The extruder motor and changes needed in extruder robustness

Im thinking of trying Duratec 750 (calcium silicate) rod in place of the PTFE. RS components do it.
Machineable, high temp, rigid, thermal insulator.
ANyone tried it ?

I have been trying some tests with Macor which is similar i think, it looks very good but is extremely slow to machine and very delicate my concerns are all regarding its tensile strength and these concerns would also apply to Duratec as it only gives compressive figures?

---

Ian
[www.bitsfrombytes.com]

It needs to be quite strong otherwise it will snap when it meets a blob of something hard like ABS. At the moment the PTFE gives and the nozzle deflects when it hits blobs caused by extruder overrun.

---

[www.hydraraptor.blogspot.com]

Duratec 750 data sheet states a flextural strength of 23 MPa. ABS is 75 MPa, so it is in the same ballpark.

How about designing an extruder mounting system that is abuse tolerant? Eg, a mechanical fuse that is easy to replace, or even self resetting. even a small ammount of compliance in the mounting when a load thereshold is reached could save the day in many instances of head collisions that I have read about in various blogs. It would be an important feature of a non-specialist machine.

I cant wait to get my machine(s) up and crashing/running

Nophead - im loving your blog. great work.

I would just like to add my two cents.

There has been a lot of talk about putting heatsinks on the heater barrel in this and in other threads. Won't such designs require more power and heat the room significantly? Trying to heat one end of a metal pipe (even if it's stainless steel) and cool the other seems like a recipe for a high summer-cooling bill. If we can find another insulator (like PTFE, but higher temp) that has the desired characteristics, it would result in a much more elegant design, IMHO.

Andromodon Wrote:
-------------------------------------------------------
> I would just like to add my two cents.
>
> There has been a lot of talk about putting
> heatsinks on the heater barrel in this and in
> other threads. Won't such designs require more
> power and heat the room significantly? Trying to
> heat one end of a metal pipe (even if it's
> stainless steel) and cool the other seems like a
> recipe for a high summer-cooling bill. If we can
> find another insulator (like PTFE, but higher
> temp) that has the desired characteristics, it
> would result in a much more elegant design, IMHO.
>

Or construct some generator that can convert the energy from the heatsink into electrical energy that can be recycled back into the system. That'd be cool. It's possible, for sure, but I don't know how easy/useful it would be.


Anyway, on to the post I've been meaning to write on granule extruding (my ideas are all silly, naive, and whimsical, and as such I do not believe they deserve their own thread; plus, this is totally related):
Adrian Bowyer Wrote:
-------------------------------------------------------
> I am in the process of incrementing the extruder
> design. This makes it both easier to manufacture
> and more robust.
>...
> Of course, this may be the time to go for broke
> and switch to a granule extruder design.
>...
> So I'm starting this thread to invite thoughts and
> suggestions.

My first thought, of course, was to simply have two extruders: the standard extruder which accepts 3mm filament and extrudes through a tiny nozzle, and the auxiliary extruder, which accepts granules or other small plastic bits and extrudes the 3mm filament for the standard extruder. The standard extruder would keep its current design.

The aux. (3mm) extruder is essentially an inverted conical container with an opening at the top and bottom (it doesn't have to be perfectly conical; my first drawing was of something that switches between a cone shape and a cylinder shape, but almost any bottom-converging shape would work). The bottom of the container would be connected to a tube. A screw (or some other pumping mechanism, I'm not sure if you can just turn a screw in a cylinder to create pumping action) would be powered from the top (wider end) of the container, and would pump matter from the container through the tube on the bottom (if a screw would work, the screw would extend through a good part of the tube). The base of the container and the tube would be heated (possibly by a nichrome wire, as in the standard extruder). Granules or scrap from a grinder (the recycler) would be poured/fed into the container. The granules would melt and the fluid would be pumped into and through the tube.

The tube is connected to an insulator (like the PTFE tube on the standard extruder), and the fluid flows out of the tube and through this insulator.
The other side of the insulator is connected to another tube, and the fluid flows into this. Attached to this tube there is a cooling element, which cools the filament to solidify it (alternately, you could just run enough tube that the fluid is air-cooled to a solid filament).

