Formlabs Form 3

(Jon Gritton) #1

Just when you thought it was never coming, it’s here!

And it’s Huge big brother, the 3L! Let’s hope they’ve also solved the resin sensing issues…


1 Like
(Whosawhatsis) #2

It’s an interesting design. Probably the most innovative thing to happen in laser-SLA for a while, though that’s kinda damning with faint praise.

I like the parabolic mirror they’re using to linearize the galvo axis. The nonlinear resolution of galvos has always bothered me. The double-bounce system sorta seems like it’s asking for problems, but it was probably necessary to keep the size down. Limiting the galvo’s range of motion may also have been a concern, since even such a small mirror will have some angular momentum as it turns at high speeds to cross the build volume. I’m sure the LPU is a sealed unit, probably with (expensive) user-installable replacements available.

One really interesting thing is that the change from a pure galvo system to a hybrid galvo-linear system will force them to rely more heavily on raster scanning. DLP is pure-raster, while laser-SLA usually has vector outlines around a raster-filled area. The vector outline gives you near-perfect surface smoothness in the X/Y plane (we benefit from the same thing with FDM, which would otherwise have far-worse surface quality), while DLP and other pure-raster systems essentially show a layering effect in all three axes. The layering effect in X and Y can actually be significantly worse than Z in many of these cases, and this system looks like it would be even more so. It’s still possible to do a full vector outline, of course (assuming that galvo axis still has positioning control like it appears to, and isn’t just a mirror wheel synced to timed laser pulses, which would be much cheaper and probably faster, allowing them to step the X axis in smaller increments), but it would be extremely slow compared to full-raster mode, especially for complicated part geometries. It’s hard kinda to see how this won’t be a big step backward in print quality, speed, or both (though skipping the peel step buys them a lot in the speed department).

They will also be limited to raster scanning in one direction rather than alternating. This could present issues for part isotropy, but probably not, especially given that they’re probably overlapping laser passes significantly to deal with the raster resolution issue (especially if it is using the mirror wheel approach). I could see this potentially causing warping stresses in the print, though.

1 Like
(Jon Gritton) #3

Good breakdown! The optimistic side of my brain is saying that they must have put a lot of thought and effort into this version so it should be better, the pessimistic side is worried by it being too great a change, rather than just fixing some of the issues with the Form 2. I had a Form 2 for about 6 months and when it worked it was wonderful, but after a few months it began to produce constant errors detecting the resin cartridge and/or the print tray. The suppliers tried new parts but the errors continued so in the end I returned it for a full refund. It looks like the resin cartridges are the same as used in the Form 2 (which would make a lot of sense, stock/production wise) but would also mean they could be open to the same problem.