threaded insert extruder filament drive

I have been experimenting with a threaded insert filament drive for an extruder. This uses a metal cylinder which has 4-40 threads on the inside (for my 2.75mm OD ABS filament). The cylinder is mounted in a gear, as the gear is turned the cylinder threads itself onto the filament and forces it through. My preliminary tests show a lot of promise. I need to get a high torque motor to keep testing as I am stalling my little DC one. I need to get a fish scale. The length of the threaded cylinder is the primary factor, shorter cylinders start to munge the threads cut into the plastic, longer cylinders require greater torque to turn. My 1 inch long cylinder seems to work but potentially something with ACME threads which are softer edged could be even better or maybe metric threads that cut into the body of the filament more. I am pretty sure this can be designed to that enough force is present to just stretch the filament of snap it).

Has anyone else played with this concept at all?

With a small hollow shaft motor (need at least 3mm ID, I have been looking anyone seen such a thing) this would make for a super simple and powerful extruder. Right now I am using 45 degree angle gears to avoid clearance issues with the motor body but a hollow shaft would eliminate all that.

Thanks.

Forrest Higgs was looking at 1/2" diameter nuts. Doing essentially the same thing, but with a large nut instead of a small threaded insert.

As I recall, one of the issues he had with it was twisting the filament. This was also something of an issue with the threaded-rod based extruders. There was even mention of getting a nut with a left-hand thread and using both to avoid twisting the filament.

Have you had issues with that? I'd think that with such a close match between filament O.D. and the I.D. of your 'nut' you'd have even worse problems. In fact, I'm somewhat surprised it even moved the filament, instead of just twisting it around and around. Do you have some method of keeping the filament from twisting?

---

--
I'm building it with Baling Wire

Currently I feed the filament through a little hole I have cut in a rubber block held down by a book to adjust resistance (like I said experimenting) but I haven't had any issues with the filament itself turning. I guess there will be a torque versus resistance consideration there. I have unwound the filament for a good length at this point to so I haven't dealt with unspooling issues. I would think that pre-winding filament with a bias against the turn of the drive could also resolve some of that.

Do you know where forest posted any information on his experiments?

I was thinking about this the other night. I imagined that you could put a blade vertically so it cuts into the filament and stops it from turning- worth a try.

---

-----------------------------------------------
Wooden Mendel
Teacup Firmware

goinreverse Wrote:
-------------------------------------------------------
>
> Do you know where forest posted any information on
> his experiments?


goinginreverse: Here you go...


[www.3dreplicators.com]

[www.3dreplicators.com]

[www.3dreplicators.com]

I put a lot of effort into getting one working and finally had to give up. It would work for a while and then perform less and less well.

I noticed that I could get it to work again for a while if I ran a tap through it to dress the edges of the threads in the nut. That led me to believe that the problem was that nuts were being made of a softer steel than screws and that it didn't take too many meters of filament running through one to dull it down to the point where it wouldn't cut into the filament for a proper grip any more.

That's a guess, mind. Good luck with getting the concept working. Just because I couldn't make it happen doesn't make it impossible.

---

-------------------------------------------------------

Hell, there are no rules here - we're trying to accomplish something.

Opportunity is missed by most people because it is dressed in overalls and looks like work.

Thomas A. Edison

From your pictures it is hard to tell, what was the 'length' of the nut from entry to exit?

I think the crucial difference is that I am using a threaded insert which has much longer traversal than a nut, the threaded area in my current experiment is 1/2 inch long but could be made as long as needed (1 or even 2 inches). There is dramatically more thread surface in contact with the filament so that the need to cut in much less.

Wear may be an issue but the inserts can be gotten in a number of materials including hardened and stainless still. Depending on the setup the inserts could be pretty easily changed out after X feet of filament.

Will let you know how it progresses. Thanks.

I wonder if there is a problem with trying to make the full cut in one turn. If you tapered the end and then tapped it again, it might not wear the threads down as much.

Or, you can try and find a pre-made threaded tapered cutting thing that is designed to be driven. [www.amazon.com]

---

---

Darwin clone, Gen 2 electronics, Arduino Duemilanove w/ AtMega328, 5D Firmware, Pinchwheel extruder
[www.codeerrors.com]

I tried a 3.5mm die on 3.1mm HDPE a long time ago, just by hand. Not surprisingly it shaved off a lot of material and produced swarf so would be very wasteful. A continuous thread should be better.

