Concept for negative-Z 3d printer
Posted by Eagleon
|
Concept for negative-Z 3d printer March 18, 2010 12:38AM |
Hey, first post. I hope I'm not in over my head here - I was extremely excited to learn about everything that's being done now in RP here, and will be looking at making my own Mendel in the near future! I have a lot of ideas, many of which are probably beyond my own skills. Forgive any ignorance! I'm a java programmer - I always thought experimenting with mechanical engineering and electronics would be a prohibitively expensive hobby, so I've never nurtured my interest in it until now.
Anyway, on to the meat of the subject. The idea is, instead of building an object up, push it down as you go. As far as I've gathered, there is an issue with sloping structures in printed objects (overhang). This concept might just avoid that problem altogether (sort of), and it also eliminates the need to move the tool-head along the z axis at all.
My idea is to use a play-doh bed as the drawing surface. Play-doh? Yes, that is correct, or something with similar properties, preferably something that can be balled up again and reused. Please keep reading, it's not as silly as it sounds (I hope
). The idea is that, as the toolhead lays down layers of material, once a particular z-level is done, the object itself is pressed into the bed such that it is flush with the whole surface. This could be accomplished by a hinged, rigid sheet, or other means I'm sure - so long as it's blocked from pressing the object too far in (a reenforced guide-bar along the edges would stop it), you could even just drop a weight on the thing, as long as it's flat enough.
After the object is pressed down, the toolhead moves back into place (from a safe location!), and resumes printing on the same level as before. As the object is pressed inward, a plate on which the malleable substance rests extrudes it out the bottom through several holes into a bag to be recycled. This is to prevent the material from wanting to compress, which would instead I think result in it seeping off to the sides and making all sorts of mess and problems.
There are a couple of benefits that I see to this method, and some possible kinks in the plan (that's a good sign, right? I'm not saying this will be easy) First, the time it takes to press the object down might be significant, and would probably change as a function of the size of the object - the problem with how much to press down for how long could, I think, be solvable by including a contact point somewhere along the guidebar and on the press, which would bridge a circuit telling the system to retract. I think this will always be slower than configurations which move the z-level, but I believe it would make, for instance, changing tools out automatically a much simpler problem to approach, and it's one less source of inaccuracy for the user to worry about.
Second, if any of the material seeps onto the top of the object, this could cause ugly pits in the final result. Worse, if there are any imperfections in the press surface, it might pull some of the material up, causing the object to shift. There has to be a careful balance between rigidity and compressability in the mixture, and it must be as non-sticky (unsticky?) as possible - it has to support the weight of the object, while still being able to be extruded out the bottom. Larger objects in the z-axis will require deeper beds, which will require more force to extrude. All of this is up in the air as to what would be best.
Third, I can see problems with certain structures where a form that tapers outward, again tapers inward, creating a cavity in the flat form (eep!) Like so: example image of problem design (forgive the crude drawing, I don't have a proper tool for it on this computer) This still makes the approach useful for some objects with extreme angles, like more complex spindle gears, because the object itself will still be a target for the tool, like normal RP machines. I'm not real sure about this problem, to be honest. The material will tend to slope inward, possibly interfering with the building of certain designs. In particular, a mushroom-like configuration would end up fairly distorted, with the rim cutting into the material, which then falls inward on the inner spire, making for an uneven surface. Maybe pausing the project to pack in additional material might work, but it would be awkward.
Fourth, the object itself might shift as it is pressed, particularly if it's non-symmetric. I do not see an easy solution to this. It's an obvious draw-back. Maybe someone else can think of a way to mitigate the problem.
I think I've bashed the idea enough, considering it's my own, haha. Some of the other advantages I see are an infinitely expandable Z-size - need more space? Just move the machine upward and increase the depth of the bed. You could make long rods this way by letting them protrude out of the bottom. An interchangeable bed design would allow you to use silly stuff like fishing nets and paint pans. Did I mention no z-motor control?
What do you guys think? Is it workable? Has it been done/thought of before? Should I get back to programming/the kitchen?
