New light extruder concept out of bolt

Hello guys

I wonder if some one has ever tested an extruder that instead of some gear rolling on the filament, is like threading the filament with a rotating bolt.

Advantages are :
- the whole perimeter of the filament is used to pull the filament.
- the longer the thread the more the filament can be pulled
- easy to insert : just make the bolt turning to screw
- easy to remove : just make the bolt turning to unscrew
- most convenient way to embeded an extruder on delta printer like kossel or rostock...

Problems are :
- as the bolt is turning and as the hot end is not a joint must be in between. I seek an oring : DF801 material type can stand 250°C : that is clearly enough compared to 200°C ABS.
- with a 1.75mm filament and a M2 thread, motor (1.8°/step) should turn at about 4000rpm for 500mm/s printing speed and 0.4mm hot end.
- cannot switch from 1.75mm to 3mm. 3mm filament seems not possible due to limited nema shaft diameter

I made a test with 1.75mm filament and M2 bolt with my screw gun : it works very fine ! It would pull harder with a longer bolt

So :
- Has anyone tested that concept before ?
- What do you think of 4000rpm ? (I 've never tried : any rex would be appreciated)


Thank you for your attention.

Emmanuel

I can't see how one can drive the bolt with the filament right next to it
I suppose the filament could bend into the drive well below long screw
under drive gear.

This would be a "worm drive"
One may want a worm gear set with filament between?

I think 4000rpm would shoot out filament in a great volume?
Let's say you used a M10 threaded bolt
1 rev would advance filament about 1.0mm inch
4000rpm 67 rps so 67mm/sec of filament

Probably sonic boom of filament flying out may be dangerous

I think you may have to do some recrackulating??

That is how the original RepRap Darwin extruder was made. [reprap.org]


It is also how the BFB touch (and earlier BFB's) extruders were built until BFB was rebranded as CubeX.


I believe they had issues with retraction and speed but I never built or used one so I don't really know.

---

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

Thanks guys

The idea is to have Nema 14 motor with its shaft threaded (making it a special motor). So the motor will itself be the extruder.

So the thread as to be about the same size as the filament. Luckily M2 is near 1.75mm (M3 is to close to D3mm). M3.5 (that exist) as a step of 0.6mm so only 0.1mm to grip : not enough.
A M2 as a step of 0.4mm that's why it needs to rotate quite fast depending of the print speed. May be 500mm/s is very fast and many users do this. About the calculation : 4000rpm/60=66rps * 0.4mm = 26mm/s @ 1.75mm so it makes 500mm/s @ D0.4

A M3 has another pb : the D5 shaft is to tiny to support a M3. With a M2 it should be limit but OK

Joined there is a section of the motor.

According to the motor torque and M2 step it seems that it could theorically pull about 500N !

The D2 helps to guide the filament inside the shaft before going to the thread.

It seems that none of this has ever been done, right ?

similar concept
[forums.reprap.org]

---

Random Precision

I don't think you will get a stepper to run at 4000 rpm with any usable torque. Stepper torque drops off rapidly with speed.

I have used the BFB screw drive and it is not great, the extruder is big and quite slow. The limiting factor is the strength of the plastic, so you can just end up cutting a big groove in the filament. Having a longer contact area probably helps, the problem is getting a the screw next to the filament and being able to drive it. The BFB extruder has the screw at an angle.
I think the screw drive was probably abandoned for good reasons.

I think having multiple pinch wheels and driven idlers would be an avenue to look at, if I had any sort of mechanical skill!

I dealt with a nema motor producer. He can produce motor with hollow shaft D3. Meaning a shaft with a D3 hole on its whole length.

At the end of the shaft I glue or fix a special M2 or M3 bolt depending of the filament size (See joined pdf). So the idea is to move the filament through the shaft. There is no side gear or side driving making it, I agree, a mechanical "non sens". The motor is directly driving the filament.

The nema 11 torque at 1000rpm is 0.06Nm that makes a 940N pull/push force for M2. I have calculated that Qubd extruder stall push force is only 50N. So I have at least about 20 times over torque to reach 4000rpm for D1.75 filament.

For D3 filament the rotor speed would be 2000rpm for 1m/s printing (who print that speed ?)

