Serial Board Project

It took me a long time to land that first design job. That's what happens when you graduate at the very beginning of a long recession.

At one point I determined that the only way to break out of the age old dilemma of needing experience to get a job and job to get experience was to do my own projects.

I needed a tool kit first. Design tools then were even more expensive than now but I did it. It helped that I could make a reasonable living doing IT at the time. Here's what I got:

Verilog simulator: Finsim
Verilog synthesis: Synplicity
FPGA backend: Altera MaxPlus+II
An old but usable 96 channel logic analyzer
Oki wire wrap gun with cut/strip/wrap bit
EETools Allmax device programmer
Dedicated PC (my primary machine at the time was an Amiga 3000
Dedicated Amiga for hardware hacking

The result was source and test bench for an Amiga high speed serial board operating at up to 460Kbps with hardware support for PPP framing.

And of course the board itself:

Front

Back showing wire wrap connections
back of serial board

The important parts. The headers in the in the middle/right are for hooking up logic analyzer probes. The missing chip on the left is the 7.37Mhz oscillator. It seems to have fallen out. The MAXIM chip is the RS232 transceiver. It is not capable of 460Kbps operation. MAXIM did made a suitable chip though it required more passives. I was still working on a source when I ceased work on the board.
Closeup of chips

Did it work? Not really determined. At the time I landed the job with Siliconians, I was debugging a "hot chip" problem. The Altera parts would burn up if you loaded a "bad" bit stream. What "bad" was or how you got one was never clear to me. Given the general bugginess of the software, the design itself may have been fine. 'Probably not. I had already squashed a few interface bugs. There ware probably more.

However, once I started working at Siliconians, I focused on their projects and never really got back to the serial board. It didn't help that the intended application had come and gone. A serial board built for ISDN isn't all that meaningful in an age of DSL and cable modems.

This is old code and has definite style issues. I suppose some time I should change all the register assigns to non-blocking and see what happens. It will probably fail simulation.