Anyway, on to the meat of the subject. The idea is, instead of building an object up, push it down as you go. As far as I've gathered, there is an issue with sloping structures in printed objects (overhang). This concept might just avoid that problem altogether (sort of), and it also eliminates the need to move the tool-head along the z axis at all.
My idea is to use a play-doh bed as the drawing surface. Play-doh? Yes, that is correct, or something with similar properties, preferably something that can be balled up again and reused. Please keep reading, it's not as silly as it sounds (I hope
). The idea is that, as the toolhead lays down layers of material, once a particular z-level is done, the object itself is pressed into the bed such that it is flush with the whole surface. This could be accomplished by a hinged, rigid sheet, or other means I'm sure - so long as it's blocked from pressing the object too far in (a reenforced guide-bar along the edges would stop it), you could even just drop a weight on the thing, as long as it's flat enough.After the object is pressed down, the toolhead moves back into place (from a safe location!), and resumes printing on the same level as before. As the object is pressed inward, a plate on which the malleable substance rests extrudes it out the bottom through several holes into a bag to be recycled. This is to prevent the material from wanting to compress, which would instead I think result in it seeping off to the sides and making all sorts of mess and problems.
There are a couple of benefits that I see to this method, and some possible kinks in the plan (that's a good sign, right? I'm not saying this will be easy) First, the time it takes to press the object down might be significant, and would probably change as a function of the size of the object - the problem with how much to press down for how long could, I think, be solvable by including a contact point somewhere along the guidebar and on the press, which would bridge a circuit telling the system to retract. I think this will always be slower than configurations which move the z-level, but I believe it would make, for instance, changing tools out automatically a much simpler problem to approach, and it's one less source of inaccuracy for the user to worry about.
Second, if any of the material seeps onto the top of the object, this could cause ugly pits in the final result. Worse, if there are any imperfections in the press surface, it might pull some of the material up, causing the object to shift. There has to be a careful balance between rigidity and compressability in the mixture, and it must be as non-sticky (unsticky?) as possible - it has to support the weight of the object, while still being able to be extruded out the bottom. Larger objects in the z-axis will require deeper beds, which will require more force to extrude. All of this is up in the air as to what would be best.
Third, I can see problems with certain structures where a form that tapers outward, again tapers inward, creating a cavity in the flat form (eep!) Like so: example image of problem design (forgive the crude drawing, I don't have a proper tool for it on this computer) This still makes the approach useful for some objects with extreme angles, like more complex spindle gears, because the object itself will still be a target for the tool, like normal RP machines. I'm not real sure about this problem, to be honest. The material will tend to slope inward, possibly interfering with the building of certain designs. In particular, a mushroom-like configuration would end up fairly distorted, with the rim cutting into the material, which then falls inward on the inner spire, making for an uneven surface. Maybe pausing the project to pack in additional material might work, but it would be awkward.
Fourth, the object itself might shift as it is pressed, particularly if it's non-symmetric. I do not see an easy solution to this. It's an obvious draw-back. Maybe someone else can think of a way to mitigate the problem.
I think I've bashed the idea enough, considering it's my own, haha. Some of the other advantages I see are an infinitely expandable Z-size - need more space? Just move the machine upward and increase the depth of the bed. You could make long rods this way by letting them protrude out of the bottom. An interchangeable bed design would allow you to use silly stuff like fishing nets and paint pans. Did I mention no z-motor control?
What do you guys think? Is it workable? Has it been done/thought of before? Should I get back to programming/the kitchen?
|
Re: Concept for negative-Z 3d printer March 18, 2010 01:11AM |
What do you guys think? Is it workable?
In theory, quite nice. In practice, imagine pushing a non-perpendicular wedge into into playdough. It's going to start to 'wander', or go off-normal.
As you say:
Fourth, the object itself might shift as it is pressed, particularly if it's non-symmetric. I do not see an easy solution to this. It's an obvious draw-back. Maybe someone else can think of a way to mitigate the problem.
