This batch of clocks took nearly a year to complete, due to many factors, but ultimately due to poor planning on my part. I'm usually OK at designing for construction, but these cabinets were just so gorgeous I neglected to consider just how difficult it would be to fabricate them. It will probably be a while before I make any more clocks!
Usually the clock guts are no big deal, I tweak my existing code for the particular interface layout, but I finally decided to implement alarm functions. And improve the rotary-encoder software. And since I now needed more inputs than the poor little PIC could handle, I had to conjure up a new input scheme. And of course I exceeded the memory capacity of the PIC -- again -- and had to squeeze code again to fit. But it all worked out fine in the end, and the interface is better than ever.
And as always, the most time-consuming part is the design, which is pretty much impossible to document; basically I fiddle with layout on one of the chassis, and when I have a decent first approximation, and have more or less chosen the interface (pushbuttons only? rotary encoder? four or six nixies? alarm? what kind of switches? do I need to make any controls? etc) I start laying out in Illustrator, which I use to make the artwork for silkscreen or etching. This time I actually have a photo of an abandoned design; far too pedestrian for these lovely cabinets.
Laying out the front panel was the easy part here, even though these cabinets have no right angles anywhere; everything was hand-carved. The photo below doesn't really do justice; these cabinets are almost a cliche of 1950's googie design.
(These chassis were unused, but pre-stamped for some unknown product; they were still wrapped in the manufacturers yellowing paper, each apparently rejected due to bad paint, but I stripped them for repainting anyways. The prototype is black hammertone, each clock in this series is a different color with different knobs.)
Basically for a batch like this I make one as a "prototype" then duplicate it. It's not exactly mass production, and they get enough planning that the "prototype" is essentially identical to the batch. Here's the prototype minus silkscreening, plus another shot of it with it's siblings parts and pieces ready for assembly.
Here you can see the snooze switch perched on the brow of the cabinets. These are hard to photograph!
These cabinets are pretty generous inside, so I laid them out internally in the age-old style. (Note that I hand-wire the nixie sockets and such, not particularly efficient, unlike David's clocks (see the Cathode Corner). This clock is a perfect example why: the tubes are not in a straight line, they are aligned with the curve of the brow. Even the horizontal spacing of the Nixies differs between clock models, depending on what looks right. I know that if I had an "e;efficient" printed circuit for this, all my clocks would look the same... and I don't want that restraint. I don't even want to think about that at design time.