Square Nozzle Orifice



  • Like many before me, I've had thoughts that a square nozzle could potentially result in improved print quality. The idea being that if it's always pointed "forward" you may reduce visibility of layer striations, or that perhaps the increase in surface area will improve part strength.

    Attached is my first attempt at a square nozzle. This is EDM finished out of an E3D "experimental" blank. The square is 0.40mm across and held to extremely tight tolerances (+/-0.0025mm).

    I have only had time to do a couple of test prints, so not much to share visually at the moment. My first impression is that the improved orifice surface finish (cross sections to come) makes a real difference in print quality. Additionally, I could go all the way to .40mm layer heights while retaining good part quality.

    Hopefully before next week I will have some good side by side comparisons to share. Though this post may be a bit premature given the lack of testing, I'm definitely curious to hear people's thoughts on the subject.

    Next up are EDM/WEDM finished round orifices, star, flower, and triangle shapes, as well as consistent ID paths (IE, .40mm from end to end, no large lead-in ID).

    IMG_0300.JPG
    IMG_0295.JPG



  • I admit this is probably a stupid question but .... how do you align the square hole to a tool path other than 0/90/180 degrees ?
    Specifically, what happens when you print a circle? Unless the nozzle turns with the direction of print, I can't see how you get a consistent print width.



  • @jens55 Good question. Right now I am limited to simple square shapes for testing. If it proved successful further experiments would require another axis to keep the toolhead pointed "forward". For now this is firmly in R&D territory.

    To align the nozzle for preliminary testing, I tightened it as normal and used a digital microscope to find the angle. I then adjusted the model in the slicer accordingly.

    EDIT: This is also why I am eager to try some other shapes. Things like the star shape will be less dependent on direction, and the circle will tell me if there is merit to the EDM finish.



  • @plasticfactory I find this an interesting idea. Is it difficult to manufacture such a nozzle, did you need special tools for it?

    The concrete printers turn nozzles in the direction of movement. The 5 Axis robot which I am currently developing https://duet3d.dozuki.com/Wiki/Configuring_RepRapFirmware_for_a_FiveAxisRobot?revisionid=HEAD provides modes P3 and P4 to turn the nozzle.



  • @JoergS5 I am exploring that option, but so far it seems slicer options are limited. Do you have any recommendations? Any pictures or information on the robot you are developing? It sounds interesting.

    As for manufacturing the nozzle, it's not so difficult if you specialize in this type of thing, but it's definitely a niche application of a niche manufacturing method to begin with. In total time, the cost of this nozzle is easily approaching $1000, but with proper tooling (i.e., making more than one at a time) that number comes down dramatically (possibly <$100). The real kicker with EDM work is that it's a two machine process at minimum as you have to make the electrode part "negatives".



  • @plasticfactory said in Square Nozzle Orifice:

    Like many before me, I've had thoughts that a square nozzle could potentially result in improved print quality. The idea being that if it's always pointed "forward" you may reduce visibility of layer striations, or that perhaps the increase in surface area will improve part strength.

    I'm curious to see the results, and I'm glad you are willing to spend the money on having the parts made...

    But the world of rheology shakes its head and wonders why people seem to forget that molten polymer is a viscous fluid...
    (Even though those folks only really deal with constant flow and boring things like that).
    🙂

    I'm really curious to see what the other shapes do.

    Have you considered building an extruder that contains a load cell? I think that extrusion efficiency is likely to be far more interesting than any effect on quality (the load cell being intended to measure required force to extrude at a given speed and temp). I've also not seen anyone since Nophead back in the day, doing any work on measuring extrusion force.
    (its been on my list of stuff to do for about 5 years...)

    The concrete printers turn nozzles in the direction of movement.

    Concrete (contour crafting) is a much less viscous fluid and the scale is very very different. Its theoretically possible to design an FDM nozzle to do what the contour crafting machines do, but it would only help for vertical surfaces, not for anything else.
    (In fact, its probably trivial, if you accept the limit of fixed layer height and probably extrusion width too... plus the 360 degree rotating head)



  • @plasticfactory thank you for explaining the manufacturing process and how much it costs.

    The firmware kinematics has to support it, there needs not be a special slicer. Every G1, G2 etc. movement has a specific direction (or tangent in case of G2), the firmware can calculate the direction without help.

    I am currently developing a robot prototype.



  • @theruttmeister Just to clarify, these parts are being finished in house.

    I've not considered a load cell. Right now I'm more focused on individual components, but hope to get there. I've always thought some sort of optical system adjusting extrusion rates would be amazing. Variance in filament OD is huge for some reason.

