Thoughts and comments of the various, intriguing shapes of 3D printer nozzles.
Great info, I would love to see some more extrusion tests, like you did to demonstrate the filament extruding back on itself, for the other features you show to get a visual idea of how they effect printing.
@James_Kao thanks! I should indeed add some “real world” pictures (will do for the filament extruding back on itself)… but I have so much stuff in the oven right now that it may take a while… 
Most of my conclusion come from “feelings” and years dealing with many different nozzles. I should change the parameters only one by one to make it more scientific, e.g. two overall identical nozzles but for the length of the terminal section.
I like your posting. But I miss some “deeper” information and various other nozzle shapes (e.g. merlin, fabbaloo or squeeze3d).
Yeah. This topic is quite interesting to me since the focus of comparison is most on extruders and the heat Chanel design of the hot end, but not much on the nozzle shape itself.
I think most of us have seen both good and bad behavior from nozzle shapes (both smoothing action and dragging action), but there hasn’t been much written characterizing the behaviors.
Would benefit a lot from a discussion of interior taper angle. Several people over the years have conclusively proven that a steeper drill taper between the full-bore section and throat decreases extrusion back-pressure and improves reliability. For example, E3D recently switched from two 120 degree bits in a stepped profile to a single steep taper drill bit. Makerbot made that switch years ago. (Molten polymers are viscoelastic and don’t like to change diameter and linear velocity fast.)
@Rene_Jurack yep I would rate this post high on my laziness scale (it was just a few hours night work, and I did not even fire the printer!)
I really wish I will spend the time for more –
I also did not talk about the extraordinary Dura-Gem for example (which I do not own) or the super buly diamond (which I own).
Hey manufacturers, feel free to send some weird stuff so I may get some motivation for a better review 
@Ryan_Carlyle I like it how you are knowledgeable each time! 
What link you would like best, as I copy/pasted your paragraph in the post (I hope it is OK with you).
The more I think about nozzles and the more I think a decent scientific workbench ought to be made…
I am unsure that (back)pressure is so much of a “modern” and serious issue as most of “serious” extruder/driving mechanism work reliably nowadays and they just create enough pressure in the heating block to properly fight the viscoelasticity of the material. This is another interesting topic and a blunt assertion I should check
@Jeremie_Francois Backpressure is an issue at higher flow rates with smaller nozzles, particularly when extruding fast enough that the filament doesn’t have enough residence time to get all the way up to the desired printing temp. You can tell by manually pushing filament through a hot end… at a slow rate it’s very easy, but try pushing faster and you’ll hit a wall fairly quickly. Better internal nozzle geometry means less force required at higher flow rates. So you can print faster without stalling or stripping the extruder, and will likely even get fewer cold-zone jams because there’s less melt pool pressure trying to push molten plastic back up the extruder.
Feel free to copy/paste/edit whatever you want from this thread, happy to help
@Ryan_Carlyle thanks! Shall I understand that the backpressure would be an issue that happens sooner than the maximum “geometrical” throughput? This would be much harder to test (well, I could compare a low-viscosity liquid to filament may be)
What do you mean by geometrical throughput?
@Ryan_Carlyle I guess it has a better name (not fluent enough, sorry!) – I meant the flow rate, limited by the nozzle diameter only (i.e. with an ideal fluid)… https://en.wikipedia.org/wiki/Hagen–Poiseuille_equation
The difference could be checked empirically by comparing the volume of plastic that can be extruded in a given time, to the volume of a very fluid liquid sent through the nozzle in the same given time.
I would (naively?) attribute the difference to the filament visco-elasticity, ignoring that of the liquid. The latter would then set an “ideal target”, in order to test different nozzle shapes and know how far we are from the limit (my guess is <<50%).
Now the main problem (for me) would still be to make those nozzles in the first place without a serious lathe.
Oh yeah, you could approximate it like that. It wouldn’t be rigorous because the viscosity of a viscoelastic fluid depends on shear rate and recent shear history, so it’s ambiguous what Newtonian viscosity you would compare it to. You’d have to just pick a number out of the range of possible viscosities.
To a good approximation, a filament-pushing extruder is a positive displacement pump, and plastic is incompressible. Until the hob/knurl loses efficiency due to pushing too hard for the grip strength and filament stiffness, volume in = volume out. So we wouldn’t get different quantities of flow when comparing Newtonian and viscoelastic fluids – we will get different resistance to flow, i.e. back-pressure. Which means more extruder force is required to achieve the commanded flow rate. Only near stripping and/or after stalling will the FLOW RATE change appreciably with viscous properties.
Now, if we had an extruder that applied constant pushing force, we would see rate variations.