At the end of the cooling tube are pinch-wheels (or toothed pinch-wheels for better grip, or perhaps just a drive screw) which pull the filament out of the tube. This is so that the fluid pump at the top doesn't have to provide the force to move the solidified filament out of the tube (which would complicate things).
The cooling and heating elements, I imagine, could possibly be combined into a single heat pump (Peltier element?) which removes thermal energy from the lower fluid and gives it to the upper fluid and granules, melting the granules and cooling them into the correct shape.

The insulator exists to prevent the cooled and heated parts from canceling each other out. It may or may not be necessary (possibly dependent upon tubing length).
The tubes and the hole through the insulator are, of course, 3mm in diameter. This creates a long 3mm barrel for the plastic to reform in.
It may be necessary to melt a certain amount of plastic in the upper container before pumping through the tubing, as otherwise there may not be enough material to form a solid filament.

Lastly, the filament coming out of the extruder can either be fed directly into the standard extruder, fed around a spool and then to the standard extruder (so that if the 3mm filament is extruded faster/slower than the standard extruder needs it, it can be stored temporarily), or fed to a spool for storage/later use.


The second thing I thought of was to connect the aux extruder directly to the standard extruder. That is, attach it so that the cooling tube from the 3mm extruder feeds directly into the drive screw area on the standard extruder. This would be a single extruder, without the mess from having two separate assemblies. It may have some problems with synchronization, but this can be dealt with using valves, etc.



Then, I thought, well, why have the 3mm filament feed through a drive screw just to be remelted at all? Why not simply keep it melted and focus it into a smaller diameter? Well, it did occur to me that this is precisely the idea of the granule extruder!
My current idea for the design is as follows:


On the top, you have the primary melt chamber, which is just a cup (cylinder with a closed bottom and open(?) top, with a heating element connected to it. The feed from the grinder (or the granule storage) is attached to the top (or granules are simply poured in), and the plastic melts. Simple. (You could have a fluid level sensor here if you're feeding automatically, so that you don't overflow.)

Connected on the bottom of this is the secondary melt chamber, in which there is a pump and an opening on the bottom. The bottom of this is connected to a tube, on the end of which is attached a conical nozzle (or perhaps part of the tube is just shaped into a hollow cone with a lathe for the outside and a countersink on the inside). The end of the nozzle is the aperture for extrusion.
A pump (like the screw pump in the aux. extruder) pumps material from the secondary melt chamber and through the tube, and the pressure in the tube pumps the material through the cone and out the aperture.
The secondary melt chamber and tube are both heated and maintained at the same temperature as each other (though possibly not the same as the primary melt chamber; not sure on this one, but for experimentation it would be a good idea to keep all 3 heating elements electrically separate and software controlled).

There is a valve on the bottom of the primary melt chamber, which can be opened (electronically) to allow fluid to flow (by gravity) into the secondary chamber. There are fluid level sensors in the secondary melt chamber. When the fluid drops below a certain level, the valve opens, and then shuts when the fluid reaches that level. The valve would either have to be small enough (and you may possibly need more than one valve) or have a filter, so that sizeable unmelted chunks would not pass through into the secondary chamber.
This essentially maintains a steady level of fluid in the secondary chamber, regardless (to a certain extent) of the conditions in the primary chamber.

As far as I know, the combination of steady pumping and steady fluid pressure (due to the steady level of fluid) would result in consistent extrusion.

And I really don't think it would have to be so heavy as some people think, given proper scaling. (In fact, the thought that the weight would be too much never crossed my mind.)


As for the recycler/grinder, I had a few ideas. First, there's the simple use a motor connected to a metal thing which spins and crushes the plastic, which other people mentioned.
Then, there's milling [en.wikipedia.org]. Particularly intriguing is grinding that does not produce significant wear on external parts (I originally heard about these on Modern Marvels on The History Channel, but I can't seem to find much information on these with light searching). Basically the material being crushed is impacted against itself, breaking itself into pieces. I think this is like the SAG (Semi-Autogenous) mill on that Wikipedia entry.
Now that I think about it, it's actually Autogenous grinding that I'm thinking of.

