How to make a better 3D printer nozzle? (Machining topic)

[strantor]

Yes, my comment was w/regard to the nozzle, nothing else.

Based on information I found on the web, the Diamondback nozzles use a composite that contains diamond particles. It's possible that the surface isn't as smooth as brass. I base my hypothesis on the fact that diamond is very hard, so any polishing agent capable of polishing diamond would likely be more aggressive toward the binder, leaving the diamond particles a bit "proud" of the surface. Depending on the binder and polishing agent, the roughness might not be all that apparent to the naked eye....or, for that matter, anything less than a high-power microscope or SEM.

Looking at the diamondback nozzle I bought, I can't tell about roughness because now it's got molten plastic on it. But in this screenshot from the video Boswell posted, it doesn't look all that smooth at all.

It would be interesting to see how a lapped/polished brass nozzle would perform, considering your observation regarding the one you ran the reamer through.

Ok I sorta have that answer now. I tested reamed and lapped barrels today, using the same 4 filaments I tested yesterday with the diamondback nozzle.



I tested reamed first, then lapped. The data is in the table below but in the course of testing my performance with the flexible filaments took a nosedive somewhere during the tests of the lapped barrel. The housing that holds the extruder together developed a lot of slop; I think the plastic got soft and creeped, allowing the bearings to move around, and this movement gave the flexible filaments new options for places (other than down the hole) to wander off too. I will need to print a new housing for it, out of polycarbonate so it doesn't happen again, and I think I will modify the OmniaDrop design to make it a little more robust.

Material:		Polymaker Polysonic PLA
Hotend/extruder	Temp	Nozzle size	mm/s of filament	Gcode	mm^3/s volumetric	measured weight (g)	over/under	Gains:
Omnia DOUBLE Volcano (drilled)	245​	0.8 brass reamed	30​	G1 E415.75 F1800	72.2​	1.214​	98%​	0%​
Omnia DOUBLE Volcano (reamed)	245​	0.8 brass reamed	33​	G1 E415.75 F1980	79.4​	1.208​	97%​	10%​
Omnia DOUBLE Volcano (lapped)	245​	0.8 brass reamed	38​	G1 E415.75 F2280	91.4​	1.214​	98%​	27%​
Material:		Overture High Speed TPU (95A)
Hotend/extruder	Temp	Nozzle size	mm/s of filament	Gcode	mm^3/s volumetric	measured weight (g)	over/under	Gains:
Omnia DOUBLE Volcano (drilled)	240​	0.8 brass reamed	27​	G1 E415.75 F1620	64.9​	1.229​	98%​	0%​
Omnia DOUBLE Volcano (reamed)	240​	0.8 brass reamed	27​	G1 E415.75 F1620	64.9​	1.229​	98%​	0%​
Omnia DOUBLE Volcano (lapped)	240​	0.8 brass reamed	27​	G1 E415.75 F1620	64.9​	1.222​	97%​	0%​
Material:		CoexFlex 30D (~78 Shore A)
Hotend/extruder	Temp	Nozzle size	mm/s of filament	Gcode	mm^3/s volumetric	measured weight (g)	over/under	Gains:
Omnia DOUBLE Volcano (drilled)	245​	0.8 brass reamed	10​	G1 E415.75 F600	24.1​	1.000​	100%​	0%​
Omnia DOUBLE Volcano (reamed)	245​	0.8 brass reamed	11​	G1 E415.75 F660	26.4581006​	0.973​	97%​	10%​
Omnia DOUBLE Volcano (lapped)	245​	0.8 brass reamed	8​	G1 E415.75 F480	19.242​	0.978​	98%​	-20% (*)
Material:		CoexFlex 60A
Hotend/extruder	Temp	Nozzle size	mm/s of filament	Gcode	mm^3/s volumetric	measured weight (g)	over/under	Gains:
Omnia DOUBLE Volcano (drilled)	230​	0.8 brass reamed	2.75​	G1 E415.75 F165	6.6​	1.173​	100%​	0%​
Omnia DOUBLE Volcano (reamed)	230​	0.8 brass reamed	3.25​	G1 E415.75 F195	7.8171661​	1.164​	99%​	18%​
Omnia DOUBLE Volcano (lapped)	Did not test - extruder falling apart

(*) Test data corrupted by extruder falling apart.

There is not quite enough data there for a solid trend, but based on the 12 hours I just spent doing this, which the table can't convey, I'm saying that Lapped is way better than reamed, which is way better than nothing. And that's weird because the lapped barrels don't look near as good as the reamed ones. I think I did it wrong or something because the lapped barrels actually look pretty bad, and I almost didn't even bother testing them. The reamer leaves a mirror shine, and the barrel laps leave what looks like tiny-scale orangepeel in a bad car paint job.

Sorry this is really hard to take a picture of, but...

Here is the barrel before doing anything:


Here is the reamed barrel (note that you can make out the individual LEDs in the COB array of my backlight):



And here is the lapped barrel:





Based on this, I am thinking maybe @Dabbler is on the right path with suggesting chaotic flow to mix up the hot and cold parts of the filament. So when I get this extruder fixed I'm going to finish/re-test the flexibles and test yet another barrel type: deliberately destroyed ID (or maybe just threaded ID).

 

[homebrewed]

I'm wondering if the lapping procedure didn't open up the bore a bit more compared to the reamer. My question is based on a comparison of the apparent ID of the bores vs. width of the lip, but that IS just a visual comparison. The two photos appear to be at somewhat different magnifications so it's a little hard to say for certain.

