Hacking the DF64 exit chute - Page 15

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

#141: Post by markmark1 »

Just printed one with PLA. Completely choked/stalled it twice with old dark roast, mainly cause I'm stupid and forgot to adjust coarser.
Cleaned it out and it was way better than the stock one (also removed the V2 clump crusher) this morning with light-medium roast.

Just a little remaining at the end when hitting the bellows. Also the grinds started coming out immediately. With stock, there was a couple of seconds delay (I guess because of the clump crusher).

GDM528 (original poster)

#142: Post by GDM528 (original poster) »

I got a clog - yay!

Input was 18g of a dark roast (Full City+, all beans lightly oily) peaberry. Grind level was set to the finer side for espresso. I had ground this roast prior, after lightly misting the beans, aka RDT. This time I thought I'd try leaving the beans dry and skipped the RDT... <cue ominous music>

I'm also using a bean input flow restrictor that blocks off half of the input area to the burrs (improves grind uniformity), so I'm not rushing the grind. Ironically, the input flow restrictor is supposed to reduce the propensity of clogging...

First third of the grind seemed to proceed normally, then about halfway through the grind the motor started straining and started stalling around the final third of the beans. I removed the adjustment collar, lifted off the dosing funnel and saw this:

If you zoom in, you can see the grinds backed up in the grind chamber. When I cleared the chamber, I could see the input to the exit chute was packed with grounds. When I removed the downspout, I saw this:

So, the clog started at the narrow portion of the neck of the expansion chamber, then accumulated all the way up and into the grind chamber. Yikes. The plastic for the downspout and expansion chamber are translucent, so if I had been paying attention I could have intervened before it got worse.

Some things I noted:
1) The grind in the clog grew progressively denser away from the start of the clog. The start of the clog was remarkably fragile, such that a light tap was enough to clear the clog in the downspout. By the time the packed grind reached chamber it had turned into an unrecoverable oily paste. The combination of oily beans and a fine grind is a recipe for disaster.
2) There was minimal compaction of grind in the burr teeth, which implies the bean flow restrictor is keeping the burr action healthy. On the other hand, the restrictor reduces the flow rate, which may in turn reduce the self-cleaning effect.
3) The dry (no RDT) grind output was very 'staticy', and sticking aggressively on the expansion chamber and catch-cup walls. I'm regretting not misting the beans before grinding.

After cleaning out the grinder, I repeated the same setup: same beans, 18g, fine grind, no RDT. The grind proceeded without any hint of clogging. So, it's not a given that it will happen every time, and unfortunately there were no obvious signs that I'm getting close to triggering a clog. A nickel is likely to land on its edge roughly every 6,000 flips (https://en.wikipedia.org/wiki/Coin_flipping) - and a grinder produces many millions of coffee particles, so lots of chances for things to get weird.

I'll be going back to misting the beans before grinding for two reasons:
1) The grind seems to flow more efficiently through the grinder, and is less prone to stick to the expansion chamber and catch-cup walls. The resulting grind looks more granular, but there's no structural integrity to it.
2) The espresso shots pull better with the RDT'd grind, with noticeably less channeling effects. After seeing what happens when the grind is compacted in the grinder, I don't want to duplicate that in the puck.

I'm feeling more dedicated to reducing all sources of fines from the DF64, as they seem like an operational hazard throughout the entire workflow. I get that fines contribute to flavor in the cup, but there may be other ways to extract those flavors without having to navigate around grinder clogging and shot channeling.


#143: Post by iyayy »

it will be fine when there is no fines :lol:


#144: Post by ltanzil »

GDM528 wrote:I got a clog - yay!

i got the same experience with the original chute.

i tried your advice to remove the declumper all together. i got lot less retention in the burr chamber. but the exit chute get fully clogged after prolong use at low rpm with dark roast.

since i don'thave a resin 3d printer i have to use old school technique. i made exit chute using 19mm brass pipe.
you are correct the metal chute does not eliminate static, but there is some reduction. i extend the chute about 3/4 inch.
the question for me now is should i build the shroud/expansion chamber?