Anyway, I was hoping that with this type of grinding we could RP a grinder, grind with it, and if it wore out, we could RP another and grind up the original. Not sure if this would work.

The same idea applies to another wacky idea I had. I was wondering if we could somehow suspend iron filings in RP plastic (or coat the surface of an RP object in it), and use that as the grinder's grinding element. When the element wore out, we could melt it down, filter out the iron filings, and reproduce it with the same filings and plastic. I don't think it's realistic though.

Recently I had a rather interesting idea. It spawned from the aux. extruder idea. When I was thinking about grinding for it, I wondered why I needed to grind at all, if the pieces would just get melted anyway. The main reasons, I figured, were that some pieces would just be too big to fit in the container, as well as the fact that smaller pieces melt more quickly. I imagined taking a large piece and just shoving it on top of the melting chamber, which would be hot. The hot chamber edges would melt the plastic partially and cut into the object, and a circle (as the chamber is conical/cylindrical) would be cut out and fall into the chamber, leaving the rest of the piece behind (in my hands). The circle cut out would then proceed to melt fully.

So why not apply this to particle size reduction (which is what grinding is)? You could have a gridded looking square which you heat up with a filament and shove the piece into (or perhaps you place the piece on it and push a flat top down on it). It gets cut into small pieces by the heat, and you then have small particles. You may need to make the width of the lines on the gridded square wider so that when the slices come out the other end they don't just stick back to each other. You could also use two more squares, all perpendicular to each other (representing the three planes of our three dimensional world), and take the output from one, feed it through the other, and so on. I also thought about using multiple sizes of square grids, taking a piece through each, from large sized grid to small, making the pieces ever smaller, but when I thought about implementation, I figured that this would just melt the pieces back together (and just be like melting them through the smallest grid anyway). Then again, if you used cooling elements at the end of each grid to quickly cool the slices, then perhaps it would be easier to use multiple grids.
Anyway, these pieces could then be either fed directly into a melt chamber or ground up into smaller pieces or powder by another machine.
(Of course pieces that don't fit into this and are too large to shove into the grid to cut out pieces can be attacked by the Flame Saber 2000. Big hand-held (with an insulated grip) metal stick heated like the grids that can be used to cut up big objects, like your neighbor's car. Batteries not included.)



Anyway, to the people who don't understand the necessity (or simple desire for) recycling... well... if you make a bunch of prototypes, would you really want to throw all that plastic away after you're done with them (or if you screw up a lot)? I think not. Plus, it's just cool [to imagine recycling your own RP stuff and recycling your neighbor's plastic doghouse... muhaha].

Liquid plastics are more like bubble gum covered in glue than like water.
There are issues with getting the air out of melted plastic, it
must be forced out, gravity will not remove the bubbles. It's
possible that forcing the plastic through a nozzle might remove
some of the larger bubbles, but the small ones will remain.

I would think that a single airtight melting container might
be able to remove the air if it was pumped to a vacuum, and
physically agitated to get the bubbles to rise. After the air
is out of the plastic it could then be extruded by positive pressure.

Hm. I'm not entirely sure I understand your post, Mr. Attwood, but it sounds like you're saying that if I melted granules in the primary chamber and transferred them directly to the secondary chamber as I described, then the fluid will be a suspension of air bubbles in the liquid thermoplastic, which, when pumped through the nozzle, will extrude inconsistently and allow bubbles to distort the build object after extrusion. Is this correct?

Also, the solution you suggested seems complicated. Is there perhaps some simpler way to restructure the design, etc, that will eliminate or reduce the problem?

Yes, bubbles in the plastic will cause the extrusion
to be inconsistent. That is the reasoning behind the
variable pitched extruder screw, it should squeeze
out the air.
[reprap.org]


It's simpler, just one melting pot, and a vacuum pump.
Under a vacuum it should be possible to heat the
plastic a bit hotter without it burning, then
"agitate" the melting pot to get the bubbles to rise.
When it is ready to extrude just reverse the motor
on the pump and open the valve to the extruder for
filament production.

