Differences between types of extruders

Hi everyone!

I'm building a new printer, and now I'm choosing the extruder to mount.

I know that there are three types of extruders:
- Geared extruder
- Bowden extruder
- Direct drive

Anyone knows pros and cons of these?

What I think is:
- Geared extruder: very precise and fast with the gear system, but a bit bulky
- Bowden extruder: reduces the moving mass, but have some precision problems caused by hysteresis (can cause a big decrease of precision compared to geared extruder?)
- Direct drive: removes the bulky gear system, but still have a big moving mass, probably needs slower speed and works better with 1.75mm filament (not considering geared motors)

hi,

I also have the same questions,
someone with more experience could explain better?

Within the direct drive type which is the best model designed so far?


Thanks
neo

In my opinion the best is hands down a bowden extruder. The acceleration of the machine has a direct effect on the surface finish and higher acceleration gives a smoother surface. But I see no reason to use one on a machine with a bed that moves in X or Y. The only machines that benefit from them are machines that can actually accelerate in X and Y really fast without introducing resonance every time it changes direction. Like in this picture [d2oadd98wnjs7n.cloudfront.net]

---

FFF Settings Calculator	Gcode post processors	Geometric Object Deposition Tool Blog
Tantillus.org	Mini Printable Lathe	How NOT to install a Pololu driver

What about retraction problems introduced with the long bowden cable?
And the mass of the motor reduces so much the maximum speed and acceleration?

Edited 1 time(s). Last edit at 03/18/2013 03:17PM by Pacca.

I think it is the mass and the springiness of the motor acting as a magnetic spring causing the issue of resonance. The retraction issue is not a much of problem with some hotends as it is with others. What I found was we want a high back pressure hotend so when we retract it stops extruding very quickly. This means a long exit hole and medium length melt zone. Amazingly enough though these are the characteristics of a properly made J-head so they are readily available. If you click on the how not to install a pololu driver in my signature and look at the videos on my youtube page you will see all kinds of prints done on machines with Bowden extruders.

---

FFF Settings Calculator	Gcode post processors	Geometric Object Deposition Tool Blog
Tantillus.org	Mini Printable Lathe	How NOT to install a Pololu driver

I cannot understand very much this resonance issue you are talking about...
I thought that the problem of the mass of the extruder was linked to the power of the motor of the X axis (considering a common prusa mendel) and to the stiffness of the structure of the printer

When you try and change direction all that mass keeps trying to move in the original direction. This cause the axis to go past the actual location you want to be at and because the motor acts like a spring it goes back and forth a little which cause the wiggles or ghosts you see in the picture I posted above. This problem gets worse and worse the faster you go. But with a low mass moving in all directions you do not get that effect.

The problem you mention looks like a little bit of Z wobble when it occurs and can be overcome with some frame stiffeners so no real need to use a bowden extruder to stop it.

---

FFF Settings Calculator	Gcode post processors	Geometric Object Deposition Tool Blog
Tantillus.org	Mini Printable Lathe	How NOT to install a Pololu driver

Oh ok, you are talking about the problems caused by the inertia of the moving mass, sure, what I meant too!
I don't really want to change the extruder I'm using, I was only thinking for a new printer

Sublime Wrote:
-------------------------------------------------------
> The retraction issue is not a
> much of problem with some hotends as it is with
> others. What I found was we want a high back
> pressure hotend so when we retract it stops
> extruding very quickly. This means a long exit
> hole and medium length melt zone. Amazingly enough
> though these are the characteristics of a properly
> made J-head so they are readily available.

Really? I have been running a 0.4 J-Head MkV-b and it has much less back pressure than 0.35 Arcol V4 (the difference in force is very obvious when pushing filament by hand). The difference is apparent in the output too, the Arcol gives sharper, more accurate output with just half the retraction compared to the J-Head. And you say that the J-Head has high back pressure...?

I did not say all others were lower back pressure, just that a real j-head has high back pressure. If you look at the measurements [jheadnozzle.blogspot.ca] of the length of the exit hole on a real j-head from hotends.com compared to a budasnozzle or makergear you will see it is longer then them which results in higher pressure. As well as a real j-head is also drilled a smaller diameter then rated resulting in higher pressure then its rivals of the same rated size. I have tried hotends with exit holes that were short like the buda and they do not work well for Bowden cable extruders.