GDM528 (original poster)

#145: Post by GDM528 (original poster) »

ltanzil wrote:... at low rpm with dark roast.

... should i build the shroud/expansion chamber?
I see you're using a motor speed controller - what rpm do you typically run at? I've been mulling over making a line frequency halver to drop my 60Hz line frequency to 30Hz, to cut the grinder rpm down to about 600rpm.

I added a shroud to capture fine particles that would otherwise go drifting away from the catch cup and stick on the grinder (I hate cleanup). If all your grind is landing in the catch cup, then no need for a shroud. The expansion part was partly simply because it's easy with a 3D printer. The overall concept was to get all the grind into the catch cup and keep it there long enough for the static charge to dissipate - more than one way to do that.

Your "old school technique" is really good! I like the worm-drive adjustment for the grind collar. Now you can add a stepper-motor and a microcontroller to the knob and create real-time variable grind profiles - I'm not aware of any grinder that does that.


#146: Post by ltanzil »

GDM528 wrote:
Your "old school technique" is really good! I like the worm-drive adjustment for the grind collar. Now you can add a stepper-motor and a microcontroller to the knob and create real-time variable grind profiles - I'm not aware of any grinder that does that.
:D thank you, i have never thought of that hehehe. that would be world first. grind size distribution profiling , combined with rpm profile control by arduino. worth pondering about.

for dark roast i typically run at 400 rpm to get abit more of fruitiness, for light roast i typically run at 700 rpm to get more balance taste and abit more body.

once again you are correct. without RDT particle would just floating away form the dosing cup and create mess everywhere. Floating up like fairy dust hahaha, so weird.... the lower the rpm the more pronounced the effect, especially at 400 rpm. extending the chute and using brass does reduce it some how, but does not eliminate the fairy dust effect. i think i have to build some sort of shroud then. i do also hate the mess when i forgot to do RDT.


#147: Post by Jonk »

Impressive :) just a cover for the VFD away from completing the Ceado Hero look.

If you're serious about variable grind setting, I'd suggest to try it manually first. I have*, and small adjustments (with breaks in between) pretty much just resulted in a worse grind (similar to inferior alignment or subpar burrs). Big adjustments might be worthwhile though, like an automated staccato espresso. Those can be pretty tasty in my experience (but also quite bad, seems like a high risk/reward concept).

*not with a DF64 though..


#148: Post by ltanzil »

this is the first time i heard about staccato esspresso. emm... very interesting, this warrant further investigation.
VFD box is still work on progress.... :D

but the one that i have not find the solution is how to build exit chute knocker. I think DF64 is missing a exit chute knocker, i also think, with it chances of the chute clogged will be eliminated.

Any idea how to build knocker for DF64 exit chute?


#149: Post by iyayy »

perhaps i may suggest printing a replacement for the whole front chute cover instead?
then u have more option to go wild.
the portafiler holder still screws to the body, no issue there i think. issue is if you want a knocker, will 3dprint material works instead of metal?
otherwise maybe you can design extended nozzle that goes further out than the front chute cover, and knock on that instead? and maybe spring load it from bottom side / body.
im not sure however how flexible a 3dprint is to multiple flexing and knock stress tho.

GDM528 (original poster)

#150: Post by GDM528 (original poster) »

iyayy wrote:perhaps i may suggest printing a replacement for the whole front chute cover instead?
then u have more option to go wild.
Funny you should mention that - it's exactly what I had before the design I posted here. A 3D printed replacement front cover actually looks pretty nice but required rethreading the power switch. The wires to the power switch are soldered on - I'm guessing because of vibration concerns. That meant cutting then reconnecting the wires to the switch, which seemed a bit too complicated for some of the hackers out there. Just replacing the chute was simpler.

I like the idea of some sort of built-in self-knocking mechanism...

I try to remember to deep-purge the chute periodically, to prevent fines from permanently accumulating inside the chute. "Deep purge" = aggressive pumps on the bellows into an empty catch cup. If I forget, the fines/oils/RDT can build up at the top of the chute. FWIW I designed the curvature of the chute so you can easily see the buildup and pop it loose with a toothpick/stick/etc. The original factory design had to be disassembled to clean thoroughly.