I think this may be unworkable.
Far better to make a tall LOM or powderbed RepRap.
-Sebastien, RepRap.org library gnome.
Remember, you're all RepRap developers (once you've joined the super-secret developer mailing list), and the wiki, RepRap.org, [reprap.org] is for everyone and everything!
In theory, quite nice. In practice, imagine pushing a non-perpendicular wedge into into playdough. It's going to start to 'wander', or go off-normal.
As you say:
Fourth, the object itself might shift as it is pressed, particularly if it's non-symmetric. I do not see an easy solution to this. It's an obvious draw-back. Maybe someone else can think of a way to mitigate the problem.
I think this may be unworkable.
Far better to make a tall LOM or powderbed RepRap.
-Sebastien, RepRap.org library gnome.
Remember, you're all RepRap developers (once you've joined the super-secret developer mailing list), and the wiki, RepRap.org, [reprap.org] is for everyone and everything!
|
Re: Concept for negative-Z 3d printer March 18, 2010 03:13AM |
rhmorrison Wrote:
-------------------------------------------------------
> For this idea to be workable your "Play-doh" would
> have to be a liquid with a high enough specific
> gravity to be able to extrude on top of it like
> mercury (although it is toxic and expensive). But
> it would flow around the object as a thick fluid
> and provide enough support to extrude onto.
>
> If a 'liquid' with the proper viscosity that is
> cheap and safe could be found I don't see how this
> pushing is going to work without deforming the
> object. Also as the object is pushed into the
> material the level of the material will rise for
> the amount of volume now under the surface.
>
This is the purpose of the uniform push per z-level, and the reason the bed has holes - both the material and object are brought into the same plane, and any displaced material is extruded out of the bottom instead of rising. Obviously it will rise a little during each step when the object is pushed in, but that's also automatically corrected for and eliminated before doing it again. There is no need for anything exotic like mercury - clay would work, except that it hardens too rapidly and probably would shrink. This is why my first thought is play-doh. An object suspended in clay will not sink, unless it is very heavy, and will retain its shape, unlike liquid. Also, deformation of the object would depend on what material you are printing with. If it's more rigid than the surrounding material (given things like surface area of the object, etc, obviously), the material will deform more significantly first. At least that's the theory
SebastienBailard Wrote:
-------------------------------------------------------
> What do you guys think? Is it workable?
>
> In theory, quite nice. In practice, imagine
> pushing a non-perpendicular wedge into into
> playdough. It's going to start to 'wander', or go
> off-normal.
>
> As you say:
> Fourth, the object itself might shift as it is
> pressed, particularly if it's non-symmetric. I do
> not see an easy solution to this. It's an obvious
> draw-back. Maybe someone else can think of a way
> to mitigate the problem.
>
> I think this may be unworkable.
>
> Far better to make a tall LOM or powderbed RepRap.
It might be worth making for symmetrical objects alone, if it could be made cheaply and easily enough. Things like gears, medium-sized rods, etc. The lack of a need for precise z-control using controlled motors (which powder-printers still require, as far as I understand) would simplify the design requirements tremendously. Here it's a simple matter of squishing something flat along a pre-measured plane. I don't know - obviously ideally something as flexible as a large reprap is going to be best, but I think this still has some advantages. I'll try and work out more details on my own time - for now, sleep
-------------------------------------------------------
> For this idea to be workable your "Play-doh" would
> have to be a liquid with a high enough specific
> gravity to be able to extrude on top of it like
> mercury (although it is toxic and expensive). But
> it would flow around the object as a thick fluid
> and provide enough support to extrude onto.
>
> If a 'liquid' with the proper viscosity that is
> cheap and safe could be found I don't see how this
> pushing is going to work without deforming the
> object. Also as the object is pushed into the
> material the level of the material will rise for
> the amount of volume now under the surface.