1000, 2000 or even 4000 rpm also means the ability to have very precise extrusion flow, because we have wide range to adjust the speed. With a 100rpm we don't have that wide range.

I am placing an order for these motors to test that and will tell what's happening.

Think of it : its a nema 11 motor ! only 28mm over all square size ! 110g=3.9oz ! and no gears, no lever, no spring. A nema 17 alone is double weight.

@erylon

Heat exchanger:
It appears you have drawn it as a thread?
Add an undercut at the base so the cutting tool can be retracted.
Or make each rib independent.

Filament twisting:
I envision the filament twisting ahead of the drive nut.
This bound up energy will cause problems.
I think you need a means of holding the filament straight as the drive nut draws the filament into the hot end.

Drive nut thread profile:
A round crest profile would aid in reducing particulate generation.
I'm concerned with the drive nut jamming up due to stripped filament caused from the drive nut.

Extruder nozzle:
You have it drawn as one piece.
Nozzles, wear, clog, and different materials need specific orifice diameters.

Heater:
Heater cartridge location, where is it?

Water cooling:
Consider adding room for future enhancement for water channels.
This will allow you to print challenging materials.

Cost:
How much are they charging, and what OEM are you working with?
Less your an engineer using the cover of your business, I'm curious how you're getting this company to mod the shaft for you

Over all I like the concept.

A2

Thank you A2 for these comments. Here are my answers :

Heat exchanger
Well I did not really work on the nozzle. I just wanted to show what the interface might look like. I am glad you comment that. To be that as it may, the heat exchanger is not a thread but layers of discs, each disc acting as a exchanger. Following your advice I would like to add an undercut but I don't know what you mean or what it could look like. Would you please give an example ?

Filament twisting
You are right : as the motor applies a torque, thanks to Newton, the same torque will appear on the filament between the drive nut and the spool. As the spool is rotating around a fixed axle and as the filament act as a spring, it will twist but a little as the torque is low. The twist, if any, will not increase in the nozzle and will remain constant after the drive nut (unless the spool is rewinded & winded several time in the printer but normally this is not done). During my experimentations with a gun screw (huge available torque compared to nema 11) I did not notice any problems. Even if it would turn in the nozzle I guess that nothing would happen in the hot end as it would be mixed once melted.

Drive nut thread profile
I am sorry but I don t understand what you mean by a 'crest profile'. (English is not my first language). Would you please add comments ? What I can tell is that the holes diameter before and after the drive nut are a little bigger than the filament size. About .1 or .2mm should be enough. I ll investigate that with my drive nuts.

Extruder nozzle
I only put .4mm as I have seem with new stainless steel prusa design. I ll investigate that. (The purpose of the design was about the drive nut / motor but thank you for this comment)

heater
The heater is a removable part as in prusa design not shown here.

water cooling
Not many place but I ll investigate that. May be the drive nut could be use as a pump with another design. A seal should then be added. Why not. Still an exchanger as to be placed somewhere. If the nozzle was designed in a another way like thermo bottles technology used in my previous job it could be passive. This means about 1mm by degree so about 200mm length folded like accordions. I should ask some friends of mine.

cost
It s about 35$ each for a sample batch of 6 motors. I just asked them ! Pb is the delay : about 4 weeks.

I was looking into that idea, with the hollow-shaft stepper. I'll let you try that... in the mean time, I'm gonna try using a regular old hex nut mounted atop my extruder nozzle (I'm using a modified Prussa-Mendel), and use herringbone gears to transfer the motion over. Then, I'll play with the size of the gears to find a good RPM.

I machined with my lathe a hollow shaft with a thread inside with same sizes of the nema I am waiting for. I am currently out of filament : it should arrive in the coming days. As soon as I get it, I will post a video on youtube showing it working with a screw gun.

@wissing
To know the good rpm is easy : it is a flow calculus. It depends on :
s : printing speed (mm/s)
T : thread step (mm/revolution)
H : diameter of the hot end hole (mm)
D : diameter of the filament (mm)
R : speed of the nut ( revolution per sec)
So the filament flow that flows in the hot end is the same that flows out :
T.R.D^2=s.H^2 that is a mm3/s equation

erylon Wrote:
-------------------------------------------------------
> The nema 11 torque at 1000rpm is 0.06Nm that makes
> a 940N pull/push force for M2. I have calculated
> that Qubd extruder stall push force is only 50N.
> So I have at least about 20 times over torque to
> reach 4000rpm for D1.75 filament.

ok, I don't know how you derived that, but I'll assume you know what you are doing.