There is a problem with heating the melt in a pot - when you pressurise it to blow it out, it may well form rat-holes - a comon in viscious and powdery product where a small section flows out leaving most of the rest just sitting there. The solution is a 'follower plate' which seperates the pressurising air and apples the pressure accross the whole of the plastic surface. Like an air pushed syringe.

.. wich then will cause that the airbubbles cannot be removed by applying a vacuum.

'sid

Wow, lot of air bubble problems...I've been thinking about his for about a year now and reading up on how industry does this extrusion thing. It is, as far as I can tell, always done with a variable drive screw. It just requires a shit ton of force to squeeze the air bubbles out of stuff and the drive screw provides the needed mechanical advantage. The problem for us is that that method is expensive and specialized and not really small enough to mount on a Darwin carriage assembly.

I have no answers but I'm continuing to read and think about it. You guys are at least coming up with thoughts and ideas. Keep it up.

Demented

as you say that ..yes,
all done by "multistage" drive screws.
but I've found some "handdrill" that is actually an extruder for plastics.

(uups, THAT IS a large picture.. sorry, now linked only)
[www.americanpwt.com]

I think one could "replicate" something like that, can't one ?

(oh and as I write that I'm sure viktor already has something similar on his shell... )

'sid

Edited 1 time(s). Last edit at 08/03/2008 12:25PM by sid.

... not in my basement yet

But here: [www.orbi-tech.de] (flyer: [www.orbi-tech.de] ) is a 'hand-injection-moulding-gun' with nearly 1000bar pressure.

AFAIK the pricing is something to 2.800 Euros for the hand-gun excl. the press ...

Viktor

Demented Chihuahua Wrote:
> It just requires a
> shit ton of force to squeeze the air bubbles out
> of stuff and the drive screw provides the needed
> mechanical advantage.

Where does the air go anyway, is there some sort of membrane or gap that can pass air but not the molten plastic?

Andromodon Wrote:
-------------------------------------------------------
> Demented Chihuahua Wrote:
> > It just requires a
> > shit ton of force to squeeze the air bubbles
> out
> > of stuff and the drive screw provides the
> needed
> > mechanical advantage.
>
> Where does the air go anyway, is there some sort
> of membrane or gap that can pass air but not the
> molten plastic?

Hehe, what a coincidence. I have some samples of PTFE unlaminated membrane filters [www.sterlitech.com], 0.2 micron pore size, 13mm diameter. I just sent an email to Sterlitech about it, asking whether a thermoplastic like PTFE or ABS would pass through these filters, and I await a reply. It should work, though, as it doesn't even allow water vapor to pass without 40 PSI of pressure.
A different membrane material (like silver or ceramic, or possibly even steel) may be better for our application, but this just happened to be what I had (and I'll test it after I get my normal extruder running).
If you'd like to acquire some of these filters for testing yourself, you may be intimidated by the high cost of the 100-qty packs. If you email them, you can order sample packs (if I recall correctly, it's rougly the same unit price). The minimum quantity is 10 units (at least for the ones I purchased). Considering their size, this makes sense; just imagine a bunch of little 13mm circles cut out of paper and put in a little zipper bag. It's not worth it for them to sell less (then again, though, I was a retail customer).

Anyway, there are probably quite a few ways these can be implemented. The first thing that springs into my mind is a sort of mini-piston.
Liquid thermoplastic enters a chamber, and the chamber closes (air-tight seal?). The chamber has a plunger which compresses the contents of the chamber against a filter, squeezing all the air out through the filter. When the plunger is almost touching the filter, a valve opens on the wall with the filter, and the plunger shoves all the thermoplastic out the valve. When the plunger is against the filter, the valve shuts and the plunger returns to the original position, and the process repeats.

There might be a continuous way to do this (you might want to look into automobile combustion engines), but I'm more concerned with other things at the moment, so this is about all I can offer for now.

Also, it may be much simpler than using a piston. I don't have anything specific in mind, I just have a feeling about this (and I would investigate further if I had the time; I will investigate later on if no one else does by then).

