I've done a few more rounds, and some further insight:
* Perfect focus may be suboptimal-- at least with this 'raster' laser-- you can see the lines between rows with some burns, but if you slightly defocus the laser, it disappears. Ramping up the burn time might help too, as that likely provides enough heat to expand the area ablated. Someone with a more professional laser might have some interesting research options.
* The dye reproducibility is poor. Dump 20 identical caps into the same batch, and some will come out notably lighter. It seems to be very prone to "dark touch surface, lighter sides" effects too, making the caps look sort of like cheap no-name "five side dyesub" sets. Maybe a longer dye soak or a more precise dying process could help to "fill out" the colour, but I worry there might be a point where the dye "overpowers" and you end up shooting past the expected colour, or just seeps in deep enough that it can no longer be reliably ablated.
* The level of laser output required for a good burn through varies *hugely* depending on the keycap stock and dyes. Some require just 10-12ms (and seemingly, they're really prone to smelling awful as it engraves) while others require 40-60ms. Overburning sometimes results in worse results-- probably charring the underlying material, or grunging up the texture It seemed like for some caps, you'd even see different performance from top-row versus other rows, or modifiers (maybe they were manufactured differently even if they came in the same bag). After some sacrificial effort, I've ended up putting the extras in bags marked with notes to improve reproducibly.
* Front legends are an option here, if you can find a good jig to mount the cap. Since I have a regular keyswitch mounted to my engraver, I mount a sacrificial cap to that, and use masking tape to attach the one I want facing front to one side.
* Beyond a certain point, running two burns of shorter duration-- like two burns at 20ms-- may produce cleaner results than one longer shot. This seems to produce a less ragged edge. Again, part of the experimental cycle.
* The melamine sponge does help to an extent in that it adds a matte factor to the overall cap surface that reduces "competition" for the contrast of the engraved characters. Some caps have a prominent grain that "sparkles" and reduces contrast.
* Dying over a non-white stock cap works to an extent, but you have to expect even lower contrast. It's never as good as doubleshot and that's even worse.
* The engraver mechanism may have resonance vulnerabilities. I found I had a lot of issues with an uppercase "M" getting distorted even after several runs-- I suspect that it was jerking the engraver head in just the right way it keep it out of alignment when the next pixel was due. This smells like a part working loose, and the whole thing is held together with the one size of Allen bolt that I don't have a driver for.
I suspect this mostly fits in a narrow economic niche.
For a full set, a proper dyesub offering like Yuzu is probably better. Yes, if you've got the process down, you could buy a $20 set of PBT blanks and a $10 worth of Rit and produce a more-or-less workable output, but the results will be B-tier and it requires a lot of handholding and tweaking.
But if you need a specific mix of rows and profiles that's widely off from what's available in vendor catalogs-- like the 25 nonstandard caps in my 130% design-- ordering from a custom service may be uneconomical-- you might have to either overbuy to get the desired coverage or buy a bunch of individual caps at $3-5 each. Similarly, it might be interesting for things like meetup souvenirs where you could bring the engraver and make caps on demand. If you had a solid supply chain and a good way to vent the fumes
Recent Posts