> 1000, 2000 or even 4000 rpm also means the ability
> to have very precise extrusion flow, because we
> have wide range to adjust the speed. With a 100rpm
> we don't have that wide range.

Of course, the other problem with high RPM is that the step rate becomes too high for the AVR Arduino to generate.

It will be great if it works. The proof is in the pudding, as they say.

First results :
I turned several bolts with my lathe for 1.75mm and 3mm filament. The D3mm has recurrent problems : as I am making a thread in the filament when turning the bolt, the PLA must creep and as it is a M3, there is no free space to go. So quite quickly the PLA breaks. And then there is another problem : when the PLA is inside the bolt, its very hard to remove. The easiest way is to hot the bolt with a welding torch to evaporate the PLA !!!! It 's absolutely not a solution inreal 3D life, isn't it ?
With the 1.75mm, there is no free space problem as it is a M2 but when I did the thread on the bold, there is to less thread remaining to grip the filament : so in many cases the filament slides in the bolt.

In conclusion, from my point of view the idea is not bad but it is absolutely not a good way to do it.
To sum up these are the requisites :
- step must be larger in order to reduce motor speed : about 10 times more a M2 or M3 thread
- do not thread the filament because if it breaks it is not removable. Instead rub it with a thread
- the thread diameter must be adjsutable...
A lot of people would stop here. How to make a thread diameter adjustable with an unexisting thread step ?
Well : just use a spring !
Instead of a bolt, an outer spring must rub the filament. You can adjsut the rub in winding on itself the spring so that the inside diameter will be reduced. I made some tests with a ~3 diameter of a compression spring (no the one that should be taken) and with one self turn I get 0.2mm of diameter reduction. I am not sure to make myself clear : to decrease inside diameter of the spring, we must fix a end of the srping and rotate on the spring axis the opposite end. This will make an axial torque.
Advantage of this is that it stays simple, we can choose the spring step and it's adjustable. The disadvantage is that it doesn't give a direct drive, just a rub drive.
So I'll try that and will give you the results once I get my springs.
I will appreciate any suggestion or comment on that. Thank you.

hi,
i'm testing it.
stay tuned

dApA extruder v2 soon on Thingiverse

Enrico

Inspired by [www.thingiverse.com]

Edited 1 time(s). Last edit at 11/28/2013 06:35AM by enricodare.

Looks interesting, good luck with your trials!

Ideas:
If you had two diagonal screws on either side of the filament it would balance the torsional force.

A globoid worm (double enveloping worm) would further increase the contact area.

Happy new year to all !

I gave up with the spring because it doesn't not allow regular and controled speed.

I updated my design to that :
.
Notice that the thread for the hot end is not designed.
The key has a pull spring to pull it back (The clever mounting seems to work fine here). The pressure against the special screw can be adjusted with the screws pressing the key.
The filament is in between : a rod (black on the picture), the screw and the Key : 3 linear contacts at ~120°, with one (key) adjustable.
The screw is slightly curved at the begining of the thread.
If the filament breaks, it's easy to remove : it can't be stucked inside.
Looks promising, no ?

2 screws counter rolling would have been better to avoid torque in the filament but price would be double and there is not much place for that. I wonder if the torque issue could be solved adding a friction before the motor.

I would appreciate comments either positive or negative !

I had no idea you could get hollow shafts. The only thing I would consider changing (consider not necessarily change) would be to replace the key guide with a grooved bearing that can be spring loaded. This could also get you a quick access port to help string the filament.

However, your design may have more than enough force that friction from the keyway is no big deal and you may never have to open it up to get the filament through.

At the very least filament diameters change through a batch so I would rather set a constant force applied to the filament instead of a constant gap. This will give you a constant feed and constant deformation regardless of filament variation. For a direct drive it doesn't matter but for a bowden it would be nice to tune this so that there is a minimum of artifacts left on the filament while still having enough pushing force. Pushing artifacts add to friction in the bowden tubes. If they get bad enough they can bind you up completely.