Doesn't need a membrane. The air is vented out the hopper end. The pressure forces the air back up the screw. You basically have a static air situation. No air really goes in, and no air really goes out. Just forces the beads together and voila! extrudate!

Actually, it's way more complicated than that. The ones I've seen have multiple stages of threading that can't even really be called threading as it is changing pitch, thread count, depth, and diameter all at once. That's why the pro extrusion machines are so expensive.

Demented

Demented Chihuahua Wrote:
-------------------------------------------------------
> Doesn't need a membrane. The air is vented out the
> hopper end. The pressure forces the air back up
> the screw. You basically have a static air
> situation. No air really goes in, and no air
> really goes out. Just forces the beads together
> and voila! extrudate!
>
> Actually, it's way more complicated than that. The
> ones I've seen have multiple stages of threading
> that can't even really be called threading as it
> is changing pitch, thread count, depth, and
> diameter all at once. That's why the pro extrusion
> machines are so expensive.
>
> Demented

Well, with a filter it could really shove the air out of the system, and it wouldn't be so complex. There's a clear path for the air, and a clear path for the plastic. After the air is shoved out, the plastic is moved to a sealed extrusion head. You don't need to route the air back and forth in the chamber; the difficulty in doing this seems to be the major source of complexity in this "pro" idea.

VDX Wrote:
-------------------------------------------------------
> ... not in my basement yet
>
> But here:
> [www.orbi-tech.de]
> ectiweld_injection_moulding.htm (flyer:
> [www.orbi-tech.de]
> ast_Injection_Mould_Prototyping.pdf ) is a
> 'hand-injection-moulding-gun' with nearly 1000bar
> pressure.
>
> AFAIK the pricing is something to 2.800 Euros for
> the hand-gun excl. the press ...
>
> Viktor

NOT? I'm shocked!

yeah I know that one but that thingy needs the same filament we need, so it's not quite what we are looking for.
The one I mentioned uses granules to extrude filament that we can use with the extruder

'sid

@Joshua

Actually, the differing diameters, pitches and what not are there for good melt quality and consistency in extrusion. Not for air removal. That is a by-product of all that. Would it happen otherwise...I don't know. Regardless, they do it for some reason and we need a work around. The membrane might be it or it might be something else. I guess the solution we need only needs to be "just" good enough.

Demented

Demented Chihuahua Wrote:
-------------------------------------------------------
> @Joshua
>
> Actually, the differing diameters, pitches and
> what not are there for good melt quality and
> consistency in extrusion. Not for air removal.
> That is a by-product of all that. Would it happen
> otherwise...I don't know. Regardless, they do it
> for some reason and we need a work around.
That's precisely the use of the dual chamber melter I described earlier. It maintains a constant amount and temperature of thermoplastic in the chamber connected to the pump, and the pump can be designed to be constant speed. Is there any reason this would lead to inconsistent extrusion?

Combined with the piston chamber feeding into that dual chamber, you'd have bubble-less, consistent extrusion. You'd need a primary melting chamber feeding into the piston chamber, so it would end up being a 4-chamber extruder.

It wouldn't have to be as bulky as one might imagine, however, as the chambers can be quite small. We could probably get it somewhere between .5 to 2 times the size of the current extruder.

Demented Chihuahua Wrote:
-------------------------------------------------------
> @Joshua
>
> Actually, the differing diameters, pitches and
> what not are there for good melt quality and
> consistency in extrusion. Not for air removal.
>
> Demented

Actually, the melt quality depends on air removal in large amounts

'sid

On a completely different route:
There may be chemical ways to remove the air too. If we in advance saturate the extruder with a gas that we can either solve in the material (without seriously adjusting material properties (in a bad way, that is)) or make something react with it so that together it becomes a solid. The original elements in there wouldn't take up any significant space anymore...

I'm didn't study chemistry since secondary school though

---

Regards,

Erik de Bruijn
[Ultimaker.com] - [blog.erikdebruijn.nl]

@Joshua

I suppose that'll work--but I never actually "know". Work up a design using parts that aren't crazy and we'll give it a go. Gotta keep costs down though.

Demented