Today marks the final day of RetroChallenge 2025, one of the reasons I've been writing about this project. Perhaps it's a good point to take stock and sketch out a few plans for where we go from here.
The first thing to say is that whilst RetroChallenge is done, the project and blog won't stop. I'll continue to investigate and write up the Computox here, so stay tuned!
Where are we?
So, what did we achieve in October's 31 days of RetroChallenge? Well, hopefully I provided an overview of the Computox and we have a sense of its main components and features.
 |
| Computox components |
We've also reverse engineered the power supply and all the main peripherals; the controllers, main game display, game counter, winner indicator and bell. In each case a circuit diagram has been constructed and how the component interfaces with the logic core of the machine has been identified.
Talking about the logic core, we've put together an inventory of the relays and reverse engineered how the current state of the game board is stored, latched and reset for both players. Whilst this is a good start it's clear there's a sizeable amount of work still to go in mapping and understanding this beast.
Finally, we've taken a brief look at some of the Computox's ancestors and peers. We've established that it's unlikely to be clone of any of the well documented machines, although it does have similarities to several, notably the Haufe design.
So what's next?
The bulk of the work remaining is reverse engineering the logic core. I suspect this will mainly be grunt work, and probably won't lend itself to regular updates until sections of logic can be isolated and explained, as has been done with input latching. Whilst we now understand roughly a quarter of the machine's relay logic, there's quite a bit to go. For example, we still don't know what the two uniselectors are for, although I have my suspicions (the main one on the left could be for computer move selection, possibly similar to the Haufe design, and the right one might be some kind of initialisation or state logic given it's triggered by the reset button). As this is the first relay machine of this scale that I've played with, the task of investigating the logic remains quite daunting.
The longer term hope is that once the Computox is documented, it might be possible to build a replica or work-a-like machine. This would allow visitors to the Museum of Computing to continue to enjoy the authentic clickerty-clack of a relay tic-tac-toe machine, whilst preserving the venerable original as a static exhibit. To this end, the Museum has furnished me with 60 double poll, double throw relays and sockets from their pile of random spares.
 |
| Museum relay stash |
As a proof of concept, I've sketched out a design that uses ten of these to demonstrate the Computox input latching behaviour we've already documented. So, look out for a write-up and video of this little side quest once some 24V bulbs and lamp holders arrive.
 |
| Computox input latch demonstration circuit |
I'm sure there will be other things too. In the mean time, thanks for taking the time to follow along, I hope you're enjoying it. Right, I'm off to buzz out some more logic!
Comments
Post a Comment