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

[strantor]

Looking forward to hearing what you think of this. Last time I looked it was still not available for a Bambu X1C but I will probably spring for one when it is.

Ok I received the diamondback nozzle in 0.8mm "volcano" form factor, tested it with 4 different filaments, and incorporated the data into my archive. Excerpts of the data which give relevant comparison to the diamondback nozzle are shown below.

Notes/caveats about the test:

• This test is performed extruding into open air just as a benchmark, to see how much filament can be physically pushed through the hotend and nozzle before the extruder begins missing steps, filament starts slipping on the gears, or in the case of super flexible filament, filament exits laterally from or gets wrapped around the gears.
• Actual printing speeds will be lower than what you see below (typically 60-90% of the volumetric values shown in the tables)
• The test procedure is as follows:
  • Extrude 415.75mm (equal to 1cc for a cylinder with a diameter of 1.75mm) into a poo pile at a "safe" speed.
  • weigh the poo pile on a milligram scale, compare to data sheet density spec (g/cc), calculate amount of over/underextrusion
  • Increase speed by 1-2mm/s, repeat.
  • Keep increasing and repeating until calamity strikes or extrusion drops below 97%
  • Repeat test 3X using the highest number successfully obtained. If any of the 3 fails, drop speed and repeat 3X.
  • The highest number which can be achieved 3X is what is shown in the tables below.
• I only set out to compare the diamondback 0.8 nozzle to the 0.8mm and 1.0mm cheap china brass nozzles that I get off amazon, but halfway through the tests I began to notice that my 0.8mm cheap nozzle was performing better than on previous tests. I figured out that the nozzle I had been using was one that I had absent-mindedly ran a #67 reamer though (0.032"/0.813mm) yesterday when I received the reamers and barrel laps that I ordered. Since I had already collected so much data with this "modified" nozzle I decided not to delete it, but keep it and present it here since it seems relevant to the comparisons being made.
• The extruder I am using is an OmniaDrop V2.1, which is designed specifically for use with flexible filaments and you will get different results with a different extruder.
• I only tested the diamondback nozzle in my double-heater hotend (but I included data from previous tests, with other hotends including the stock one, for comparison). Tests with a more orthodox hotend setup may or may not reflect more favorably on the diamondback, I don't know.

Now that's out of the way, here is the data:

Material:		CoexFlex 60A
Hotend/extruder	Temp	Nozzle size	mm/s of filament	Gcode	mm^3/s volumetric	measured weight (g)	over/under
Omnia Volcano	230​	0.6 CHT	1.50​	G1 E415.75 F90	3.6​	1.139​	97.4%​
Omnia Volcano	230​	0.8 brass unreamed	2.00​	G1 E415.75 F120	4.8​	1.174​	100%​
Omnia stock	230​	0.8 brass unreamed	2.25​	G1 E207.875 F135	5.4​	1.158​	99%​
Omnia stock	230​	1.0 brass unreamed	2.50​	G1 E207.875 F150	6.0​	1.15​	98%​
Omnia DOUBLE Volcano	230​	0.8 brass reamed	2.75​	G1 E415.75 F165	6.6​	1.173​	100%​
Omnia DOUBLE Volcano	230​	0.8 brass unreamed	2.75​	G1 E415.75 F165	6.6​	1.154​	99%​
Omnia DOUBLE Volcano	230​	0.8 diamondback	3.00​	G1 E415.75 F180	7.2​	1.17​	100%​
Omnia DOUBLE Volcano	230​	1.0 brass unreamed	3.00​	G1 E415.75 F180	7.2​	1.17​	100%​
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
Omnia stock	245​	0.8 brass unreamed	7​	G1 E415.75 F420	16.8​	0.952​	98%​
Omnia DOUBLE Volcano	245​	0.8 brass unreamed	9​	G1 E415.75 F540	21.6​	0.939​	97%​
Omnia DOUBLE Volcano	245​	0.8 brass reamed	10​	G1 E415.75 F600	24.1​	1.000​	100%​
Omnia Volcano	245​	0.8 brass unreamed	11​	G1 E415.75 F660	26.5​	0.977​	100%​
Omnia DOUBLE Volcano	245​	0.8 diamondback	11.5​	G1 E415.75 F690	27.7​	0.967​	99%​
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
Omnia stock	216​	0.8 brass unreamed	8​	G1 E415.75 F480	19.2​	1.222​	97%​
Omnia Volcano	216​	0.6 CHT	8​	G1 E415.75 F480	19.2​	1.221​	97%​
Omnia stock	240​	0.8 brass unreamed	9​	G1 E415.75 F540	21.6​	1.217​	97%​
Omnia stock	216​	1.0 brass unreamed	12​	G1 E415.75 F720	28.9​	1.221​	97%​
Omnia Volcano	216​	0.8 brass unreamed	12​	G1 E415.75 F720	28.9​	1.23​	98%​
Omnia stock	240​	1.0 brass unreamed	15​	G1 E415.75 F900	36.1​	1.22​	97%​
Omnia Volcano	216​	1.0 brass unreamed	18​	G1 E415.75 F1080	43.3​	1.219​	97%​
Omnia DOUBLE Volcano	216​	0.8 brass unreamed	21​	G1 E415.75 F1260	50.5​	1.225​	98%​
Omnia Volcano	240​	1.0 brass unreamed	21​	G1 E415.75 F1260	50.5​	1.21​	96%​
Omnia DOUBLE Volcano	216​	0.8 diamondback	24​	G1 E415.75 F1440	57.7​	1.235​	98%​
Omnia DOUBLE Volcano	240​	0.8 brass unreamed	27​	G1 E415.75 F1620	64.9​	1.229​	98%​
Omnia DOUBLE Volcano	216​	1.0 brass unreamed	27​	G1 E415.75 F1620	64.9​	1.214​	97%​
Omnia DOUBLE Volcano	240​	0.8 diamondback	29​	G1 E415.75 F1740	69.8​	1.239​	99%​
Omnia DOUBLE Volcano	240​	1.0 brass unreamed	34​	G1 E415.75 F2040	81.8​	1.225​	98%​
Material:		Polymaker Polysonic PLA
Hotend/extruder	Temp	Nozzle size	mm/s of filament	Gcode	mm^3/s volumetric	measured weight (g)	over/under
Omnia stock	245​	0.8 brass unreamed	12​	G1 E415.75 F720	28.8633825​	1.22​	98%​
Omnia Volcano	245​	1.0 brass unreamed	18​	G1 E415.75 F1080	43.2950738​	1.257​	101%​
Omnia DOUBLE Volcano	245​	0.8 brass unreamed	28​	G1 E415.75 F1680	67.3478925​	1.216​	98%​
Omnia DOUBLE Volcano	245​	1.0 brass unreamed	28​	G1 E415.75 F1680	67.3478925​	1.214​	98%​
Omnia DOUBLE Volcano	245​	0.8 diamondback	29​	G1 E415.75 F1740	69.7531744​	1.22​	98%​
Omnia DOUBLE Volcano	245​	0.8 brass reamed	30​	G1 E415.75 F1800	72.1584563​	1.214​	98%​

