Support material candidates

Nophead had a recent blog about being able to print without any support material. Once the plastic is part of a base and hardens, there appears to be little chance for it to move much. The sand idea was based on this method and would give support to long or large overhangs. I was also thinking of making darwin do SLS or SLA if I could get the right type of laser at a price I could afford.

As for the multi head system, nobody said anything about the head design I had on the PCB page (http://forums.reprap.org/posting.php?14) but here they are again (currently designed for PCB production, but could be tweaked for support).


Yeah, it works well for Zcorp because they only do stuff for prototyping. You do have to admit it's a good starting price though for a commercial unit.

I think the snap off parts could work well. Thin upside down pyramid structures or lattice of sorts that wouldn't take up too much material and be easy to take off. I saw a great pic of this method somewhere but can't find it again.


As for a new idea. reusable mold making compound. It takes a while to harden, but can be melted off at a low temp and reused. [compositherm.com]
Someone might even be able to make a recipe that hardens fast enough to print even if it compromises the same kind of clarity. I have found some other recipes for this, but it contains glycerin and we might not want to put it through here. I'm not too sure how flammable or volatile it is.

Glycerin is very safe. You will find it in any Pharmacy and it can be heated up to over 200

Dylan

[forums.reprap.org]
Just brings up a reply box, I can't see any posts.

In that case, here is one link I found for the recipe [forums2.gardenweb.com] .

Odd, I thought the PCB link would work. [forums.reprap.org]
that should work now.

Hopefully some sort of 'gel' could be made to work well enough to take off easily and be reused also. This is why I suggested the reusable mold compound. I don't know if that recipe can be changed to dry fast enough.

I like the agar idea.

And initial experiments seem to have been positive

My feeble attempts with Crystallization and Epsom salts came to zip. Too unwieldy and as likely to jam the nozzle as any aggregate.

Yup adding a simple anti bacterial agent would help too (Detox or something equally noncorrosive)

Dispenser options.

1. Heated syringe and needle. (nichrome element)
2. Adrian's fields metal melting pot thingy.
3. An existing extruder with the temperature turned down (does it need modifying ??)

Yup a second head on a two head machine would be great. We would have to find some way to move the unused head up out of the way when not used though so it did'nt foul the work piece. maybe a solenoid lift ????

I know that support is a perennial and contentious one.

Thoughts for what they are worth.

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

this is a real head-banger. My thoughts was to find a decently hard material that can then be rinsed of with some kind of catalyser that doesn't affect the plastic.

how about gelatine? to wobbly?

mimarob Wrote:
-------------------------------------------------------
>
> how about gelatine? to wobbly?

Doesn't set fast enough. Agar is better in that respect and is less wobbly.
It works fine for "cold" deposition. But for molten plastic, you need a support material that sticks to it so that it prevents warping.

The best candidate, IMHO would be the same plastic, deposited in support columns and pyramids that have to be removed after the whole process.

Has anyone suggested solder yet? Or any other metal?

It does cool pretty quickly, and would cool quicker if you had a fan or something actively blowing on it.

Quote

Has anyone suggested solder yet? Or any other metal?

Wouldn't you have problems with there not being a handy viscous phase? It might be difficult to build vertical structures with solder if it melts together into a big blob, or flows before solidifying.

And again, the usual important question, will the filament stick to it? I'm gonna hazard a guess here and say 'no', but feel free to correct me

Ru Wrote:
-------------------------------------------------------
> Wouldn't you have problems with there not being a
> handy viscous phase? It might be difficult to
> build vertical structures with solder if it melts
> together into a big blob, or flows before
> solidifying.
I imagine you could control it such that you could get a fairly flat surface for a layer with some tweaking (or at least a predictably curved one, based on the properties of the solder mixture).

> And again, the usual important question, will the
> filament stick to it? I'm gonna hazard a guess
> here and say 'no', but feel free to correct me
I don't really understand the need for the filament to stick to the support material.

If the filament is extruded on a piece of metal that it doesn't stick to, when it cools, it will contract, leading to a smaller than expected piece. It will also warp, depending on the design of the piece, because there could be more contraction in one area than another connected area, due to differences in amounts of material (with different shapes, etc). (I suppose the other problem would be the difference in rates of contraction... now that I think about it, this would probably be the main cause of warping, if not the only one.)

Is that the correct description of your warping problem, which supposedly necessitates adhesion between the extruded material and the support material?


If it is, it may be possible to predict the contraction of the extruded material (which may be a challenge on the software end, but not necessarily impossible/unfeasible), and then, accordingly, lay out the extruded material in such a way that when it contracts, it forms the correct size and shape.

Is this physically possible? Or is the contraction too unpredictable/uncontrollable (if this is the problem, then please explain why)?
Or is my understanding of the necessity of adhesion flawed?