---

FFF Settings Calculator	Gcode post processors	Geometric Object Deposition Tool Blog
Tantillus.org	Mini Printable Lathe	How NOT to install a Pololu driver

That's interesting, I thought the J-Head had particularly low back pressure, since it's lower than the others I have used (Arcol V4 and my late Mendelparts V9). The Arcol must have a really long orifice length or short melt zone then. The maximum output per second is much less than with the J-Head, but on the positive side, it also has quite a bit more accurate output than the J-Head Mk-Vb:s I have been using (originals from Reifsnyder Precision Works).

There are also other odd differences. For some reason the Arcol's overhangs curl up a lot more than J-Head's, even with everything else being the same (including the specific roll of plastic). Maybe something to do with ABS's elastic properties after being pushed through a small hole?

The feed resistance has two components. One is imposed by the nozzle. If this is high, but you can still extrude consistently, then that is a good thing. The other component is the resistance to feeding the plastic due to the glass transition plug. This will vary quite a bit with the extruder geometry and thermal gradient in the transition zone and may also not be very consistent or predictable. This component of the feed resistance should be minimised for good prints.

Edited 1 time(s). Last edit at 03/23/2013 09:06AM by Greg Frost.

I can't found anywhere a quantitative answer to my problem: how much does the weight of the motor really affects on the print speed and movement?

This is the difference of the light weight carriage on Tantillus vs my Prusa.

---

FFF Settings Calculator	Gcode post processors	Geometric Object Deposition Tool Blog
Tantillus.org	Mini Printable Lathe	How NOT to install a Pololu driver

Thanks very much!
The weight of an extruder motor (nema 17) is approximately 0.35 kg. I will have the same problems with a light weight motor like this one that weights only 185 g?

That would make the X axis nice and light but you will still end up with some resonance from the bed weighing a lot.

---

FFF Settings Calculator	Gcode post processors	Geometric Object Deposition Tool Blog
Tantillus.org	Mini Printable Lathe	How NOT to install a Pololu driver

Now I'm thinking to build a coreXY printer. So the bed is fixed and the extruder moves in the xy plane.
I prefer to have a better control on extrusion and retraction, and that's why I'm trying to understand if also a lightweight motor can reduce very much the maximum speed and acceleration!

What are the pros and cons of CoreXY versus the Ulitmaker / Tantillus crossed bars scheme?

---

[www.hydraraptor.blogspot.com]

Personally I don't like very much the Ultimaker design because it has too many pulleys (CoreXY uses many pulleys too, but you can change them with bearings, I think) and belts (also two small continuous belts), and it needs two more bars to work.
Tantillus reduces the things that I don't like using fishing line and gears to connect motors with axes.

Don't forget about Delta geometry it is quickly gaining popularity and the ones that work well work really well.

Also do note the issues my picture shows is extreme. My Prusa was a year old at the time and the speed of 45mm/s was chosen above the speeds I know did not leave the ghosts on purpose to show the effect of going fast with a heavy carriage and bed. if you keep a Mendel based machine below 30mm/s even with a threaded rod based machine you do not really have these issues. I would also imagine that machines like Nopheads Mendel 90 are far less prone to it due to the added strength of the frame and better belts.

---

FFF Settings Calculator	Gcode post processors	Geometric Object Deposition Tool Blog
Tantillus.org	Mini Printable Lathe	How NOT to install a Pololu driver

I like the Delta geometry because it doesn't need Z threaded rod, but to have a big printing surface you need a very tall machine (with problems of stability and frame stiffness)

@Pacca,

The problem with the motor that you linked is that they're rather cheaply built, with sintered bronze gears, and have 2 failure modes.

The first is that if you use the same acceleration settings that you would use on a Nema 17 motor, one of the small bronze gears will break when retracting or replacing filament.

The second (and IMO, worse) is that in some of the motors, the interference fit of the main (steel) drive gear is poor, and the gear can simply spin off of the drive shaft under stress.

- Ron