>
This is the purpose of the uniform push per z-level, and the reason the bed has holes - both the material and object are brought into the same plane, and any displaced material is extruded out of the bottom instead of rising. Obviously it will rise a little during each step when the object is pushed in, but that's also automatically corrected for and eliminated before doing it again. There is no need for anything exotic like mercury - clay would work, except that it hardens too rapidly and probably would shrink. This is why my first thought is play-doh. An object suspended in clay will not sink, unless it is very heavy, and will retain its shape, unlike liquid. Also, deformation of the object would depend on what material you are printing with. If it's more rigid than the surrounding material (given things like surface area of the object, etc, obviously), the material will deform more significantly first. At least that's the theory
SebastienBailard Wrote:
-------------------------------------------------------
> What do you guys think? Is it workable?
>
> In theory, quite nice. In practice, imagine
> pushing a non-perpendicular wedge into into
> playdough. It's going to start to 'wander', or go
> off-normal.
>
> As you say:
> Fourth, the object itself might shift as it is
> pressed, particularly if it's non-symmetric. I do
> not see an easy solution to this. It's an obvious
> draw-back. Maybe someone else can think of a way
> to mitigate the problem.
>
> I think this may be unworkable.
>
> Far better to make a tall LOM or powderbed RepRap.
It might be worth making for symmetrical objects alone, if it could be made cheaply and easily enough. Things like gears, medium-sized rods, etc. The lack of a need for precise z-control using controlled motors (which powder-printers still require, as far as I understand) would simplify the design requirements tremendously. Here it's a simple matter of squishing something flat along a pre-measured plane. I don't know - obviously ideally something as flexible as a large reprap is going to be best, but I think this still has some advantages. I'll try and work out more details on my own time - for now, sleep
|
Re: Concept for negative-Z 3d printer March 18, 2010 02:49PM |
Larry_Pfeffer Wrote:
-------------------------------------------------------
> @Eagalon,
>
> If nothing else, , I give your idea a point or two
> for creativity and use of an inexpensive support
> material. I don't know whether playdough would
> work well as a support material (of course I don't
> know that it wouldn't, either.)
>
> However, the issue of the part wandering in the
> horizontal is an issue, as has already been
> pointed out. Similarly, I think playdough springs
> back somewhat, so it may not stay pressed down.
> If it works, adding some sort of
> anti-fungal/anti-bacterial to the dough is
> probably a good idea. I'd try a few drops of
> bleach, or a little rubbing alcohol (not too much,
> we don't want the support material to ignite!)
>
> If you can, try some simple experiments pushing an
> object into playdough, and measuring how much it
> wanders and/or springs back. Please report
> results, whether positive or negative. One of the
> benefits of suggesting blue-sky ideas (even if
> they don't work) is that doing so prompts others'
> thinking -- and one of them may work.
> If the horizontal shift could be measured after
> the push-down (e.g. by a vision system), then it
> might be compensated for in software, by shifting
> the next layer to match. Rotations (about the
> vertical axis) would take more computation to fix,
> but would still (ahem, theoretically) be doable.
>
> Perhaps an electro-rheological fluid (or a
> magneto-rheological one) might work as an
> easy-to-apply, easy-to-remove support medium.
I'll give it a shot, will try to find some tonight and make some measurements. Whatever the bed would contain it would have to resist compression quite well to avoid the spring-back effect, as you said. I hadn't thought about something like the electro-rheological stuff - thought that sort of thing was still sci-fi. Using that would on the surface eliminate the need for a press entirely, assuming it's responsive enough and the fluid has a lower density than the plastic and the viscosity is low enough that it would sink down - just drop the field momentarily when you want the object to sink in further. I think the shifts would be even more of a problem than the press method, though - you'd see the object roll dramatically about both horizontal axes, especially objects with a large surface area compared to their weight (i.e, the first layer of nearly any project.) Too bad
Thank you guys for responding seriously
I knew this would have more problems than I initially thought of, and it saves me from obsessing over it in my head so that I can work on practical things.
-------------------------------------------------------
> @Eagalon,
>
> If nothing else, , I give your idea a point or two
> for creativity and use of an inexpensive support
> material. I don't know whether playdough would
> work well as a support material (of course I don't
> know that it wouldn't, either.)