Sorry, I miss typed the motor ref : it's Nema 14 with hollow shaft. You can get them on ebay.
They look like that:


@Nicholas: Thank you for your feed back. Well, I choose not to place a groove bearing because it would lead to a contact spot instead of a linear contact as with the key. Nevertheless, you are right: I should apply a constant force instead of a constant gap. So I updated my design with your recommandation. I placed 3 ball plungers :
-2 to press the key with .8mm movable ball to adapt the changes in diameter (of course the screw position is adjustable),
-and one ball plunger roller that is designed to be able to roll continuously. The purpose of this plunger is to give friction to compensate torque created by the screw so that there would not be any torque on the filament geting through the motor.

So the updated design comes to:

People interested should have a look at Werner Berry similar test.

What is stopping the filament from rotating?

Might be cheaper to modify a 10 lead Acme lead screw.
I would use a Dia 1.00" lead screw.
This would allow you to pass 608 bearings into the bore of the housing for the rotating shaft.
Using a sintered bushing might work, but it's a poor practice.

Dual pinch worm drive
The Kraken Dual Pinch Wheel Extruder - How it Works
[www.youtube.com]

Edited 2 time(s). Last edit at 01/07/2014 07:12AM by A2.

Thank you A2 for your comments. On the proposed design, 105 and 115 series bearing are used. The filament is stopped rotating, I should say the torque on filament before the motor is largely reduced, by the M6 anti torque ball plunger roller. You can see them better on the section view:
Section.0.2.pdf

Kraken is very interesting. Thank you for this information. He could improve his design by removing one motor, using a longer worm (guided) and a reverse Pinch wheel. I don't undestand how he manages to adjust the pinch since it seems that the wheels are guided so not movable.

On my design there is only one motor.

Do you like this?

[www.youtube.com]

---

Enrico

[www.dapa3dservice.it] [www.studiodapa.it]
repstrap dApA, Ramps 1.4, MarlinKimbra / scanner 3D Cubify Sense / Formlabs Form 2

@enricodare:
I like your design

The way you have the worm captured between the bearings looks like it eliminates all backlash, is this correct?

A herringbone gear would be an incremental improvement to eliminate backlash.

It is interesting to note that you are using a larger gear on the stepper motor, that is the opposite of most designs.
What is the pitch of the worm (thread)?
Any issues of filament stripping?

You cleaned off some filament dust on the extruder, would a finer pitch worm (80 pitch thread) reduce the shredding?

Fine Screw Adjusters
[catalog.monroeengineering.com]

3:09 min into the video, particulate was cleared off the extruder:
[www.youtube.com]

Do you think this design is extrudes more accurately?

If you were to design another one what changes would you make?

Edited 1 time(s). Last edit at 01/16/2014 08:54PM by A2.

I am prototyping a nut drive extruder and should be able to test later this week if the bearings arrive. My design uses counter-rotating nuts to cancel the torque that tries to twist the filament when a single nut turns on the filament. See attached images.

A 6-32 nut will twist onto a piece of 3mm filament with finger torque but will not allow the filament to slip. With 6-32 threads, the motor will rotate about 1.26 times to move the filament 1mm. I am hoping a NEMA-17 motor has sufficient torque to drive two nuts on the filament. I'll know soon enough.

This image is a 6-32 nut twisted onto a piece of 3mm ABS filament. The "threads" are barely marking the surface, yet it grips tightly. I suspect the plastic conforms to the nut then springs back after the nut passes a given section of filament.



I drew the thing in Sketchup (my CAD skillz aren't quite where they should be by now) and printed it.



These are the printed parts for the prototype. There will be two gears on brass tubes with bearings mounted in the blocks. There will be a 6-32 nut soldered to the end of one tube and a left-hand threaded nut soldered to one end of the other tube. Since the threads are opposite and the rotation of the tubes is opposite, the twisting torque will be canceled and the filament motion will be in one direction through the two tubes. I have all parts except the bearings right now, including the left-hand tap to make one of the nuts. I'll try to put the nuts in the middle, between the gears, so that getting the filament loaded will be easier (when only one nut is on the filament, it's going to try to twist).



