C64 mini is very popular and I'm envious. Because nobody is going to make a cute little Vector-06c, I'm going to build one myself. Today the original ones in good condition are very rare. Wikipedia has a summary of what it is: https://en.wikipedia.org/wiki/Vector-06C
I have recreated its logic previously in both FPGA and software. The main challenge of this project is to also create a somewhat faithful reproduction of the physical appearance. What exactly is going to represent the functional part is TBD.
Had to take a picture because I have a bad feeling about the decals...
Top to bottom: RAL7034 (online swatches look nothing like it, I think it was a cruel joke from the colourist, but they warn you that you need real swatches so joke's on them), RAL1019, RAL7032.
Now that I have such a beautiful stencil, the rest should be easy, just spray paint through it. I remember from some previous experiences that spray particles would try to reach behind the stencil and that would make the edges fuzzy.
For the first test, I used a couple of drops of silicone glue to hold the stencil as close to the key as possible. The result looked promising:
However I didn't like using the glue. I decided to stretch the foil across the keys using two test prints as supports. Here's just a nice picture of sun projecting through the stencil:
(We have sun)
I used some tape, perhaps some other things, but the foil would never stay put. In the end I decided to just use some weights (don't put too much weight on it).
I tried to make the coats as thin as possible using a spray can. After 5 or 6 coats with some 5-7 minutes between I decided that enough is enough and it's time for the grand reveal.
The stencil prints always have these annoying connecting bits which I also had to add, or the stencil would not hold itself together. I touched them up using acrylic white paint on a tip of a smallest brush that I could find.
I'm going to seal it with transparent lacquer when it's fully set. If I were to try it again, I'd account for the fuzziness and use a thinner font.
As it turns out, such thing exists. I wasn't going to buy such an exotic thing for just a few key legends though. But it shows how desperate people can get when solving this seemingly trivial problem. Come think of it, all printing tech around you is built around one fundamental principle, you take away the colour. If you begin with white, you can have everything. If you begin with the brown however, you can make it a bit browner or black, but that's it. Toner transfer will not work here. So what can we do?
Here's a brief summary of technologies that can be used to transfer white on black.
Screen printing / Silk screen. Possible to achieve very nice results, used professionally, including for printing legends on (flat) keycaps. Requires a set up of a shop size of Ben Krasnow's hangar and a metric gallon of specialized chemicals.
Pad printing. Another amazing tech which gives good results, especially on curved surfaces, and that's why it is frequently used for making keycap legends. Unfortunately not possible without some pretty specialized equipment and chemicals.
Laser ablation. Make a white key, coat it in black paint, use a laser to evaporate the paint. If I had a laser engraver handy, I would definitely try this one.
Woodblock printing, linogravure, etc. Beautiful antique arts well represented on youtube. Very inspirational, but I don't have skills to produce the woodblocks of reasonable quality at the miniature scale.
It's worth noting that a local souvenir printing service would be happy to pad print a few keys (for the price of 1000) for me, but it's not sportsman-like. Another constraint is that this being 2020 I prefer to use what I already have at home or can have delivered to the door. So I had some pretty imaginative ideas.
Apply film photoresist to the key surface, a thicker layer than usual. Expose so that the text glyphs are washed out, forming pits. Fill the pits with thick acrylic paint. After the paint will have fully set, use a hard solution of NaOH to remove the remaining resist. Did not work because my roll of photoresist was 12 years old. Nice try.
Apply spray can photoresist, the rest would be the same as (1). Amazingly, not only the spray can variety also have expiration dates, the cans themselves may develop holes in them and that's nasty. Did not work out.
Perhaps a UV-curing solder mask would work? I had some red one, but I could order a tube of white mask if it works. Unfortunately this mask is essentially an oil paint, and it would not work well with previous coats of paint. Probably there are other variants that are worth investigating?
UV nail polish. A strange choice, but it's available. It sounded like a plan and nothing could fail. There's a catch though. These things are UV-opaque. They are designed to form a kind of film on top of your nails, but under that film they don't really cure. Another problem is that I'm using a contact method of photo template and UV nail polish sticks really well to the printer transparency, but, because it's UV opaque and does not cure in depth, not to the target surface. So close.
Found a roll of copper foil that I remember finding in a closet when I was a little child. Unaccountably I've been holding on to it for all these years. Let's see what we can do.
