The Possibilities of 3D Printers

[djolds1]

Cut from the Mach-Effect thread on the News sub-board:

I've had my eye on 3D printing since it was called laser sintering and applied to the ceramics we're always interested in. The problem is that in all likelihood, the geometry we would use is based upon many, very-thin layers of ceramic separated by electrodes. 3D printing only applies to items constructed all of the same material.

Not true. Multi-material printer heads have been around for going on a decade now. But THIN layers imply vapor-deposition - a somewhat different process than using ink-jet heads or laser sintering. OTOH, vapor-deposition should be integratable into an all-up fabber.

But yeah, we even identified the kind of laser sintering machine we'd need to do our own in house work. Lock-Mart has one. they're $1 Million each. Only superficially like your $500 3D printer.

Bootstrap up to an equivalent starting from a RepRap or similar? I do love the ISRU philosophy.

Practical 3D printers don't need to produce bleeding-edge; they just need to produce good-enough. They can extrude fiber-reinforced polymers or even amorphous metals as structural basics, for one.

In addition to RepRap, I almost think there needs to be an Open-Source Everything project for simplified but robust and capable versions of other modern basics.

Open Source Operating System
Open Source Hardware
Open Source Medical (a potentially VAST topic)
etc.

[GIThruster]

Cut from the Mach-Effect thread on the News sub-board:

. . .THIN layers imply vapor-deposition - a somewhat different process than using ink-jet heads or laser sintering. OTOH, vapor-deposition should be integratable into an all-up fabber.

That was the conclusion I came to in a white paper in 2008. There's no reason to not combine laser sintering with CVD. The problem is that CVD relies upon the proper chemical affinities between materials in order to work at all, so you're restricted in the kinds of materials you can use. However, in order to push the boundaries of what is possible so far as thin layers and high frequency, CVD is likely the answer.

[kunkmiester]

Here's the project I mentioned in my post:
http://www.reprap.org/wiki/MetalicaRap
They're trying for micrometers of resolution, not sure how likely that'll be. All you really need is .0001 inches though, anything beyond that is specialized--most bearings, optics, etc. that would probably have their own fabber.

One of the holy grails of 3D printing is metallurgy--printed parts don't have forged metal strength. Being able to control crystal structure like that is years away, but will pretty much make everything else work out.

One way I was seeing to do multiple materials was an inkjet type head that carries solutions that can be boiled off, leaving the dissolved/suspended material.

The other, which is probably more applicable here, would be controlled ion deposition, where a beam of ions is fired, grounding and attaching to the target where desired. This could use just about any material that can be ionized, provided it will actually attach to the part. Was thinking a microchip fab could use it instead of lithography, eliminating the chemicals and stuff. Not sure if it'd even work well enough though. Too poor to do much, even if I had all the brains needed(failing calc doesn't help).

We can compare this to the information revolution. Reprap is an Altair(if that), Stratasys makes IBMs. The generation that will really get fun is when Pentiums break a gigahertz.

[ladajo]

One of the holy grails of 3D printing is metallurgy--printed parts don't have forged metal strength. Being able to control crystal structure like that is years away, but will pretty much make everything else work out.

How about printing it, then cooking it?

[GIThruster]

I'm pretty sure all the stuff used so far is thermal epoxy based. Whatever it is, it's not like bringing a metal up to melting point and letting it cool. I suppose it would be possible one day to do this with a liquid metal or amorphous glass alloy, but like all such alloys it would be allergic to heat. Such amorphous glasses fracture easily as well. They're about the last thing you'd make a gun part out of.

You can check but it's not just that epoxies don't have the necessary strength. IIRC the real trouble is they lack the hardness to survive wear and tear. Polymers are amazing things though, and I'm sure there's a huge push on now to improve 3D materials.

[ladajo]

I have seen they are printing and firing some fairly exotic ceramic mixes now. I am thinking metals can not be far behind.

Replicators in the kitchen are soon to be me thinks...

[Betruger]

Can't happen soon enough.

[GIThruster]

There are lasers capable of melting metal. I'll bet there are plenty of folks working on it right now.

[kunkmiester]

djolds mentioned selective laser sintering. You can make parts out of metal, in fact most articles on "additive manufacturing" mention compaies using the process to make production parts. The problem is, even if you manage to completely melt the metal to fuse it together, you're getting lower quality than a cast part. The two advances metalicarap offers are being able to scan the printed layer and make cuts to correct errors from the process and the powder size, and provide enough heat to actually melt the metal to get a higher quality part.