>
> However, the issue of the part wandering in the
> horizontal is an issue, as has already been
> pointed out. Similarly, I think playdough springs
> back somewhat, so it may not stay pressed down.
> If it works, adding some sort of
> anti-fungal/anti-bacterial to the dough is
> probably a good idea. I'd try a few drops of
> bleach, or a little rubbing alcohol (not too much,
> we don't want the support material to ignite!)
>
> If you can, try some simple experiments pushing an
> object into playdough, and measuring how much it
> wanders and/or springs back. Please report
> results, whether positive or negative. One of the
> benefits of suggesting blue-sky ideas (even if
> they don't work) is that doing so prompts others'
> thinking -- and one of them may work.
> If the horizontal shift could be measured after
> the push-down (e.g. by a vision system), then it
> might be compensated for in software, by shifting
> the next layer to match. Rotations (about the
> vertical axis) would take more computation to fix,
> but would still (ahem, theoretically) be doable.
>
> Perhaps an electro-rheological fluid (or a
> magneto-rheological one) might work as an
> easy-to-apply, easy-to-remove support medium.
I'll give it a shot, will try to find some tonight and make some measurements. Whatever the bed would contain it would have to resist compression quite well to avoid the spring-back effect, as you said. I hadn't thought about something like the electro-rheological stuff - thought that sort of thing was still sci-fi. Using that would on the surface eliminate the need for a press entirely, assuming it's responsive enough and the fluid has a lower density than the plastic and the viscosity is low enough that it would sink down - just drop the field momentarily when you want the object to sink in further. I think the shifts would be even more of a problem than the press method, though - you'd see the object roll dramatically about both horizontal axes, especially objects with a large surface area compared to their weight (i.e, the first layer of nearly any project.) Too bad
Thank you guys for responding seriously
I knew this would have more problems than I initially thought of, and it saves me from obsessing over it in my head so that I can work on practical things.
|
Re: Concept for negative-Z 3d printer March 18, 2010 03:08PM |
WHAT IF . . .Current powder printers are complex because after each layer is fused (usually with laser sintering) a new perfectly even layer of powder must be spread over the area and when the object if finished you must brush and vacuum all the excess powder (and of course return it to the hopper - too expensive to throw away).
Now what if we had a liquid plastic instead of powder!
It should have the same specific gravity as the sintered plastic such that it remains on the top! Then the bed is lowered by 1 mm (or whatever your Z resolution is) and the liquid plastic automatically flows in from the reservoir to fill up the empty space. Then we sinter the next layer (probably with a laser). This is repeated until the object is complete. Now we can simply remove the object as the liquid automatically separates from the solid model. Then it is only necessary to raise the bed back up to the 0 position and push the liquid plastic back into its reservoir.
Bob Morrison
Wörth am Rhein, Germany
"Luke, use the source!"
BLOG - PHOTOS - Thingiverse
|
Re: Concept for negative-Z 3d printer March 18, 2010 10:28PM |
Guys, check out this stuff here on photopolymerizable resins:
[forums.reprap.org]
-Sebastien, RepRap.org library gnome.
Remember, you're all RepRap developers (once you've joined the super-secret developer mailing list), and the wiki, RepRap.org, [reprap.org] is for everyone and everything!
[forums.reprap.org]
-Sebastien, RepRap.org library gnome.
Remember, you're all RepRap developers (once you've joined the super-secret developer mailing list), and the wiki, RepRap.org, [reprap.org] is for everyone and everything!
|
Re: Concept for negative-Z 3d printer March 19, 2010 01:56AM |
buzzard192 Wrote:
-------------------------------------------------------
> @Eagleton - Interesting idea. In order to use the
> play-doh as support material, it would have to be
> fluid enough to flow into the holes that are being
> made (instead of the teardrop holes). I'm not
> sure if play-doh can be made fluid enough to flow
> into a model and remain rigid enough to support
> the piece as a whole. Perhaps it can...
