Granule extruder/filament maker - some new thoughts

An introductory note: as it turns out, I'm currently building a RepRap as an official university project, with funding and all. It will be used primarily to build monolithic shells for minisumo-class robots, as a light, durable and cheap replacement for laser-cut aluminum, but other uses are surely going to emerge. One of the first things I thought of was contacting and engaging some fellow students from other fields, especially those related to polymer chemistry and polymer processing technology (I'm studying industrial automation and robotics), and maybe even get some research staff interested in the project. Now, it looks like that was a good idea.

Last friday, I discussed the granule hopper design with the members of one of the students' science clubs that deals with polymer processing. They pointed out all the flaws they could find in the devices tested by RepRap users and concluded that correcting them should produce a working, reliable device. I was able to take a close look at a few commercial injection moulding and continous extrusion machines, too, including a disassembled one.

The most important conclusion is that, while making a tabletop (read: the size of a RepRap or somewhat smaller), stationary extruder to produce a continuous filament suppy is definitely possible, fitting such a thing directly on a RepRap is out of question for all sensible intents and purposes and should not be attempted if the goal is having an efficient, practical rapid prototyping setup, and not doing something useless and overly complicated just for the heck of it.

The specifics, in no particular order, are:
- a drill, no matter what type, will not be a good auger, period. Even if some drills appear to be somewhat effective, that's more due to magic fairy dust and sunspots than dynamics of molten polymer. The augers in commercial machines are made the way they are for a reason - surely they wouldn't be that complex, were it possible to use something simpler. An effective auger probably can be made without any serious calculations, using just the "gut feeling" and some common sense, but its general shape must mimic the commercial designs. That is, the thread pitch gets smaller towards the tip, and the shaft diameter gets larger (in other words, the shaft should be somewhat conical.)
- a strand of thick, twisted copper wire soldered onto a copper or brass rod and turned down, as posted by someone on the forums, is probably the best bet for a homemade auger. The rod should be slightly tapered and the thread made with variable pitch, as noted above. Using a lathe is probably inavoidable for that, I'm afraid.
- the rear end of the auger, where the granules are fed, must be actively cooled to prevent the device from clogging up with half-molten granules. Without active cooling, even most commercial machines will clog up all the same and those that don't have special, expensive (because of their heat resistance) non-stick coatings on the auger and inside of the body, and some serious passive heatsinking integrated within their structure. One or two coolant channels passing through the body of the machine will do, with a pump (e.g. a garden pond fountain pump) and a heat exchanger (an old car radiator is OK). Plain tap water or mineral oil can be used for a coolant.
- the orientation of the auger should be horizontal, with the nozzle mounted at a right angle and pointed downwards. Such a configuration allows for an upwards-pointing vent to remove any trapped air, it will just travel up the vent, not down the nozzle. I'm not sure yet how exactly the vent is made to prevent it from leaking plastic upwards under such pressures, but my first guess is that there's always enough trapped air that constant venting will keep the plastic from going this route. I forgot to ask about that, I will on the next occasion.

For now that's all on the granule hopper. I'll forward any comments and questions, of course, and post replies or any further suggestions here.

It's a big undertaking. Good luck and keep us posted. :-)

---

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

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

Enleth Wrote:
-------------------------------------------------------
> The most important conclusion is that, while
> making a tabletop (read: the size of a RepRap or
> somewhat smaller), stationary extruder to produce
> a continuous filament suppy is definitely
> possible, fitting such a thing directly on a
> RepRap is out of question for all sensible intents
> and purposes and should not be attempted if the
> goal is having an efficient, practical rapid
> prototyping setup, and not doing something useless
> and overly complicated just for the heck of it.

Do not underestimate the inventive capabilities of someone who refuses to be limited by the state of the art.

If we keep trying to build one, keep failing, and keep improving the design, we may some day accomplish building the extruder which has been, herein, circumscribed.

I don't see the extruder as being nearly the bother to make as all the ancillary equipment like heaters and godet stations and the winder. I can't for the life of me how we will get all that on a tabletop, unless it is a really big table.

---

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

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

I'm in it for the heck of it :-)
For me, this is a fun project, very geeky, that fits in with my professional software training and my forgotten Mechanical Engineering degree. It's definitely a hobby activity, and I've got no illusions about that.

