Metamolds: Creating ideal 3D printed silicone molds faster and cheaper | Hacker News Reader

Metamolds: Creating ideal 3D printed silicone molds faster and cheaper

3ders.org

71 pointsceoloide29 comments

It would be great to try the algorithm out if they ever release the code. I've made molds through "traditional" CAD software like Solidworks and the process takes about 30 min. It works pretty well for the typical stuff but I have not tried it with something with a lot of organics shapes like a figurine. The biggest downside about silicone casting though is that its super messy. The costs adds up too since you either need a small vacuum chamber for the silicone or you have to buy the more expensive variant with platinum in it.

Side note: If any one is looking for a lowcost (~$500) "SLA" (it's really UV DLP) printer, you can go for the Anycubic Photon[1]. It should be good enough to make a master pattern without the visible layer lines. A caveat though, the resin based prints are super messy and is a pain to use.

[1] https://all3dp.com/1/anycubic-photon-3d-printer-review/

I'd be concerned about layer lines ending up in my final molds. If you look closely they'll end up in your final cast repeatedly. Perhaps this would be best accomplished with a resin type 3d printer to minimize the issue.

You've got several options here, none of them great.

1) Use a resin printer and print the meta-molds at very high resolution, then lightly sand with high-grit paper or use a gap filling primer paint.

2) Use ABS, HIPS, or another acetone smoothable plastic and vapor bathe the meta-molds (would probably slightly mess up the split lines though).

3) Use a paint-on epoxy to manually smooth out the meta-mold interiors.

4) Manually sand as best as you can.

The normal process for this would be to print the 3D model you want and then do a bunch of post-printing cleanup and treatment to get it to mold-ready quality, then cast a silicone mold of that. It's not super fun, but you can get good results without years of practice.

Re.: 1) The latest generation of 'cheap' DLP printers can do 25 micron (e.g. the Phrozen Shuffle). After curing, dipping in resin again or priming and then coating with 2-3 layers with an airbrush results in a perfect surface finish. No sanding required.

Are you actually getting a perfect surface finish that way? My Phrozen Make printing at 20 microns leaves noticeable lines and definitely requires cleanup around support attachment points. I tried doing the second dipping method and ended up with melty details on finer parts of prints. The only thing I haven't tried is sandblasting.

The Make can't do 20micron. It can do about 47 max (or rather, min), I think.

Re-read my post: I was talking about the Phrozen Shuffle which has a different screen/display unit.

Check the FB group for some 27micron examples done with the Shuffle.

> 25 micron

Hey where did the 0.4 micron go?

5)Have one of the $40k clip printers that don't leave the lines.

Heh, true. But if you've got a $40k printer you're probably sending out for professional molds already.

The paper authors used FDM machines (Ultimaker 2+, a Stratasys J750, and a Stratasys Fortus 450 MC). But even the cheapest $500-ish resin printers can do 25 micron layers with 50-ish micron XY resolution and it is very very hard to see those layer lines.

FDM machines are almost invariably resin printers; ABS and PLA are the most common resins for them to print in, although others exist. In theory you could do FDM in non-resin materials such as, for example, metals, but I don't know of any examples.

Your parent comment is talking about DLP/SLA printers. Though you are technically correct about FDM being a resin printer, it's not the way hobbyists talk about it.

Maybe hobbyists should say "SLA" if they mean "SLA". This is like people who say "wheat bread" as the antonym of "white bread" — what do you suppose they think white bread is made from? (And it sets them up to be ripped off, as there plenty of manufacturers out there who are happy to sell them "wheat bread" made from 75% white flour.)

Lost wax casting.

Post printing a very small amount of time with sandpaper (or acetone for certain materials) would smooth things out. It wouldn't work well for small details though.

Where can I get this?

It’s a research paper so most probably there won’t be any software delivered. Sadly it’s not the norm for researchers to release the code or an user-fronting application when doing computing research (or any research). Open source exist but is even rarer.

There’s dozens of very useful and advanced creation tools created in labs that never turned into usable application. In some domains big players pick up on the research and include it as a feature, but many advances just stay in labs for ever.

The rare one that are released as usable applications by the authors are most of the time not supported and end up abandoned—quite similar from many non-research software in that respect, but a little more sad as they are often more unique, like for example the free-form plush creation tool from a Japanese lab I remember from several years back.

The most irritating is when tax-money funded labs don’t incentivize their researchers to open source their code, and we end up with a innovative software tool, closed source and with proprietary licenses, that is effectively only the research playground of one researcher, mimicking a product but never really turning into a product.

> It’s a research paper so most probably there won’t be any software delivered. Sadly it’s not the norm for researchers to release the code or an user-fronting application when doing computing research (or any research). Open source exist but is even rarer.

> There’s dozens of very useful and advanced creation tools created in labs that never turned into usable application. In some domains big players pick up on the research and include it as a feature, but many advances just stay in labs for ever.

I call papers like that "Show & Tell"—it's not science, it's just "hey look at this cool thing I did". Which, yeah, it's cool, but as a paper it's completely useless.

But 3d printing is not food safe, so using this method for chocolate molds seems unwise.

It can be food safe. You can even print invisalign-style braces.

https://formlabs.com/store/us/form-2/materials/dental-lt-cle...

PETG polymers, printed with sufficient fidelity are food safe, and available on FDM type printers.

Some is, I've seen a few decent examples using an Ultimaker [0] like Dinara Kasko's work [1]

[0] https://ultimaker.com/en/stories/19498-using-3d-printed-shap...

[1] http://www.sogoodmagazine.com/pastry-blog/pastry-chef-articl...

The molds are not 3D printed. Rather, the molds for the mold are. The actual molds used in production are silicon, obtained after pouring silicon into the 3D mold-of-the-mold.

FWIW mold for a mold is usually called a pattern

Pattern is the object you want to have produced. Metamold is a mold which gives you mold which then can produce your pattern.

Ah! My mistake. Thank you.

The 3-D printed molds are used to make silicone molds. It doesn’t really matter how bad the 3-D printed molds are as long as they don’t ‘infect’ the silicone molds while you’re making them.

The video shows it being used to make fancy ice cubes.

Can you make silicone sex dolls with that?

http://makerlove.com/

Reply on news.ycombinator.com