Screw drive in general wastes a lot of energy cutting the thread, produces dust, and tries to make the filament rotate. It works quite well on soft slippery plastics like HDPE and PCL, but is not good with very hard plastics like PCL.

The threaded pulley I use has nearly as much grip but makes the thread mark in the filament by pressure, rather than cutting it. That is more mechanically efficient and does not make dust. It does need a lot of torque at very low speed though.

---

[www.hydraraptor.blogspot.com]

Quote
goinreverse
With a small hollow shaft motor (need at least 3mm ID, I have been looking anyone seen such a thing) this would make for a super simple and powerful extruder. Right now I am using 45 degree angle gears to avoid clearance issues with the motor body but a hollow shaft would eliminate all that.

I've been taking CD-ROM drives apart, to see if any of the motors could be used a repstrapper, what you are looking for sounds more or less like the spindle motor, a BLDC motor, here a link to a guy who modified a motor for his RC plane, but I think the same principle could be used for an extruder motor.

I built this prototype using an internal screw approach. The bearings do a pretty good job of keeping the filament from rotating. Creating a mechanism to pre-thread the filament might make for an interesting experiment...

Rick

Well there you go, pretty much what I had in mind... Have you had any of the issues mentioned previously?

Could you make a post somewhere in detail, what parts were used, etc.

Looks like great work, you makerbot guys are totally schooling us.

Thanks, that's good information on motors but I am all about the standardized parts. If it isn't available off the shelf (or at least in chinese batch manufacture) I don't think it is a good fit for what I am trying to do. Commercially there are a lot of hollow shaft motors, they are just giant, such as 2HP size. That would probably be overkill

I played with dies a little bit, much to sharp and I saw the problems you described. If threaded pulleys were commercially off the shelf I would be all over that, I don't think it makes sense to require people to have a lathe to build an extruder though I think your R&D is still extremely valuable.

I can't speak for other plastics, at this point ABS is really all I care about in terms of material for its durability and cost parameters. There is some distortion but not really very much cutting, the threads in the insert are fairly dull and my filament is 2.75mm OD which leaves a bit of room against the 4-40 threads. It does not take much force at all to get the threading started, it is the friction that builds up over the length that requires the torque. I think acme threaded coupling nuts may also be a win because they have softer threads, more nudging than cutting. I think the insert length is the critical part, finding the right value that optimizes all the competing forces. I am not saying it is a magic extruder that will solve all the problems. My goal is a simple, cheap, easy, fast extruder which can be built without any printed parts, minimal fuss, adequate results. If it turns into more then great.

My final ace in the hole is the potential of some of the barrel heat to raise the temp of one end of a longish insert. That I think offers numerous advantageous qualities and brings us very close to industrial screw extruders. Long term I would like to move away from filament and see commercial resin pellets as feedstock. They are dramatically cheaper, right now we are buying filament which is just pre-extruded injection molding pellets and we are paying for the convenience.

My last post just started showing up, actually written this morning... weird.

I saw your detailed post rick, ignore my query for more info.

After reading the makerbot forums I think I may be contributing my efforts to the wrong project... that looks like division A to me and possibly a better fit for my perspective. Thanks Rick!

I posted the dxf on thingiverse...

Do you yourself have a laser cutter or access to one? If so what make/model, I have been looking to invest in one for work. I really appreciate your contribution, if you are ever in Phoenix ring me and I'll buy your a beer.

Forrest, I thought you'd given it up because the pinch wheel thing was working so well, not technical difficulties.

My first thought right after "it's getting dull" would have been "Let's heat-treat it."

---

--
I'm building it with Baling Wire

@goinreverse

Thanks!

For some reason I can't log into this forum with FF. I can log in fine with Chrome though.

Yes, I have a laser and access to lasers. I'll write something up...just need to figure out where to post it.

There is a lot of activity in the MakerBot Google Group but there is a lot more depth and history here. It will take a long time to read through the RepRap related blogs to figure out what has worked, what hasn't and what still needs to be tried...

Rick