I don't expect this to be particularly efficient, fast, or practical - It's more interesting that I *can* do this at home. If I manage to produce some useful things out of it, then that's a bonus! I've learned a lot building mine and it's getting more interesting playing with extruders and drive systems. Selection of components is determined by what I can find, buy cheaply, or scrounge, and I don't have lots of time to make custom stuff, particularly for experimenting.

Of course, everyone's in it for a different reason. If the RepRap is a secondary goal (like to produce a paticular item, or against a time limit) then you're going to find it more frustrating when it goes wrong.

For myself, the experience of the journey is the reward, not the destination.
YMMV

Annirak Wrote:
> Do not underestimate the inventive capabilities of
> someone who refuses to be limited by the state of
> the art.

Quite often I'm using this phrase myself, but in this case, I'm not sure if its appropriate. Maybe for the sake of just doing something really hard and complicated - I understand and even symphatize with this approach - it is, but my current goal is to make - or at least help someone make - something that will just perform its job, and do it the easy and cheap way.

Making such a complicated device so much smaller is a tough job: you have to use stronger materials (expensive, hard to tool and machine) and more accurate tools (expensive, requires serious skills to use), some of the physical phenomena involved change (capillary action becomes significant, phase boundaries in molten material end up in different places relative to the device and to each other, heat transfer occurs at different rates, thermal expansion of the parts becomes a significant problem, trapped air bubbles become huge air cavities relative to the size of the melt chamber etc, etc), some ratios and sizes can't be scaled linearly no matter what - all kinds of problems await, none of them really understood in the particular context of a device that never before was made that small.

Maybe there is someone skillful and determined enough around here to do that, but I'm not this person and I'm not even going to try, when simply replicating something tried and working will do for the most part and should be in the range of my skills and, hopefully, resources. Should anyone actually attempt to use the information I provided in making an extruder, I'll be glad to have helped, of course.

As for the size of a filament maker, scaling an industrial extruder to a RepRap-like size is not a big deal, that's still roughly in the same order of magnitude and there already are machines just slightly bigger.

We're really glad to have you on board. Good luck with your RepRap and let us know if you run into any issues - we'll try and help.

If you have any more info on extrusion augers, then please post - photos are good - because it might help us in the right direction.
Thanks

---

---
Reprapping blog and other rants: [renoirsrants.blogspot.com]
My Reprap: [sites.google.com]

I, too, am thrilled to have someone onboard who is taking a such a rigorous approach to improving the designs. Good luck!

Enleth Wrote:
-------------------------------------------------------
> Making such a complicated device so much smaller
> is a tough job: you have to use stronger materials
> (expensive, hard to tool and machine) and more
> accurate tools (expensive, requires serious skills
> to use), some of the physical phenomena involved
> change (capillary action becomes significant,
> phase boundaries in molten material end up in
> different places relative to the device and to
> each other, heat transfer occurs at different
> rates, thermal expansion of the parts becomes a
> significant problem, trapped air bubbles become
> huge air cavities relative to the size of the melt
> chamber etc, etc), some ratios and sizes can't be
> scaled linearly no matter what - all kinds of
> problems await, none of them really understood in
> the particular context of a device that never
> before was made that small.

In some cases, e.g. miniaturizing a jet engine, things become much more difficult as you try to miniaturize a system. In others, e.g. microfluidics, making something smaller makes it easier to do.

I'm hoping that all it takes is a different approach to the matter. For instance, what would happen if we were to use something other than an auger as a drive?

Edited 1 time(s). Last edit at 04/07/2009 02:54PM by Annirak.

ive been working on a design that has some potential. its actualy using a industrial glue system's pump that im working on attaching to a melt tank. the Idea is to have the whole mass melted and hopefuly you can pump the liquid from the bottom (as we do with the hot glue) and run it through a extrusion/forming process. Its still very much a work in progress and by its self will not realy be practical as a whole for the reprap comunity (no access to the parts) but it could give ideas that could be build by many.

Ill post more when i have something actualy working (still building my repstrap and quite a bit of a procrastinator).

most of the possible pump configurations have been invented and exploited already.
A gear pump would be the most practical for generating the extrusion force, but it wouldn't have the bulk granule transport capacity unless either the gear pump was made way oversized, or the granules were caster sugar sized, instead of wheat sized.