>
> The other concern I have is if you are pushing the
> model into the play-doh after a layer extrusion,
> wouldn't the top layers of plastic still be soft?
>
>
> Also, the major fluid in play-doh is water. How
> will the play-doh react to having hot plastic
> (>100 Celsius) extruded on top of it?
>
> Just some questions to ponder...
>
> Buzz
Dunno. I don't imagine it would catch it on fire really, just bake somewhat. The other ingredients are flour and salt, some binders, and preservatives. It's non-flammable unless dry, and that amount of plastic would probably cool very rapidly in contact with the water, which has a high specific heat. I don't have a Mendel yet so I don't know the details as far as how hot things get and stay, but I imagine the object would be cooling off much faster than normal plastruder-built models do as it is in full contact with the surrounding mixture. Is it a design consideration that the plastic has to remain somewhat molten for the next layer to fuse properly, or can you stop and start up a 'cold' project later? If the latter, it would be possible to time the compression so that the soft plastic isn't deformed, but obviously that would lengthen the time required to "move upward"
@Sebastien, thank you much for the link! The more I look at this place the more I feel at home Grew up reading Science News, SciAm, Discover, etc., transhumanist/singularitarian with two geologist parents, too many interests for my own good, and completely sick of the current education system and job competition. I'm going to look at what's been done on the software side. I'm wondering if anyone's made a network-friendly frontend for Skeinforge - I have the strong itch to start a business that sells people time on my (to be built) printer remotely, with a webcam view of the action and a 'help' button that lets me know if there are problems via cell :3 I'd ship whatever they make to them as it comes out. Big step up from janitorial work, and I wouldn't have to make much doing it to get my current average monthly pay, heh.
RE: Play-doh/polymer clay (which might be more suitable) deformation measurements, work interfered with my time as usual. Will post when I have them.
-------------------------------------------------------
> @Eagleton - Interesting idea. In order to use the
> play-doh as support material, it would have to be
> fluid enough to flow into the holes that are being
> made (instead of the teardrop holes). I'm not
> sure if play-doh can be made fluid enough to flow
> into a model and remain rigid enough to support
> the piece as a whole. Perhaps it can...
>
> The other concern I have is if you are pushing the
> model into the play-doh after a layer extrusion,
> wouldn't the top layers of plastic still be soft?
>
>
> Also, the major fluid in play-doh is water. How
> will the play-doh react to having hot plastic
> (>100 Celsius) extruded on top of it?
>
> Just some questions to ponder...
>
> Buzz
Dunno. I don't imagine it would catch it on fire really, just bake somewhat. The other ingredients are flour and salt, some binders, and preservatives. It's non-flammable unless dry, and that amount of plastic would probably cool very rapidly in contact with the water, which has a high specific heat. I don't have a Mendel yet so I don't know the details as far as how hot things get and stay, but I imagine the object would be cooling off much faster than normal plastruder-built models do as it is in full contact with the surrounding mixture. Is it a design consideration that the plastic has to remain somewhat molten for the next layer to fuse properly, or can you stop and start up a 'cold' project later? If the latter, it would be possible to time the compression so that the soft plastic isn't deformed, but obviously that would lengthen the time required to "move upward"
@Sebastien, thank you much for the link! The more I look at this place the more I feel at home
Grew up reading Science News, SciAm, Discover, etc., transhumanist/singularitarian with two geologist parents, too many interests for my own good, and completely sick of the current education system and job competition. I'm going to look at what's been done on the software side. I'm wondering if anyone's made a network-friendly frontend for Skeinforge - I have the strong itch to start a business that sells people time on my (to be built) printer remotely, with a webcam view of the action and a 'help' button that lets me know if there are problems via cell :3 I'd ship whatever they make to them as it comes out. Big step up from janitorial work, and I wouldn't have to make much doing it to get my current average monthly pay, heh.RE: Play-doh/polymer clay (which might be more suitable) deformation measurements, work interfered with my time as usual. Will post when I have them.
Sorry, only registered users may post in this forum.