Hacking the DF64 exit chute

Grinders are one of the keys to exceptional espresso. Discuss them here.
GDM528

#1: Post by GDM528 »

I've designed a new exit chute for the DF64 to address retention and static issues with the stock grinder. I imagine this may be of interest to readers that either have a 3D SLA printer or know someone that does. There are a lot of makers out there, and I bet most of them drink coffee.

After cleaning my DF64 too many times, I made a couple observations:

1) The closer I held the catch-cup/portafilter to the exit port, the more debris accumulated on the grinder. If I set a bowl on the counter, well away from the exit port, I saw far fewer grounds accumulate around the exit port. My theory is the static 'spray' force is largely between the individual particles of the grounds, and the accumulation around the exit port is simply an aftereffect of the grounds self-repelling. I tried charging the portafilter to about 3KV to see if I could electrostatically attract the grounds, but the grounds were too busy avoiding each other to react to the charged portafilter - besides, the safety issues of a 3,000-volt portafilter would be, well, problematic.

2) Grounds were semi-permanently accumulating and lining the direct path of the exit chute. Between what gets backed up behind the clump crusher and the nooks and crannies of the factory stock exit chute, there could be up to a couple grams of fines and oils that could break loose and make an unwanted contribution to the grinder output - yum. I think root cause is from pressure fluctuations along the path of the grounds: high-pressure at the clump crusher, low-pressure at the top of the factory exit chute, and increasing pressure as the exit chute narrows. This creates pressure pockets that encourage grinds to accumulate.

I addressed the observations in two ways:

1) Redesigned the exit chute so there is a constantly-increasing cross-section along the path, from the spinning rotor to the exit port. Any pressure build-up in the port will find the path of least resistance to be in the direction to the output:



This reaches into to grind chamber and clears the rotors by just a half-millimeter. The elliptical aperture is about 40% the area of the original opening, and similar to the opening of the original design after a build-up of coffee grounds. Lest one thinks this is choking the path of the grounds, the velocity of the grounds exiting the chute is noticeably faster than the original version with the clump crusher.

This new chute drops right into where the clump crusher and factory exit port went - that's right, no more clump crusher!



The new chute extends a bit past the cover plate, and incorporates little magnets for attaching the next half the redesign:



At this point I have a new exit chute that no longer accumulates old coffee grounds. It completely clears with a light tap on the bellows - no more pumping.

However, coffee grounds will still drift around and stick to the grinder and surrounding countertop. So the second half of the solution is to control the space the grounds pass through with a magnetically-attached downspout and expansion chamber that extends into the dosing cup (or portafilter).



As with the bellows, a single light tap on the wall of the expansion chamber will knock loose any grounds in case I was too light with the RDT. My countertop is now spookily clear of grounds as if I wasn't using the grinder every day.

Not missing the clump crusher. Grounds are now significantly fluffier, to the point where I had to use a taller dosing ring on my portafilter.

If there's any interest, I can post the design files online (Thingiverse, GrabCAD).
★ Helpful

Giampiero

#2: Post by Giampiero »

I did a transparent tube chute 6 months ago, and for what i saw i was not really interested to pursuing further test, if you got such good results will be nice to see some videos.

ronner
Supporter ♡

#3: Post by ronner »

Interesting design. If this were made with conductive material, then the screws would ground to the chassis. Would be curious to see if that would reduce static. Keep us posted.

CM00

#4: Post by CM00 »

:shock: :shock: :shock: Just to express my appreciation for the efforts and creativity! Thank you for enriching the community with it!
Could you please upload a video with the flow? And also if you can share the design files, very very grateful! I am ready to give it a try.

Cheers!
mc

GDM528 (original poster)

#5: Post by GDM528 (original poster) »

Giampiero wrote:I did a transparent tube chute 6 months ago, and for what i saw i was not really interested to pursuing further test, if you got such good results will be nice to see some videos.
I couldn't think of a way to convey the difference using a video, but perhaps these pictures will help:

Before the design shown at the top of this thread, I had built an easily-detachable exit spout system so that I wouldn't have to tear apart the grinder every time I wanted to deep-clean it - plus it allowed me to experiment with different exit chute designs. This next photo is from one of my detachable spout designs, that was modeled to resemble the exit chute of the original DF64:



The above photo was taken after several grinding sessions and shows how the grounds can accumulate over time. The entire upper half of the exit path had filled up with grounds that resisted aggressive pumps on the bellows. Ironically, retention after all this accumulation was pretty low at this point, just a few tenths of a gram. The downside to cleaning out the debris is higher retention until the debris builds up again - how very Sisyphusian...

This next photo shows the aperture of the grinder with a similar build-up of grounds along the rough, flat surfaces of the opening:



The grinder's expeller blades move from right to left in the photo, so the grounds are being slammed against the wall of the opening, which is cast - not machined, and the resulting rough texture is conducive for the grind to stick.