Gear pumps have huge ineffeciencies when pumping viscous fluids

Did you ever figure out how they vent the air bubbles? I can't believe that if you had a vertical vent at the output that plastic wouldn't come shooting out, it's under such high pressure.

Greetings all,

Although a direct granule-to-part extruder for reprap would be very useful, I don't see that happening soon.

Enleth's notion of a separate, stationary device to make 3mm diam rod has merit, and I think is a more achievable goal. Toward that end, I'll pose a question:

Must a table-top rod-extruder use an auger?

An auger would be nice (among other things, to get a continuous rod), and I believe the commercial rod-extruders do use auger feed. However, an extruder could be made using a piston/cylinder -- or a couple piston/cylinders -- one set to reload while the other extrudes. Using pistons would probably lead to discontinuous rods, but can be spliced after extrusion.

If memory serves, Prof. Boyer did an experiment with a piston-based extruder, and (so long as the extrusion rate was low), the air between granules could be kept out of the melted polymer. I think a vertical piston, extruding downward will help exclude the air. Managing the transition zone (AKA keeping plastic flow and heat flow in balance), is probably critical.

Note: this ability to simply get the air out is an advantage of going slowly -- an advantage that mass-production injection moulding machines don't have. IMHO, we should take maximum advantage of this happy circumstance.

Such a device might lend itself toward a plastic recycler. I think a recycler will need a "packing screen" a fine metal mesh (properly supported), to strain bits of labels/crud out of the polymer -- to keep from clogging reprap extruder nozzles down the line. Packing screens are used in commercial injection moulding machines. (And at high flow rates, getting viscose plastic to flow through them takes extra power.)

---

Larry Pfeffer,

My blog about building repstrap Cerberus:
[repstrap-cerberus.blogspot.com]

Quote

Its actualy using a industrial glue system's pump that im working on attaching to a melt tank. the Idea is to have the whole mass melted and hopefuly you can pump the liquid from the bottom (as we do with the hot glue) and run it through a extrusion/forming process.

I had a quite similar idea with a melting tank too. The idea (not mine actualy) is to recycle bottles and other plastics. The problem with this like Enleth allready posted are the trapped air bubbles.

My Idea is to splitt the process into 4 steps.
1) cut or whatever the bottles to very small pieces
2) press them with a (heated ?) spindle in a melt-tank. Let the air move out till its filled enough,
3) open a small (3mm) hole at the bottom of the tank and let the the plastic flow out while still pressing the cutted plasitik into the tank. You have to cool it immidiatly after it with eg. CPU-coolers and roll it in a water-tank (for additional cooling).
4) => now you have a filament for a classic Mendel-Extruder


Its only an Idea ... don't kill me please

koallalays Wrote:
-------------------------------------------------------
> Its actualy using a industrial glue system's pump
> that im working on attaching to a melt tank. the
> Idea is to have the whole mass melted and hopefuly
> you can pump the liquid from the bottom (as we do
> with the hot glue) and run it through a
> extrusion/forming process.
> I had a quite similar idea with a melting tank
> too. The idea (not mine actualy) is to recycle
> bottles and other plastics. The problem with this
> like Enleth allready posted are the trapped air
> bubbles.
>
> My Idea is to splitt the process into 4 steps.
> 1) cut or whatever the bottles to very small
> pieces
> 2) press them with a (heated ?) spindle in a
> melt-tank. Let the air move out till its filled
> enough,
> 3) open a small (3mm) hole at the bottom of the
> tank and let the the plastic flow out while still
> pressing the cutted plasitik into the tank. You
> have to cool it immidiatly after it with eg.
> CPU-coolers and roll it in a water-tank (for
> additional cooling).
> 4) => now you have a filament for a classic
> Mendel-Extruder
>
>
> Its only an Idea ... don't kill me please


I woke up this morning with this exact idea playing around in my head.

I've only just ordered my first Mendel kit (lasercut from Botmill) and am anxiously awaiting its arrival, but already my mind is running wild with all the possibilities.

I'll be keeping an eye on this thread and hopefully once my RepRap is up and running I'll be able to begin contributing to this concept.

I love the idea of recycling your own plastic junk...