Joshua Merchant Wrote:

Quote

I imagine you could control it such that you could get a fairly flat surface for a layer with some tweaking (or at least a predictably curved one, based on the properties of the solder mixture).

You can only extrude paste like substances. Liquids will just flow out of the nozzle uncontrollably and run away. There is a metal extruder but that is for filling channels in plastic.

Quote

Is that the correct description of your warping problem, which supposedly necessitates adhesion between the extruded material and the support material?

The main reason for warping is that hot layers are extruded onto cold layers that have already frozen. They in turn freeze but they are still hot so they contract afterwards. This puts the layer below under stress leading to it curling upwards.

That is not the only reason for needing adhesion though. Simply extruding a line and then turning a right angle would cause the filament to drag about if it did not bond to the support.

The warping is predictable using finite element analysis that keeps track of the temperature of each small element as it cools and calculates the stress on its neighbors.

Compensating for it is another matter. I don't think all shapes can be compensated for and it requires 3D extrusion paths with variable velocities and support material.

---

[www.hydraraptor.blogspot.com]

nophead Wrote:
-------------------------------------------------------
> You can only extrude paste like substances.
> Liquids will just flow out of the nozzle
> uncontrollably and run away. There is a metal
> extruder but that is for filling channels in
> plastic.
Oh, I didn't mean you would use a standard extruder head. It would probably have to be some complex piece of equipment (which I won't discuss unless the material itself is usable).

As for the rest of the information, it's good to know. I'll mull over it later (in a few days, I have to leave town this weekend) and report back when I come up with a solution.

How about simply injecting a given quantity of a suitably viscous liquid into areas that need support? It would not work for external support (i.e overhangs) but could be used to fill in vertical holes.

Perhaps a bit of a long shot but whatever solution you use, it would have to be fairly compact to avoid reducing the usable work area.

it would work for exterior support with a printed frame around the actual object.

But you'll have to make sure that the warping doesn't release the object from the build tray, else it wont help either way

calculating the ammount of liquid (i'd guess water) will be quite thrilling

'sid

Quote

Oh, I didn't mean you would use a standard extruder head. It would probably have to be some complex piece of equipment

Not so good. Complex implies 'full of bits which you can't replicate', I'm thinking. It also suggests 'hard to build' and 'awkward to start using', neither of which are really desirable :/

Quote

How about simply injecting a given quantity of a suitably viscous liquid into areas that need support

I love suggestions which include the word 'simply' They seem to involve a lot of handwaving away of serious complexities. Simple support material ideas like this might work to support spans, where there will be long straight filaments forming a new horizontal surface between two existing support pillars of extrudate.

They probably won't work to support protrusions and overhangs where filaments do not describe straight arcs from one solid point to another, but instead must form corners on the support material, because they won't stick to that material.

In the case of fluids, you've got additional restrictions whereby the support volume must be enclosed, and you want something that is denser than the extrudate, and doesn't cool the extrudate too quickly, and you want to squirt out precise measures of fluid...

And so on.

I'm skeptical, anyway, as you can probably tell

... instad of fluids you can use small glass-spheres, as used for sanding - it's cheap, reusable, behaves nearly like a fluid, but didn't cool down the surrounding plastic-body.

I have some 100 gramms of this stuff - the size of the spheres is something around 100 microns, but it's not uniform - i found some bigger ones with nearly 1mm size, very fine spheres smaller then 10 microns and some ovoids, coagulated grapes or broken ones.

So you have to pour them through a 200micron-size or finer filter-sieve, but then you have a good filling- or support-medium ...

Viktor

Edited 1 time(s). Last edit at 07/26/2008 04:03PM by Viktor.

I did not believe that it would be simple. Perhaps I should stick to building a working repstrap first to have something to test with .

Quote

I did not believe that it would be simple. Perhaps I should stick to building a working repstrap first to have something to test with smiling smiley.

That's pretty much the conclusion I came to with my own ideas

Happily, I can still heckle on things where I feel I am on slightly more solid ground. Building a repstrap is going to be a long and expensive process

I guess I should turn my hand to some object design, or something else that doesn't require any hardware...

Ru Wrote:
-------------------------------------------------------
>
> I am on slightly more solid ground. Building a
> repstrap is going to be a long and expensive
> process
>

Possibly not.

Quote

Possibly not.

Well, yes and no. The cost of the electronics alone was slightly surprising (it seems to be about twice the cost of the same bits in the US; the usual premium we pay). I've just been looking at the cost of steppers too...

I think I'm just living in the wrong part of the world A few thousand miles west and everything is half the price it is here, or cheaper still. A few hundred miles north, and people seem to be throwing away old machine tools for pocket change.

