Hacker News new | past | comments | ask | show | jobs | submit login
Casting Glass from 3D Printed Molds (amosdudley.com)
147 points by dezork 78 days ago | hide | past | web | favorite | 25 comments

Also see 3d printed aluminum casting molds:


Myfordboy is an excellent mechanic and engineer, and one of my favourite youtube channels.

VOG [1] runs another channel that does a lot of (artistic) metal casting with mostly 3D printed positives, called "lost PLA casting".

He uses casting plaster and built his own vacuum chamber and furnace to improve the quality of his casts.

[1] https://www.youtube.com/channel/UCkEYj8wtK3aEW8vSGhlB43g

I taught a similar course at Pilchuck in 2015, which was the first 3D-printing/digital fabrication course taught there, and was also up there taking a flameworking course when Amos was TA'ing for Yoav and Angela.

First off, Pilchuck is amazing, and if you're at all interested in glass, you should seriously consider taking a course there.

Second, modulo a few tricks for clean burnouts, if you can print it, you can investment-cast it in whatever material you want. It's amazing how powerful cheap printers become when you combine them with a little casting skill - you can access nearly arbitrary geometry in some useful, high-performance materials without a $100k printer.

Do you know what kinds of results are possible with a standard consumer 3D printer? (A PLA printer?) The page says, "Any layer lines in the print end up transferring through to the glass, which makes polishing a lot more difficult," but does "a lot more difficult" mean impossible?

(I have an Ender 3 printer, and a microwave, so it looks like I can possibly get started for <$100. But both because of price and turnaround time, I am not very interested in using online resin printing services.)

Yes, your surface finish is limited by your print quality - but smoothing surfaces post-print goes a long way. "a lot more difficult" doesn't mean impossible at all, but you will have to remove more material to flatten the surface before you can start polishing. Coldworking (cutting and polishing glass) is a field in of itself, and requires some non-trivial equipment to do efficiently, so I would recommend designing your initial pieces to avoid much coldworking. Don't expect an optically clear surface anywhere the glass touched the mold - the default for investment cast surfaces is matte, and clear surfaces either need to be polished with abrasives or HF+H2SO4, or were the top of an open-face casting.

Unfortunately, most kiln-casters prefer very high lead glass (40%+), because it flows very well and doesn't devitrify easily. This, however, makes coldworking an environmentally-dicy proposition, as it generates lots of micron-scale lead dust.

A microwave may suffice to fuse small glass jewelry (...kinda, the annealing will be pretty horrid...), but isn't going to cut it for casting. Stuff at the scale presented in the article would require firing times of at least 12 hours, with much better temperature control than a microwave can provide.

I've got no practical experience with 3D printing, but surely dremel + grinding/sanding bits = no more layer lines?

Yep, sanding is an option. I've also seen people fill the valleys with some sort of goop filler and then start sanding from there, just to get it a little closer to smooth. But I can't say if introducing different materials like that might interfere with casting. They were doing it to paint over.

Another approach is vapor smoothing - you can use acetone vapor to smooth ABS prints. Acetone won't do anything to PLA though.

I've heard of other chemicals being available for smoothing PLA, but they're nastier than acetone and I don't know anyone who's actually tried them.

Wax is a very good material for modifying/smoothing prints pre-casting, and doesn't mess with the casting process. UV-cure resin is also a good choice, as you can set it fully, sand it down, and apply a gel-coat for a shiny finish.

ABS is a very bad material for casting if you're planning on burning it out, but vapor smoothed parts do make very nice plaster molds if you can remove them. Chloroform works very well for PLA, but it is much much nastier.

You should be able to use acetone. See the solubility table [0] for PLA. It takes longer, but it works.

[0]: https://onlinelibrary.wiley.com/doi/pdf/10.1002/app.38833

Acetone really don't work well for PLA. You need tetrahydrofuran to smooth PLA in the same way that acetone smooths ABS.

Hi! Did we meet during that Pilchuck session? I wonder if I remember you.

Either way, feel free to send me an email ping if you like, I'm always looking for folks to share knowledge with.


Glass is definitely my favorite material to work with. I started messing with it after following the instructions for setting up a lab in “The Golden Book of Chemistry Experements.” (I can’t believe that’s a children’s book haha)

> Once you have a model, the problem you need to solve is how to make a model with no undercuts. Undercuts will make it impossible to remove the 3D printed positive

Is there a mathematical term for solids with this property? It reminds me of the vertical line test, but in 3D.


I'd call it "monotonically expanding" along some particular direction vector.

In die casting, you have to do a little better than the vertical line test and stay within the specified draft angle. A vertical wall would be a 0 degree draft angle, and not suitable for casting most materials.

I would call it self-occluding.

Convex perhaps?

There are non-convex shapes that don't have undercuts. I think convexity implies that you don't have undercuts, though.

This is great stuff!! As a hobbyist glassblowing and 3d-printer, this is exactly the kind of content I’ve been looking for. Thanks!

I did some pate de verre lost wax casting to make glass armguards a while back but they turned out too rough to wear. 3d printing might be the answer! (Check out my attempt here: https://www.instagram.com/p/Bk45IFqBhbL/?igshid=16fgxi4oubho...)

I'm curious about the too-rough-to-wear bit. Would an extra (leather/plastic) layer below the glass be impossible?

Pâte de verre is typically rougher than kilncasting from a larger billet, because you're fusing glass granules without fully melting them together. So you could get a smoother surface just by adjusting the peak kiln temperature.

But yea, certainly you could add a leather wrap underneath.

Here is another artist who uses the technique.


This is fantastic. I've been wanting to make some custom liquor bottles. One step closer :)

This would be fun to get with our company logo in the last photo of the snake head. If only we weren't paperless and could use a paper weight...

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact