any consideration been given to using vibration to level the grounds? Similar to the manner in which concrete may be leveled and distributed by vibration. I’d imagine this would create a vertical strata of particles, with the fines sinking to the bottom, but should evenly distribute the material horizontally.

(nicolas) #5

is there a market for automated distribution tools? works in the same format as a PuqPress? Or is the human error side of it too small to warrant it?

(Eisen Lai) #6

For your second tool, how do you make sure the volume are the same since the wall of the basket isn’t vertical?

(Matthew Perger) #7

@Eisen_Lai That’s explained in the text below the video! :slight_smile:

(Matthew Perger) #8

Maybe? I think that level of sophistication would be better invested in a full super automation TBH

(Matthew Perger) #9

I know a lot of people have tried. But it’s never created optimal extractions, and is very tricky to achieve a horizontal bed.

Tricky, noisy, electrical (read: expensive). We’re aiming for something cheap and usable by anyone in any situation!

(Andrew) #10

Have you thought about taking one step backward and just 3D printing a fake basket with perfect vertical walls that will fit the pizza annulus, a tamper and all dosing tools perfectly ?

(Matthew Perger) #11

3D printing one that fits a portafilter properly, can be tamped, and won’t deform at all throughout the testing is probably beyond the strength of the materials I have access to. I thought it’d be another explosion of variables I don’t have the time to take care of, especially when simply reducing the diameter of the tool by a few mm’s will make everything fine. We can also measure the total weight of the very outer area (what’s left after the tool) to see if it’s consistent dose-to-dose as well (so, another data point!).

(j l) #12

As a secondary test, just use a bottomless portafilter to visually observe flow. This connects to “real” life for determining even extraction and at a minimum provides an independent check on the above tool.

I hope the tool works but possibly will have issues with applying tool with even force, edges being sharp enough, and also static issues. Won’t know till trying, though maybe simulate with a razor blade first?

(j l) #13

I am going to order a tuning fork and see what happens :smiley:

(Simon Ho) #14

According to Socratic’s test, the OCD was like decreasing the Extraction level, will we also explore this zone(relationship between Extraction and evenness)?

Here is the link of the test:

(marcus delponte) #15

Something I do occasionally, as sort of a poor man’s distribution tool, is let my tamper rest level on the coffee and rotate it several times before pressing down.

Anecdotally, it feels like it is improving the evenness of my extraction, but I hope you guys can test this as well.

(Paul Westfall) #16

Matt- If I understand correctly, you intend to weigh each section of the puck to test for uniformity of distribution. Certainly this will be useful data, but shouldn’t you also measure grind distribution of each section?

(Matthew Perger) #17

From the debate I was in on twitter, some people don’t care about the lower extraction yield and say it’s all about taste. I disagree. If you can taste, you’ll find higher extractions ceteris paribus are more delicious in every circumstance except when grinding too fine.

(Matthew Perger) #18

This is called ‘nutation’. Totally works, and I’m pretty sure it would be a “perfect” distribution method. Definitely going to test. But for it to “win” we’ll have to figure out how to do it consistently and quickly!

@landodafree with today’s PSD technology this is impossible. The sample sizes would be too small to measure. Also, there’s so much friction between coffee grinds that it’s highly unlikely to get any kind of consistent “migration” of particles with any normal distribution method. IMO Vibrating table, p=0.9. OCD, p=0.01.

(Matthew Perger) #19

“simulate with a razor blade” - Hard to do without building the whole tool for them, and then we’re done. It’s a lot tricker to use a razor blade (from a design perspective) than I initially thought. They bend, but you can’t build a frame to hold them or it’ll influence the grinds as you push it down.

3D Printing the tool v0.1 today!!!

(j l) #20

I think I need a bigger fork…

My tuning fork experiment for distribution of grinds with a note of “C” did not work so well. Oddly, even singing at a note of “C” did not help either when harmonizing with the tuning fork.

Cat joined in for a bit, but not on key.

Grinds did move very briefly but that was it. So either a bigger fork and/or full Opera required.