The following data seems a bit inconsistent and I am not yet sure why:

• The 0.8mm reamed brass nozzle performed better than the diamondback and the PLA filament but worse than both of them with Coexflex 30D and CoexFlex 60A.
• The reamed nozzle performed worse than or the same as the unreamed on the flexible filaments.
• The 0.8mm unreamed brass nozzle performed better with the 78A filament with only 1 volcano heater, while the double heater arrangement was vastly better for every other filament test, and that goes for the other dozens of filaments I've tested as well.
• The stock hotend performed better than the single volcano hotend for Coexflex 60A but like above, that has not been the case for any other filament I have tested.

I have not actually used the diamondback nozzle to print anything yet but in my experience this open-air test is a direct indicator of how the given filament and heater/nozzle setup will perform, and I am not seeing anything here that indicates that the diamondback offers any quantifiable advantage. I will try printing with it next time I am running a part that is appropriate for a 0.8mm nozzle, maybe there is some hidden magic that only manifests with a moving print head, but my assessment for now is:

....meh. Not worth a hundred bucks. Infinitesimal advantage over a normal nozzle.

I think the glowing reviews of it are like the reviews of shoes, watches, pocket knives and handbags that cost too much. When you spend that much on something, you want to feel like it was money well spent, and you want everyone else to think so too. So you deliberately avoid any kind of quantifiable testing of it, and instead say vague things like "wow, night and day difference!"

 

[Inferno]

I'm in the preheat camp. Getting the filament to 100C in the first stage will probably be about right.

 

[homebrewed]

Interesting information there. It looks like the main decision you need to make is how close to 100% you want (in the over/under category), because most of the nozzles' performance is pretty close --- with the filaments you evaluated.

Perhaps the main question is what kind of variance you are willing to tolerate, and the mechanical properties you need.

 

[Inferno]

Here's an idea.
Duplicate your extruder.
If you run parallel extruders you can print parts twice as fast. You just sacrifice the max size you can print.
Set your part to print on one side of the bed and the other side is the clone.

 

[Dabbler]

Here's an idea.
Duplicate your extruder.
If you run parallel extruders you can print parts twice as fast. You just sacrifice the max size you can print.
Set your part to print on one side of the bed and the other side is the clone.

Thse are produced commercially. they are called IDEX printers.

 

[strantor]

Here's an idea.
Duplicate your extruder.
If you run parallel extruders you can print parts twice as fast. You just sacrifice the max size you can print.
Set your part to print on one side of the bed and the other side is the clone.

I'm making large parts. I just had to buy a bigger printer (420x420mm) to make the next batch which will be 16" robot tires (1-piece wheel/tire). So I really need to just be able to print flexible filaments really fast.