Two key takeaways from these images:
1) The grounds accumulated where they weren't getting in the way - which I didn't find very reassuring given how easily they could break loose.
2) Grounds were accumulating in the corners. Rounding the corners actually increased the accumulation - lol.
Those points were the guiding principle in the latest design: no flat surfaces and no unused volume. That meant reaching all the way to the expeller blades by extending into the exit port and even matching the curvature of the rotor - exploiting the capabilities of high-resolution SLA printing.

This next photo is looking up the latest spout design (too lazy to tear apart the machine just for this photo). It clears the expeller blade by about half a millimeter:



After about two weeks of daily use the accumulation is trivial and looks virtually the same as it did after the first grind. There's essentially no build-up in the path now. I had started out thinking I needed to create something what would be easy to clean - but that implies a system that gets dirty.
This new design appears to be maintenance-free so far, with virtually no retention effects. And the area around the grinder is still spookily clean.

The resulting grind is lightly granular; I suspect due to localized static accumulation effects. I've observed the texture of the grinds is affected by the amount of RDT I use prior to grinding: more water = more granular. The next photo shows the grinds in the dosing cup after gently tipping the cup to level the surface. 18g of a Full City roast fills the DF64 dosing cup to the half-full point.


GDM528 (original poster)

#6: Post by GDM528 (original poster) »

ronner wrote:Interesting design. If this were made with conductive material, then the screws would ground to the chassis. Would be curious to see if that would reduce static. Keep us posted.
My thinking exactly - I even got some conductive 3D printing filament so that everything in the path of the grind would be electrically tied to the grinder body. I tried using a metal tube for the spout that was grounded to the chassis... Didn't make a difference. Only a small fraction of the grind actually comes in contact with the downspout.

I even tried electrostatically attracting the grounds to the portafilter - also didn't help, not to mention dangerous :)

I think the static (technically, the triboelectric effect) is kinda inevitable. There are things I can do to that will mitigate it, such as RDT, but I'm not gonna make it go away. BTW, after I spray the beans and drop them into the grinder, I let them sit in the chamber for about a minute to give the beans more time to electrically equalize with the grinder - not a huge difference, but does help a bit.

The mass of the grind particles is very low compared to the amount of charge they can hold, and because they have the same polarity charge they're aggressively repelled from each other - like a static electricity explosion. Hence the funky downspout/expansion chamber that's nested inside the dosing cup - let the grinds do what they wanna do and just control the space in which it happens.

GDM528 (original poster)

#7: Post by GDM528 (original poster) »

CM00 wrote::shock: :shock: :shock: Just to express my appreciation for the efforts and creativity! Thank you for enriching the community with it!
Could you please upload a video with the flow? And also if you can share the design files, very very grateful! I am ready to give it a try.

Cheers!
mc
Welcome to the active-posters, I'm honored to be (or at least what appears to be) your first!

Workflow for using the grinder is the same as a stock unit, so I'm a little fuzzy on how capturing a video would help. There's no special usage tricks beyond RDT, drop in beans, let sit a minute, push power button, tap-tap bellows a few times to flush the system, and done.

I'll look into posting the design files. In the meantime here's a few 3D printer-nerd points:

1) It was designed specifically for SLA printing in order to capture the level of detail and surface smoothness required. I'm using a Form 2 at 100um layer resolution, but I imagine any comparable SLA printer would work as well. I get that SLA printers aren't as common, but you might know someone that has one. Alternatively, there are hobbyist-friendly printing services out there.

2) Sub-100um tolerances for a snug fit to minimize gaps that would accumulate stray grounds. There was a bit of filing/scraping involved to tweak the fit. For example, the holes for the magnets needed to be cleared out with a 1/8" drill bit.

2a) Speaking of magnets, I used N52 (highest grade) magnets for the downspout attachment. K&J Magnetics D22-N52.

3) The exact dimensions and location of the exit port in the DF64 may not be stable if the manufacturer is tweaking their design. For example, the exit port on my machine isn't centered (dunno why), so I had to accommodate by offsetting the plug that goes into the opening (a PITA, BTW). Likewise, the position of the opening in the black cover plate relative to the grind chamber might shift around - it took me several test prints to get that positioning correct.

4) The last part of the downspout is FDM printed, 'cause it's more rugged, cheaper, and looks way cooler. The two parts are designed so they could be merged into a single body if one is so inclined.

5) All parts are designed to print directly on the printer's build platform with no supports required - makes for much cleaner prints. Direct on-platform SLA printing will crush the first millimeter of the model however, and that might vary from one SLA printer from the next. That 'crush' will affect the fit between the main chute and the downspout - not a showstopper, but could create a small gap that would trap a tiny bit of grinds.

GDM528 (original poster)

#8: Post by GDM528 (original poster) »

Design files for the exit chute are posted here:
https://grabcad.com/library/df64-exit-c ... ownspout-1

Posted as .step files to enable modifications, hacks, etc.

spopinski

#9: Post by spopinski »

Thank you!

CM00

#10: Post by CM00 »

Big thanks!