    I'm curious how the star/flower shapes perform. I really think it may improve tensile strength, but pull tests will tell.

    I appreciate posts like this, there is a lot to learn.

    EDIT: for clarity



  • @plasticfactory said in Square Nozzle Orifice:

    I have only had time to do a couple of test prints, so not much to share visually at the moment. My first impression is that the improved orifice surface finish (cross sections to come) makes a real difference in print quality. Additionally, I could go all

    Switching from machined stainless to sintered Tungsten carbide with much better internal geometry netted a noticeable improvement in quality.
    I suspect that a laminar flow nozzle would be the best thing if that is even possible.



  • @plasticfactory said in Square Nozzle Orifice:

    @theruttmeister Just to clarify, these parts are being finished in house.

    I've not considered a load cell. Right now I'm more focused on individual components, but hope to get there. I've always thought some sort of optical system adjusting extrusion rates would be amazing. Variance in filament OD is huge for some reason.

    Its hard to get a tightly controlled OD on an extruded polymer... you need to control a lot of factors! Some of the cheap filament extruders are basically working in warehouses without any kind of climate control, not drying their feedstock, building rubble on the ground (cheap Chinese filament is cheap for a reason).

    I'm curious how the star/flower shapes perform. I really think it may improve tensile strength, but pull tests will tell.

    I'm curious if you get a different surface finish. The matte finish filaments are really nice, it would be very cool to be able to produce the same effect from a nozzle shape.

    there is a lot to learn.

    So much so! I'm very interested in what you can teach us with this experiment!



  • I've finally had time to sample this nozzle a bit. A few initial considerations:

    • other than swapping nozzles, printer hardware and settings remained the same, except the Z offset
    • photos were taken under the same lighting conditions and with the same settings
    • there are other minor print quality issues that I am aware of (obvious in the pictures), but did not affect the goal of the experiment and were thus ignored
    • these tests were certainly not exhaustive and could be improved upon with more time
    • 250* (PETG), 60mm/s with no reductions or overrides, 25% infill
    • Part size at base is 50mm

    First impressions: the square doesn't make much of a difference, but the EDM finish does. Whether or not it would be considered significant is purely subjective, but the change is objectively there. It appears to do the following, to some degree:

    • hide ringing and other imperfections
    • slightly more matte finish

    I will be performing additional testing with a coarser EDM finished round nozzle to confirm the improvements were not orifice shape related. Further, I'd like to quantify the change in surface finish. Can anyone point me in a helpful direction?

    Next: because the majority of the nozzle was predrilled, I believe a larger EDM finished surface area may further improve results. After all, the total land for the major orifice ID is only .60mm. It will be a bit, but these results will also be shared. A star shape nozzle is still on the shortlist for the existing nozzle blanks.

    Other observation: the square nozzle did allow me to more easily print higher layer heights. The single attached picture was printed at .40mm layer heights. I did not put much effort into it, but the .40mm round nozzle could not complete a print under the same settings/conditions. This may be attributed to the larger surface area of the square vs. circle.

    Please see the picture (dropbox link at the bottom) and share your thoughts. The image is quite large (source quality) for better assessment.

    Additionally, I would love to share a square nozzle with someone qualified to spend some time experimenting with it, should anyone be interested. In exchange for said testing, the nozzle would not have to be returned.

    Please let me know if you have any questions, I'd be happy to elaborate on anything.

    Full Comparison Photo
    https://www.dropbox.com/s/h90btw6l11z34y0/Finished.jpg?dl=0

    .40 Layer Height
    https://www.dropbox.com/s/kk4213mvsp8ytfe/01.jpg?dl=0

    EDIT: For clarity and to add certain print settings



  • I am sorry to say that I can only see a very minor difference in quality (which is exceptional in both prints). Given the cost and effort to produce the result, it doesn't seem worthwhile.
    I wonder if the nozzle might be more effective in less perfect printers ....



  • No apologies necessary, this is why I shared it. For me it's somewhere more in between, but I may be biased. At this point I'm mostly looking for any change or improvement. And I appreciate the comment on print quality. I would definitely like to try these on some off-the-shelf entry level printers; good idea.


  • Moderator

    Maybe on shorter layer heights it would make more of a difference. I suspect that the filament will want to return to a rounded state as soon as it's able to.

    What does the extrudate look like in free air? Does it retain a square shape or does it return to a round shape?



  • @Phaedrux Surprisingly it retains a square shape. I'll see if I can find some purge to take a picture of.