So a roll of copper foil, a ribbon rather, about 2 cm wide. Probably same thickness as used on a PCB substrate, maybe a bit thicker. From this pint on, it's almost as if it were a PCB. It's a foil without substrate, so I took a piece of 1mm thick modelling cardboard. Wrapped the ends of copper ribbon around the edges, taped them down. Then took a printed template (in negative, the letters would be etched) and wrapped it around, secured with a bit of tape again. Ran through the laminator. The other side of the foil needed to be protected so I used a thick coat of PLASTIK spray (acrylic) to cover the reverse side. Soaked the paper in water, peeled the paper.
The keys don't really look proper unless they have legends. I have tried a few techniques so far.
This is a picture of two interesting tests.For the reference, the key pitch here is 10mm. So the key tops are roughly 8mm across.
The bottom one has the legends embossed in the design, then printed normally in white PLA on Ender2 with a regular 0.4mm nozzle (although I think I had extrusion width set to 0.3 for this test). Then I sprayed it over with a black paint and sanded down the tops. It's definitely usable, but it's brutally ugly.
The top one in yellow is more interesting and that's the method I'm counting on. The legends are printed on regular glossy paper, the kind you'd use for toner transfer PCB making. The keyboard and the paper are then sprayed with alkyd-based lacquer. A few moments after spraying, so that it's already not too runny but not yet hard, the paper is placed on the plastic, toner side down, adjusted (as you can see adjusting it is trickier than it seemed), and let dry overnight.
Then it's as if it were a PCB. The paper is soaked wet, peeled off, the remaining bits of paper are brushed off using an old toothbrush. There will still be some visible bits when the water dries, but here comes the magick with a k of another coat of transparent lacquer. It all becomes transparent and it looks amazing.
I make a point of using an alkyd-based lacquer because I also tried an acrylic variant and it uses a base that immediately dissolves printer toner.
This is really nice so far, but what about those brown keys? Black legends will not look good on them.
The C64 mini fake keyboard is a joke. My keyboard is going to be functional. However it's easier said than done. I've looked around, in fact I spent innumerable hours browsing for a passable switch that would work fine in a mini keyboard. I don't intend to compete with a real modern pc keyboard, but it should be usable for entering a small BASIC program, or doing some assembly hacking. At some point I decide to just call it quits and stopped on STPMAME010 switch by ALPS.
ALPS claim 160gf in their softest variant. 160gf is perhaps soft for a vandal-proof elevator button, but it's at least 3x stiffer than your typical keyboard switch. However this is the softest I could find. Upon arrival I measured the tripping force using IKEA scales. As it turns out, they trip at full 250g, but push back 160g. They have nice long travel though, and they're still softer than a typical black 6x6 button.
Another problem is that they don't have any stem. When pressed, the stump sinks in the case entirely. So the keycap should somehow sit on top of it, but have some frame that would not let it fall off. It turned out to be rather difficult to design a keycap that would sit freely, not fall off, and not bind when pressed, all at the same time. After some 50 attempts I think I have dialed in the dimensions.
The general shape:
And then there's a frame that sits on top of this and holds them all together. Small test prints show that this works and feels nice. We'll have to see how it works with the complete keyboard though.
It's worth mentioning another discarded idea, the photo of which was used for colour tests. In case I wouldn't be able to find a nice switch, the alternative variant was to have the keys permanently attached to the frame from the top side, using little compliant hinges. They would press on tiny flat buttons. The model proved that this would work and probably it will come in handy in another project. But compliance in a 3d print is not a very reliable thing, besides it would feel perhaps a bit calculator-like. I decided in favour of long travel switches.
For me one of the most striking visual aspects of v06c is the keyboard. A strange palette of greenish-beige colours that are pretty difficult to match. On top of that, even the keyboards coming from the same factory used to vary wildly in colours. Other factories used different plastics entirely, and would often use plain white keys.
I did some amateur colour matching, checking out what's available locally and came up with this list of colours, which is pretty close to my particular sample but may be very different for others.
Ordered the paints in epoxy variant, in 250ml spray cans. Time passed by quickly, and here's the colour probe on a mockup keyboard: The spacebar seems strangely off, I even suspect that there was some mistake at the paint shop. But it's also of the strange baby stool variety, so it's going to work.