More to come...

I ordered 25 LEDs, lost/destroyed the first couple, then tried to connect them to a dead-bug shift register (top right). During testing I fried a bunch of the LEDs, and the Dremeled trace work wasn't consistent enough to really function, so I tried again (bottom). Dead-bug plus 0.65mm pin spacing was tough to get working. Also shown: tiny connector!

Time to shift gears, skip the shift register (note the attempted dead-bug ATtiny88, also abandoned), and mount the microprocessor right-side-up. X-acto trace time!

This chip is a bit more forgiving at 0.80mm, but I still had trouble with alignment. To make sure the pins lined up, I soldered the device halfway to the board and carefully cut the remaining traces using the installed chip as a guide. Not shown: an unbelievable number of continuity tests:

After finally getting the damn processor to listen (more continuity checks!), I toggled the LEDs on and hit a few snags. One was wired to the RESET pin (dumb) so I temporarily disconnected it, and another had a short. But more importantly... one was also the wrong color:

The LED worked but the single yellow element really bothered me. I was (and still am) also bothered by the fact that I pulled all the LEDs out of the same tape, meaning there was a sorting issue at the factory. At this point the LEDs I hadn't soldered down had pretty much all evaporated off my workbench due to moving air currents, so I had to desolder one from the other ring (which took 5 tries to get right). Finally done:

To do: dendrites, axons, more indicators, and a bit more time defluxing. Also programming, I suppose.