Bah.

Quote

A few hundred miles north, and people seem to be throwing away old machine tools for pocket change.

Where's that then?

---

[www.hydraraptor.blogspot.com]

Quote

Where's that then?

I've been looking around on ebay for a lathe for a little while... if a lathe is small and potentially courierable for a sensible sum of money, it tends to sell for

I'm back, yay.

nophead Wrote:
-------------------------------------------------------
> The main reason for warping is that hot layers are
> extruded onto cold layers that have already
> frozen. They in turn freeze but they are still hot
> so they contract afterwards. This puts the layer
> below under stress leading to it curling upwards.
Is heating the object feasible? That is, heat the entire thing (or maybe the most recent layers in gradient) to a max temperature a bit below the melting temperature (at a point where it is at least semi-stable and won't flow), and when you extrude the new layers, the temperature difference between the new layers and the layers they sit on will be much lower, and so warping will be minimized. Could this work?

> That is not the only reason for needing adhesion
> though. Simply extruding a line and then turning a
> right angle would cause the filament to drag about
> if it did not bond to the support.
Perhaps we could sever the connection between the filament in the extruder and the filament outside the extruder? Could the valve(?) that will be used to stop dripping on the extruder when not extruding be used to do this? The way I'm imagining it is closing it and reopening it every time you make a turn beyond a certain angle.

Or perhaps, before turning, the extruder could put another layer on top of the layer that rests on the support (giving time/cooling for the first layer to dry?) which starts the separate line (that is, instead of directly turning, going vertical and then turning). I don't like this second idea, though, and I don't think it would work well, if at all.

> You can only extrude paste like substances.
> Liquids will just flow out of the nozzle
> uncontrollably and run away. There is a metal
> extruder but that is for filling channels in
> plastic.
What I'm thinking is that if the support (solder/low melting point metal/etc) needs to be on the outside portion of any layer, then a sort of container-barrier tube would be extruded around the build area, such that any solder on the outside of a layer would be contained in an area between the side of the layer and the wall of the tube. The tube would be shaped such that it could be easily removed at the end of building and creates as little space between the object and the tube's interior walls (to not waste support material/time to extrude support material).


I imagine you could have either of two setups. First, you could have a build preparation that is setup by the extruder at the beginning, consisting of the tube to contain the support connected to a channel for the support substance to a heating element. You then extrude a small amount of the support at the bottom of the tube and through the channel. The heating element heats the support substance, which conducts heat to the build (maintaining the higher temperature to prevent warping, as above). Personally, I don't like the idea of having to setup such a heating element, channel, etc.
The other way would be to build the tube with an extra space on one side (or all around) in which a small amount of the support material is deposited into on every layer. A heating element is attached to either the extruder or the bottom of the tube or independently controlled (or whatever, this part is less important at the moment than the concept itself). It heats the column of support element (or rather the top/bottom of the column; if the heating element is on top, then it can heat the uppermost part of the column, which leads to the gradient in temperature I spoke of earlier), which conducts to the rest of the build.


In any case, the solder would have to melt at lower temperatures than the plastic being used (so that it could be melted away at the end; if anyone has better suggestions for removing it, do tell), and the temperature maintained at the newest layers would have to be slightly below the solder's melting temperature (or the solder would melt and not support the newly extruded layers).
The thing that concerns me is that I think the solder might melt when the melted plastic is extruded on top of it (as the solder would melt at lower temperatures)...

The only alternatives to melting the support material away at the end, that I can think of right now, are to either dissolve it or to use some chemical that reacts with it and has no solid products (such that the products are fluid and flow out of the build). I don't know how either could be used, but if either are possible, then a solder could be chosen which melts at a higher temperature than the plastic.


I think that's all... my brain is a bit fried from other stuff, so I might have left something out.





Edit: Now that I think about the complexities of using a support material...

It may be better to simply do away with the idea totally. It would probably be simpler and cheaper to pivot the building platform (and/or build object). The goal, of course, would be to change the direction of gravity relative to the build object, thus allowing the creation of any shape (eliminating the overhang problem). The biggest problem with this, that I see, is the possibility of the build object sliding/falling off the build platform.
Alternately, I suppose you could orbit the extruder around the object, and use a flat piece of support material that you stick out under the extruder nozzle to support the material while it's being laid down, and then when it cools and is stuck to the object, you remove the support (possibly preventing breaking by heating the support material), I'm thinking of a metal blade or rod or wire or something. I think the rotating platform would work better.

Edited 1 time(s). Last edit at 07/28/2008 01:36AM by Joshua Merchant.