If it looks like it's going to work, I'll invest in some real gears and do it right. I may have to embiggen the holes in the nuts or design split nuts with spring clamps to allow for variations in filament diameter/roundness.

The cheesy plastic gears press fit onto the 5mm motor shaft and brass tubes so there's always a possibility of them slipping, but they seem to hang on a lot tighter than the torque required to twist a nut onto the filament so I'm not too concerned about it.

Great !
I agree with the fact that there will not be any torque in between the two nuts, but there should still be a torque outside of the nuts. (The torque is between the nut and the whole filament not considering somewhere in paritcular) As there is a counter nut torque, there will be no resulting torque in between but that's all.
So you should see a torque before and after your extruder.
Please, keep us informed !

Yeah, it's a little hard to wrap one's brain around it. My thinking is that the torques applied to the filament by each nut gets applied to the whole filament, including beyond the other nut, therefore the two equal and opposite torques should result in a net zero torque on the filament. If the torques aren't balanced- say one nut is tighter on the filament than the other, there will be a nonzero net torque on the filament, so I may have to do some playing around with that to make it work. I think a small net torque on the filament is OK, especially with 3mm filament which is pretty stiff. It will twist just so much before the filament's springiness provides an equal and opposite counter torque to stop the twisting. I think it would be a bigger problem with 1.75 mm filament that might twist a lot with even a small net torque.

Progress: I fabricated the left-hand threaded nut today and soldered the nuts to two pieces of 5mm brass tubing. When the bearings arrive I'll mount them and test the whole thing.



The left nut (no jokes, please) is a normal, right hand threaded 6-32 nut. The right nut (oh come on!) is a left hand threaded 6-32 nut. The gears mesh well and alignment seems to be OK for testing purposes. I'll post some video when I test it.

Edited 2 time(s). Last edit at 01/26/2014 11:16PM by the_digital_dentist.

@the_digital_dentist :
That looks super promising! I am crossing my fingers for you.

BTW, we can do this with 4 spur gears and get the nuts closer together and compact the whole design. It might even be possible to use this for 1.75mm. I get ahead of myself. I eagerly await your results.

Is there any reason you didn't just buy the left hand nuts?
[www.mcmaster.com]

---

ConceptFORGE
Wally, GUS Simpson, LISA Simpson, THOR Simpson, Sextupteron, CoreXZ

Progress! The bearings arrived, I did a slight redesign of the base and cover to eliminate some impingement on the bearing races, printed new base and cover and assembled it. The gears mesh perfectly and turn smoothly with no bad behavior and minimal backlash. I ran a piece of filament through it manually and found there is a small and variable residual torque that tries to twist the filament coming into the first nut. I believe the variability in that torque has to do with imperfect centering of the nuts on the tubing. I think the torque arises mainly from the different threads on the two nuts- one nut is thicker than the other and the thread profiles are different- the left-hand threaded nut cuts the filament a little deeper than the other. I don't think the torque will be a problem because the filament will be coming off a spool and will not be free to twist. The spring force of the filament will balance out the torque very easily. I will make a new set of nuts with equal thickness and try to make equal thread profiles, then figure out a way to align them better with the centers of the tubes. Electrical tests tomorrow- gotta find out if the NEMA-17 motor I am using has sufficient torque to drive the gears and filament.

The filament loaded very easily. I had to hold it pretty tight to prevent twisting until the end of the filament found the second nut, which it did by itself, and after that the torque twisting the filament was very small.

I didn't buy left hand nuts because I anticipate having to make a few runs at this to get it right. A tap allows me to make nuts of any thickness and to vary the cutting depth (into the filament) by tapping different sized holes.

The nuts can be put arbitrarily close together simply by moving the tubing through the bearings and gear. There is no reason they can't be 1mm or less apart.

Since filament diameter may vary a bit, the torque required to twist the nut on it will probably vary. Some 6-32 nuts slip right over a piece of 3mm filament, some bite into it, depending on the hole size that was tapped. I am thinking that maybe a longer nut with a larger hole size will accommodate the variability of the filament diameter while still allowing a firm grip on the filament because more threads will be biting into it.

Edited 1 time(s). Last edit at 02/01/2014 11:21PM by the_digital_dentist.