  • Not a very good photo, but you can definitely see the square shape. This is a photo from the first day of testing. If I put it back on I will cut some with a razor to better show its profile.

    Capture.PNG



  • @plasticfactory said in Square Nozzle Orifice:

    First impressions: the square doesn't make much of a difference, but the EDM finish does. Whether or not it would be considered significant is purely subjective, but the change is objectively there. It appears to do the following, to some degree:

    • hide ringing and other imperfections
    • slightly more matte finish

    Fascinating.
    How does the internal geometry differ on the two nozzles? With the carbide nozzles I've used before there was no throat, the internal faces tapered all the way to the actual opening. We never did real testing to compare to our standard steel nozzles, but there was a noticeable improvement in quality.

    I will be performing additional testing with a coarser EDM finished round nozzle to confirm the improvements were not orifice shape related. Further, I'd like to quantify the change in surface finish. Can anyone point me in a helpful direction?

    Ooo. That might be tough, most surface roughness measurement tools are going to pull out the layers just as much as anything else. But I'm not a metrology engineer...

    Next: because the majority of the nozzle was predrilled, I believe a larger EDM finished surface area may further improve results. After all, the total land for the major orifice ID is only .60mm. It will be a bit, but these results will also be shared. A star shape nozzle is still on the shortlist for the existing nozzle blanks.

    Other observation: the square nozzle did allow me to more easily print higher layer heights. The single attached picture was printed at .40mm layer heights. I did not put much effort into it, but the .40mm round nozzle could not complete a print under the same settings/conditions. This may be attributed to the larger surface area of the square vs. circle.

    Can you clarify what you mean by 'easily'?
    I'd love to get these nozzles on a load cell, as I assume what you are describing is lower extrusion force. Certainly surface area of the nozzle has a huge impact on extrusion force.
    The 0.4mm round nozzle is 0.126mm2, the square is 0.16mm2. The relationship between extrusion force and cross-sectional area of the nozzle isn't simple, but in this kind of range I'd expect it to be reasonably linear. So I would expect the square nozzle to require 25-30% less force for a given volumetric flow rate.

    It would be interesting if a square or otherwise non-round nozzle would let you extrude significantly narrower paths than the major diameter of the nozzle, but without a loss in quality. That might be the trick for getting down to silly size extrusions. Although it would probably need something other than sinker EDM to make it even close to economical.
    What extrusion widths were you using?

    Please see the picture (dropbox link at the bottom) and share your thoughts. The image is quite large (source quality) for better assessment.

    Additionally, I would love to share a square nozzle with someone qualified to spend some time experimenting with it, should anyone be interested. In exchange for said testing, the nozzle would not have to be returned.

    Please let me know if you have any questions, I'd be happy to elaborate on anything.

    Full Comparison Photo
    https://www.dropbox.com/s/h90btw6l11z34y0/Finished.jpg?dl=0

    .40 Layer Height
    https://www.dropbox.com/s/kk4213mvsp8ytfe/01.jpg?dl=0

    EDIT: For clarity and to add certain print settings

    I do find it very interesting that what I would have assumed was ringing from the X/Y stage, clearly must be some sort of ringing in the extruder.
    Could be laminar low, could be a masters thesis for the eggheads over at UC Berkeley.



  • @theruttmeister said in Square Nozzle Orifice:

    Fascinating.
    How does the internal geometry differ on the two nozzles? With the carbide nozzles I've used before there was no throat, the internal faces tapered all the way to the actual opening. We never did real testing to compare to our standard steel nozzles, but there was a noticeable improvement in quality.

    Internal geometry is identical to an E3d V6, but with a square orifice. Within 60 days we should be sampling nozzles without that major ID. Then we will be able to try different internal geometries. In your case, were you using the nominal ID, with a draft (cone shaped end to end)? Just want to make sure I'm understanding correctly.

    Can you clarify what you mean by 'easily'?
    I'd love to get these nozzles on a load cell, as I assume what you are describing is lower extrusion force. Certainly surface area of the nozzle has a huge impact on extrusion force.
    The 0.4mm round nozzle is 0.126mm2, the square is 0.16mm2. The relationship between extrusion force and cross-sectional area of the nozzle isn't simple, but in this kind of range I'd expect it to be reasonably linear. So I would expect the square nozzle to require 25-30% less force for a given volumetric flow rate.

    By easily I mean I changed no settings going from .20 to .40 other than the layer height.

    I don't currently have the ability to do that kind of testing. Though I'd certainly be happy to toss one of these to someone who could.