It still would be cool to have an idex printer though, for the multi-material aspect. I bought the Neptune 4 Max specifically because it's big, because has Klipper firmware which I've grown familiar with, and because it uses bed slinger cartesian motion with POM rollers on V-strut. It's basically a giant ender 3 with bells and whistles - "so last year" technology now that every "serious" printer is going to coreXY, but I have extra long pieces of v-strut and an idea for turning it into a multi-material machine by hanging multiple gantries off the X-axis strut. The plan is to have a continuous belt on the x-axis and the tool heads just grab onto it whenever it's their turn. You could do the normal Idex stuff (multi-material, duplicate, mirror (by grabbing the opposite side of the belt)) but not limited to just two tool heads. Might have two parked on the left side, one parked on the right side, and one cruising around over the bed. When the left-most tool is called, the middle one parks to the right, secon-to-left goes to right, then the left-most tool comes out to play. I could immediately use this ability to print the upcoming job but it wouldn't be ready in time. I could print the wheel hub out of a hard material and the tire printed in a softer material in the same print, fused to it, and a 3rd even softer material to to make an airtight membrane (not an inner tube, long story) in the center of the tire.

 

[strantor]

Interesting information there. It looks like the main decision you need to make is how close to 100% you want (in the over/under category), because most of the nozzles' performance is pretty close --- with the filaments you evaluated.

Perhaps the main question is what kind of variance you are willing to tolerate, and the mechanical properties you need.

What I showed was just the tests which passed, for each nozzle/hot end/filament combination. My exclusion of the failed tests (for the sake of brevity, there are 100 more entries for every one entry you see in the table) might make it look like all the combinations performed the same, if you are looking at the extrusion % column. To see the difference, check the "mm^3/s volumetric" column. In that column you will see, for the Polysonic PLA for example, with the stock heater and nozzle I could print 28mm³/s without going under 97% extrusion, but with my double heater setup I was able to print 250% faster at 72mm³/s without going under 97%. 97% is what I have defined as my lower acceptable limit, so anything below that is not shown. 72mm³/s is probably "crazy fast" in the opinion of most people who 3d print but still with the scale of the parts I need to make and materials I need to make them out of, the parts could take 24hrs or more, which is risky. I would like to double the speed again from here. If the lapped barrel doesn't get me there (I will be trying it this weekend) then I think I have found the limits of what is possible with 1.75mm filament extrusion and need to step up to 2.85mm or maybe even a pellet extruder.

EDIT:
I think I may have read something that you didn't write. You said the performance of the nozzle is pretty close but I took that to mean the whole nozzle/hotend assembly. If you meant just the nozzle (i.e. 0.8 diamondback vs 0.8 brass) then yes, I agree, there is little discernable difference. In fact the brass outperformed the diamond more than once.

 

[strantor]

I'm in the preheat camp. Getting the filament to 100C in the first stage will probably be about right.

I really don't think it will work but I will try it. It should be easy to try since I already have two heaters in series, I'll just set the top one to 100C and see what happens. I'm not sure it will be able to get down to 100C though, since it's thermally coupled to the lower heater which is at 215C+.

 

[homebrewed]

What I showed was just the tests which passed, for each nozzle/hot end/filament combination. My exclusion of the failed tests (for the sake of brevity, there are 100 more entries for every one entry you see in the table) might make it look like all the combinations performed the same, if you are looking at the extrusion % column. To see the difference, check the "mm^3/s volumetric" column. In that column you will see, for the Polysonic PLA for example, with the stock heater and nozzle I could print 28mm³/s without going under 97% extrusion, but with my double heater setup I was able to print 250% faster at 72mm³/s without going under 97%. 97% is what I have defined as my lower acceptable limit, so anything below that is not shown. 72mm³/s is probably "crazy fast" in the opinion of most people who 3d print but still with the scale of the parts I need to make and materials I need to make them out of, the parts could take 24hrs or more, which is risky. I would like to double the speed again from here. If the lapped barrel doesn't get me there (I will be trying it this weekend) then I think I have found the limits of what is possible with 1.75mm filament extrusion and need to step up to 2.85mm or maybe even a pellet extruder.

EDIT:
I think I may have read something that you didn't write. You said the performance of the nozzle is pretty close but I took that to mean the whole nozzle/hotend assembly. If you meant just the nozzle (i.e. 0.8 diamondback vs 0.8 brass) then yes, I agree, there is little discernable difference. In fact the brass outperformed the diamond more than once.

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.

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.

 

[strantor]

I'm in the preheat camp. Getting the filament to 100C in the first stage will probably be about right.

Ok I tested this. As I was concerned might happen, I wasn't able to get the "preheat zone" (upper hotend) down to 100C. Since the lower hotend is dumping heat into the same brass barrel that the upper hotend is installed on, heat from the lower corrupts the upper. I would need to install a heatbreak between them with a fan, turning the upper hotend into a kind of second cool end, with heating ability. I am not going to invest time on that right now, as I did get a chance to see what would likely be the result. Here I did a few extrusion tests and the temperature of the "preheat zone" drops each time I do it, as the filament draws the heat out. I tried running a longer test, extruding a few meters of filament to see if I could get the temp down to 100C that way. At the beginning of the test it was extruding just fine and as the upper hotend got colder and colder, there started to be issues. The stepper started missing steps, filament stripping in the gears, and it stalled before it got 3000mm out.