You still have little control over the metallurgy--what you get is a cast part, period. Until you can manipulate the crystal structure the way forging and rolling do, the process is of limited utility. Heat treatment is of limited utility here, there's a reason wrench sets brag about drop forging.

The second grail is multiple, disparate materials. Say, a circuit board, but you're printing out the substrate, layering traces in it, and building up the components that go on it, all in one machine. The key there is finding low energy deposition methods for the various materials, so laying down the copper doesn't melt the plastic base among other issues.

The third is microchips, which ties in with precision--modern chips are at what, 40 nanometers or so? Building devices at this level is so much different than anything else, but sets off the true industrial revolution. A reprap can make some of its parts, a metalicarap could probably do even more, but they both need computers, which neither can make.

Some people don't quite realize that you don't actually need one Star Trek Replicator that can churn out anything--multiple machines can be used, as long as the shop is reasonably small. That's one reason I see open source as the group that will make the big strides--for now plenty of stuff is too easily mass produced too cheaply to appeal to the commercial side, until it's been proven to work.

[djolds1]

djolds mentioned selective laser sintering. You can make parts out of metal, in fact most articles on "additive manufacturing" mention compaies using the process to make production parts. The problem is, even if you manage to completely melt the metal to fuse it together, you're getting lower quality than a cast part. The two advances metalicarap offers are being able to scan the printed layer and make cuts to correct errors from the process and the powder size, and provide enough heat to actually melt the metal to get a higher quality part.

You still have little control over the metallurgy--what you get is a cast part, period. Until you can manipulate the crystal structure the way forging and rolling do, the process is of limited utility. Heat treatment is of limited utility here, there's a reason wrench sets brag about drop forging.

The second grail is multiple, disparate materials. Say, a circuit board, but you're printing out the substrate, layering traces in it, and building up the components that go on it, all in one machine. The key there is finding low energy deposition methods for the various materials, so laying down the copper doesn't melt the plastic base among other issues.

Agreed in all particulars about quality limitations and range of materials. That's why we don't need perfect - we need 'good enough.' Perhaps with an intentionally restricted range of raw materials inputs.

The third is microchips, which ties in with precision--modern chips are at what, 40 nanometers or so? Building devices at this level is so much different than anything else, but sets off the true industrial revolution. A reprap can make some of its parts, a metalicarap could probably do even more, but they both need computers, which neither can make.

You're thinking too far along the development curve. Think the equivalent of 286 chips or even 8086 chips - get the basics and work up from there.

Some people don't quite realize that you don't actually need one Star Trek Replicator that can churn out anything--multiple machines can be used, as long as the shop is reasonably small. That's one reason I see open source as the group that will make the big strides--for now plenty of stuff is too easily mass produced too cheaply to appeal to the commercial side, until it's been proven to work.

Excellent point. And that is something I've realized for awhile. Multiple dedicated modules around a common construction crucible seems workable. Each module as compact as practical.

[hanelyp]

Some things you can do to control metallurgy with laser sintering are:

- how much the metal melts, from a partial melt that sinters the particles together, to a full melt that more resembles building up metal with welding beads.

- how fast the metal cools, and how much time crystals have to grow.

Nano-crystalline forms of some alloys are reported to have some very good properties. The fast cooling possible from a very thin melt layer may be good for this. But it would require a slower build to keep the bulk cool.

[paperburn1]

Some things you can do to control metallurgy with laser sintering are:

- how much the metal melts, from a partial melt that sinters the particles together, to a full melt that more resembles building up metal with welding beads.

- how fast the metal cools, and how much time crystals have to grow.

Nano-crystalline forms of some alloys are reported to have some very good properties. The fast cooling possible from a very thin melt layer may be good for this. But it would require a slower build to keep the bulk cool.

http://www.substech.com/dokuwiki/doku.p ... on_process
Melt metal into ceramic, problems solved

[DeltaV]

Better quality resin objects at a lower cost:
http://formlabs.com/

Form 1 can print layers as thin as 25 microns (0.001 in) with features as small as 300 microns (0.012 in) in a build volume of 125 x 125 x 165 mm (4.9 x 4.9 x 6.5 in).

[choff]

On one of last visits to my dentist he used a 3D printer to make a ceramic crown, it was interesting to watch. Lots of little niche applications for them to turn up in.

[Betruger]

Lots of little niche applications for them to turn up in.

Absolutely... the sooner kids get their hands on these..