    It would be interesting if a square or otherwise non-round nozzle would let you extrude significantly narrower paths than the major diameter of the nozzle, but without a loss in quality. That might be the trick for getting down to silly size extrusions. Although it would probably need something other than sinker EDM to make it even close to economical.
    What extrusion widths were you using?

    I was thinking perhaps a trapezoidal shape could help things, almost like a paintbrush effect. Though I'm not sure where to begin on geometry, and because it would only have one "direction" it's not really feasible for me to do anything but make one. Perhaps it's time to start considering another axis, maybe something open source already exists here.

    I was printing .50mm width.

    I do find it very interesting that what I would have assumed was ringing from the X/Y stage, clearly must be some sort of ringing in the extruder.
    Could be laminar low, could be a masters thesis for the eggheads over at UC Berkeley.

    lol, I hadn't considered it would be ringing from the extruder. Based on my jerk/accel settings I'm still inclined to believe it's from the XY. It's throwing a Hemera around without much regard for anything at the moment.



  • @plasticfactory said in Square Nozzle Orifice:

    . In your case, were you using the nominal ID, with a draft (cone shaped end to end)? Just want to make sure I'm understanding correctly.

    Yes, the nozzle went from 2mm ID, down to the 0.4mm orifice as a single cone shape progression. I don't recall the angle, but probably 20 degrees. Actually quite reasonably priced!

    I don't currently have the ability to do that kind of testing. Though I'd certainly be happy to toss one of these to someone who could.

    I have a load cell... and a setup I was intending to use for benchmarking some extruders.
    It's not finished right now, but it wouldn't be too hard to finish it.
    I don't have any extruders that your nozzle would fit though. All the ones I have on hand are... custom, shall we say.

    I was thinking perhaps a trapezoidal shape could help things, almost like a paintbrush effect. Though I'm not sure where to begin on geometry, and because it would only have one "direction" it's not really feasible for me to do anything but make one. Perhaps it's time to start considering another axis, maybe something open source already exists here.

    Before that, I'd be interested to see the results of printing more organic shapes and general things not aligned with the nozzle orientation. Because:

    I was printing .50mm width.

    The major dimension of your square nozzle is ~0.56mm (diagonal). My interest is can you get the quality of a smaller round nozzle, but using a larger surface area and hence lower force, square nozzle.
    Can you try some prints at say, 0.45mm width?

    The paintbrush idea seems sound, but that brings back needing infinite rotation of the nozzle/hotend.
    Square is a good place to test ideas for now. A 4 pointed star might also be interesting.

    lol, I hadn't considered it would be ringing from the extruder. Based on my jerk/accel settings I'm still inclined to believe it's from the XY. It's throwing a Hemera around without much regard for anything at the moment.

    Its an axis like any other, and steppers act like a spring-mass. I see no reason why we shouldn't be thinking about the things discussed here:
    https://forum.duet3d.com/topic/20373/a-software-solution-to-eliminate-ringing
    For the extruder as well.
    My hypothesis being that the square nozzle, requiring less force, generates less slip in the motor and so what looks like ringing being eliminated is actually cogging of the extruder motor as it accelerates the filament (less load, less cogging).



  • @theruttmeister said in Square Nozzle Orifice:

    Before that, I'd be interested to see the results of printing more organic shapes and general things not aligned with the nozzle orientation. Because:

    I was printing .50mm width.

    The major dimension of your square nozzle is ~0.56mm (diagonal). My interest is can you get the quality of a smaller round nozzle, but using a larger surface area and hence lower force, square nozzle.
    Can you try some prints at say, 0.45mm width?

    I will print some more traditional parts this weekend. I see what you're saying now. I will try it with the part pointed "forward" and at 45*. I settled on going side to side because of the ideas presented in the original post, but I think you are onto something here.

    I could also make another one at .40mm for the major dim, but that seems counterproductive. A squircle would also be possible.

    lol, I hadn't considered it would be ringing from the extruder. Based on my jerk/accel settings I'm still inclined to believe it's from the XY. It's throwing a Hemera around without much regard for anything at the moment.

    Its an axis like any other, and steppers act like a spring-mass. I see no reason why we shouldn't be thinking about the things discussed here:
    https://forum.duet3d.com/topic/20373/a-software-solution-to-eliminate-ringing
    For the extruder as well.
    My hypothesis being that the square nozzle, requiring less force, generates less slip in the motor and so what looks like ringing being eliminated is actually cogging of the extruder motor as it accelerates the filament (less load, less cogging).

    This explanation makes a lot more sense to me now. It's definitely changed the way I'm thinking about this.


Log in to reply