Quote

Is heating the object feasible

It is tricky, to say the least. And you have to have some extruded structural parts *inside* the heating volume (ie, the extruder head), which means you'd have to make them out of higher temperature plastic such as HDPE which is itself more prone to warping... and for a reprap, you'd need to be able to fabricate most of the heated volume container, and so on

Quote

It would probably be simpler and cheaper to pivot the building platform (and/or build object).

I was pondering this the other day, having been looking at some stuff about 4 and 5 axis milling machines.

Mechanically, building such a reprap would be somewhat harder; making it cope with the limited precision that reprap parts have is even harder. But the major problem is toolpath generation. Trying to calculate head trajectories in 5 axes to do sparse infill on a multi axis platform is going to be nightmarish.

It becomes somewhat easier if you have 3+2 indexed axis positioning whereby the A and B axes are moved into position, but lock in place when you're laying out filament. We already do this with the Z axis, of course.

Now you've chopped down the problem domain quite a bit, but you now need to dice up the STL into a series of subparts which can be printed using just XYZ movement. Each subpart can be sliced and diced as normal, but with the following additional requirements:

1) You need to know something about toolhead geometry so you don't crash it into a previously printed subpart which is not totally rotated out of the XY plane of movement of the head.

2) You need to plot joins of two subparts at a point *extremely carefully* for the same reason... you don't want to end up having to fill in a gap which is too narrow for the print head to fit into without crashing into the item.

2) You need to make sure that the parts you're building are solid enough that they won't flex when rotated (so they don't hit other parts, or the print head, or bend out of the way when you expect to print another subpart which attaches to them) and don't sag because they haven't cooled enough after printing to resist gravity when they are rotated out of the XY plane.

Anyway. Tricky. I've put a bit of thought into it, cos it would be terribly cool, but it is a problem of a similar order of magnitude to support material choice, if not more so.

A cursory search for 5 axis milling software suggests that such things are over $10k for a single license, for a cheap tool. It isn't an easy problem to solve

So did we ever find a good support material? I found the Paste [Support] Extruder documentation [reprap.org], but I can't seem to find any posts/blog entries about testing it as a support extruder.

So far as I understand it.......no

Pasta dough is a possible someone needs to try.

A partialy crystalised alchohol suspension of Epsom salts failed miserably clogging up a fine orifice. (Not planning any further work on this until after I have a Darwin to work from)

Anything you can heat extrude and then wash off which comes in at the 1$ per bucket level, either suffers the same crystalline clogging issues, short pot life or melts/degrades at too low a temperature for the plastics we are currently working with (Not just Capa)

Nop Head has had quite some success using ABS and trailing how far he can push certain things without support. (ABS being higher temp sets up more rapidly) and using sparse joining layers to make for easy separation (Rafts and components built on top). I think Adrian had a play with the software to add this in and expressed some suggestions that it would be good to try using the plastic itself as support but using the sparse separation layer technique to make the support easily removable after the fact. (great for accessible areas but would struggle with internal voids particularly those with smaller access than the void dimensions.

ABS prints better than some other plastics without support, in some cases teardrop shaped holes are not necessary.

I think we need to see some folk with a functioning strap/darwin and a viable tool head for putting paste and other things down ie Adrians pneumatic pop bottle or a syringe depositor for example and trialling potentials with it.

Even wacky things would be steps forward, a negative result is still a step forward in learning. (This is how the Fab@Home folk found pasta dough was doable)

Hope this helps.

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

What's wrong with the various(?) water-soluble polymers?

I stumbled upon polyvinyl alcohol [en.wikipedia.org], which "is a water-soluble synthetic polymer" with "excellent film forming, emulsifying, and adhesive properties". Perhaps one of our thermoplastics will stick to it? I'm not sure how difficult it will be to extrude this without decomposition, unfortunately. Then again, I can't figure out what it decomposes to, so it may or may not be a problem.

Anyway, whether PVOH specifically would work or not is not so important, as there seem to be quite a few water soluble polymers to investigate (well... actually... I could only find two or three that might work). One possible candidate (from my sort searching) seems to be Aquabond Technologies' thermoplastic adhesives [www.aquabondtechnologies.com]. It comes in a few different forms and dissolves in hot water (they say 100% solubility at 80

Edited 1 time(s). Last edit at 08/23/2008 08:53PM by Joshua Merchant.

Quote

What's wrong with the various(?) water-soluble polymers?

PVA looks interesting... relatively easy to get hold of, safe to work with, high enough melting point that ABS won't melt PVA it is extruded on to, etc etc. The cost isn't too much of a problem if it is reclaimed after each build.

I'm pretty certain that the time it takes to set would be the killer here. As you have some handy, try toasting it with a hairdryer or something and seeing how quickly it congeals?

Quote

Mr. Bowyer

That should probably be Dr. Bowyer I'm sure he didn't spend all those years at evil replicating robot school just to be called 'Mr.'