Sequencer Construction: 8-cycle FSM
Note there that I did say sequencer 'cards' and not 'card' because there's two of them this time. This is the first part of the computer that will be constructed across two cards stacked together rather than all on a single card (the upcoming control unit will also be spread across two cards). This is for two reasons: firstly the sequencer needs access to more connections than a 'regular' card but secondly there'll be so much to fit in that it physically wouldn't fit on one card.
Working across two cards does present additional challenges in the construction ... mainly around how to get the required signals that are private to the sequencer between the two cards. I considered many options but in the end a system of stacking header pins seemed the best route forward. So, to start construction I soldered down the ribbon cable connectors and board interconnects ... once done it looked like this:
To stack the cards the upper card header pins are aligned and inserted into the respective sockets on the lower card producing a unit that looks like this:
For the LEDs I've continued using the newer method of soldering the LED cathodes together with bits of trimmed off diode/LED legs as it makes soldering much easier. This is the first time, however, that I've tried this technique on double height LED holders. Generally the concept stays the same ... join all the cathodes together ... but for the holders that have LEDs in their upper slot I need to add a small Kynar wire link to get that LEDs anode to the wire wrap post. It's all a bit fiddly but as long as you have patience and a steady hand it usually comes off OK.
Next up were the relay sockets and associated wire wrap posts:
Next job was to solder in the flyback and feedback diodes (which ensure produced outputs don't feed back into parts of the sequencer's finite state machine):
There's some temporary Kynar wires in the power and ground lines and these are just to 'hop' over future parts of the card where the lines will eventually go through. The Kynar wire, of course, can't handle all that much current but it should be fine for now.
With the soldering done it's on with the wire wrapping. The sequencer probably has the most complicated wiring yet (second only, maybe, to the ALU arithmetic unit).
That's it for the sequencer ... at least for now. It produces all the pulses that will be needed for the computer to perform copying values between registers, loading values from the opcode and performing ALU operations. The final step in making these operations a reality is to construct the control unit ... again though, just enough to operate these three 8-cycle instructions. The control unit is similar to the sequencer in that it will also be spread over two cards although fortunately it's wiring will be quite a lot simpler as most of it is just combinatorial logic.