 

[strantor]

I'm wondering if the lapping procedure didn't open up the bore a bit more compared to the reamer. My question is based on a comparison of the apparent ID of the bores vs. width of the lip, but that IS just a visual comparison. The two photos appear to be at somewhat different magnifications so it's a little hard to say for certain.

They were the same ID as best as I can measure. I determined they were, by inserting a tapered carburetor jet reamer in to the reamed one (from both ends) and seeing where it stopped, then doing the same with the lapped ones, and they stopped at roughly the same spot.

 

[strantor]

I got a little off the track of testing different bore surfaces because I was convinced that the bigger issue is just heat... not enough heat. So I added a 3rd 115W heater in series with the first two.




That resulted in a pretty dramatic gain with the firmer filaments (still working with the sloppy extruder so couldn't test the more flexible filaments). The Polysonic PLA I was able to take up to 113mm³/s. That's 24% higher than the best of any of my previous tests with this filament and 4x faster than just about any 1.75mm FDM printer on the market. Overture HS TPU I got up to 76mm³/s which is a 49% improvement over any previous test, and an unreal number for any kind of TPU.

From that I concluded that barrel ID surface is irrelevant, I just need more heaters. I received the new extruder I ordered and I was so confident in my theory that i didn't even bother repeating any flexible filament tests with the new extruder, I just skipped all that and invested 2 days in building this 600W absurdity:





It sucks. It's heavy as hell and it isn't any faster than 3 heaters. Well, a little bit faster for the PLA but for the 95A TPU the performance is exactly the same as 3 heaters, and it won't consistently extrude the 60A or 30D filaments. Those super flexibles come out in spurts, stops, and dribbles like the tail end of a urinal visit. What a disappointing waste of time!

Well, I guess it wasn't a total waste of time as I learned a few things:

• A big heater is better than a small one
• Two big heaters are better than one
• Three big heaters are better than two
• Five big heaters are NOT better than three.
• You can support a hotend by the heater (or by 3 heaters) with a conduit strap and it doesn't take a half dozen layers of stainless/mica sandwich to keep the heat from getting into the gantry. One stainless sheet with mica on either side is sufficient. I went overboard.
• Don't get ahead of yourself
• Bore surfacing methods deserve more investigation with a two or three heater setup and running the flexible filaments.

This isn't over but I have other things that need attending to, so I am pausing for now. When I get back to it, we will have a definitive answer about lapped, reamed, destroyed, and threaded barrels.

 

[homebrewed]

Heaterzilla!

 

[strantor]

Yesterday I finished up the projects which drew me away from this, so today I was back at it, with a new idea. The idea was teflon... doesn't get much smoother than that. Teflon all the way from the bottom of the extruder gears to the nozzle tip.

I ordered a "supervolcano" nozzle meant for 2.85mm filament, and its bore just barely fits a (2mm ID x 3mm OD) teflon tube.



Drilled out and threaded the cool end of a Neptune printer to accept a larger heatbreak that is meant for teflon liner.



The supervolcano nozzle isn't quite long enough for 3 of the 115W heaters but two of them plus a 40W heater fit just right.



The 5/32" copper tube upstream of the heatbreak is just a bushing to keep the Teflon tube centered because the passage between extruder gears and cool end is ~5mm so the 3mm OD tube was flopping around.






I figured this would be (at best) better than any of my previous tests or (at worst) valuable data on whether a smooth or rough barrel is best. Now that I have the data, I can't decide if it is relevant. It didn't work as well as hoped. It is a big improvement over a stock hotend but a step backwards relative to my previous tests. I am printing 43mm³/s with this setup but previously I was hitting 65mm³/s with no teflon, only two 115W heaters, and using a different extruder.

The fact that it's a different extruder, I think does not effect the result much. This is 95A TPU, pretty firm for a flexible filament, and this Neptune extruder should be just as capable with this filament as the OmniaDrop extruder. The teflon is not a great conductor of heat. In fact, it's a thermal insulator. I knew that before I started, but the little voice in my head said "with a melt zone that long it won't matter, just do it." Well now I'm cursing that voice because I spent 12 hours putting this together and in the end I don't know if it's slower than previous setups because teflon is smooth or because teflon is an insulator.

I need to quit getting distracted by shiny things and just test what I set out to test (smooth vs rough metal barrel). I have a load of parts to make so progress in testing will be slow, but I have more printers now so I can do testing while other printers are printing.

 

[MikeWi]

At this point, I'd be wondering if the heat break will be able to handle all that extra heat. Water cooling may even be appropriate.

 

[strantor]

At this point, I'd be wondering if the heat break will be able to handle all that extra heat. Water cooling may even be appropriate.

I had the same thought, and for that reason I ordered a water cooling kit from Biqu and two different water cool end options, but so far haven't needed them. I don't know how the physics works there. I'm not running any hotter than normal, actually I'm running at 215C instead of the recommended 220-240C because I found it reduces stringing and I don't have any parts cooling fan.

The heat isn't any hotter, there's just more of it. Or... is that even a valid thing to say? I think (someone correct me if I'm wrong) the heatbreak is like an orifice valve tapped into the bottom of a bucket with 12" of water in it. If you exchange the bucket for an above ground pool then still, if there is only 12" of water in the pool, the same amount of water will get through the orifice.

In any case I feel pretty good about putting that single pathetic sheet of mica between the hotend and cool end, even though it's probably doing absolutely nothing.