Believe it or not, I think this pig may actually fly! Things are starting to get interesting. After several more days of wiring, I’ve reached the point where I can execute straight line code (no branches or jumps), with no RAM, and no registers. If you think about it for a moment, you’ll realize that given those restrictions, you can’t really do anything with the computer at all. There’s only one piece of state (the current program address), and there’s no way to change that state other than by sequentially executing instructions. It may even be a stretch to say that it’s “executing” instructions, when they don’t change any state. All the instructions might as well be NOPs.

Despite the outwardly boring appearance, I’m actually very happy with this result. The ability to execute straight line code means:

• The program counter works, because it steps through the program instructions correctly.
• The address bus is wired correctly.
• The address decoder works, because it enables the boot ROM for addresses that are mapped to it.
• The boot ROM is set up correctly, because program instructions are being read from it.
• The external data bus (memory bus) is wired correctly, because instructions are transferred on it to the opcode register.
• The microcode and control system work (although I knew this already from earlier tests).

I didn’t exactly stress test it, but everything seemed very solid and reliable during my experiments, with no weird glitchy behavior at all. The couple of signals I examined with the oscilloscope looked pretty clean.

From here, it’s only a couple more steps before I have something interesting working. I think it’s time to start thinking about my minimal definition of a computer, so I can pinpoint the date of the first successful boot. If I had conditional branching, and two registers, that would be enough to write a simple program to compute a factorial or some such. The result would need to be read out with the logic analyzer, since there’s still no human-readable output, but it would be good enough for me to declare BMOW officially up and running. If all goes well, I should be less than a week away from that goal!

Mmmm, 24 bits of delicious address bus.

I also managed to cram in a ZIF socket for the boot/program ROM, and a shiny red reset button.