When the mood to tinker strikes ...

[Jove]

I have been putting together a replacement ISA riser for my Model 25-286 and decided I had enough room to fit a small SRAM to provide UMBs.

In theory it will work so I sent the gerbers out for fab. I got in a mood yesterday and decided to mock up the UMB circuit. A floppy controller card from Sergey was sitting nearby and has most of the signals in mostly the right places.

What you see is 128K of memory, mapped by a GAL in 4K blocks. C800-DF00 for my Model 25 (with XT-IDE at DE00). It also works in my Sergey 8088 providing UMBs from C800-F000, lasting through 30 minutes of continuous ram tests.

Don't judge my soldering/wiring, this is a quick throw together to prove a schematic, and I was trying to make use of the solder that came with my new soldering iron. I know better, but I tried anyway!

The important part to this is that it works.

 

[VERAULT]

Thats excellent. Was 128Kb your goal or just a jumping off point?

 

[Jove]

The target platform only has a small open window for UMBs and add-on BIOS images of 96K. It has a very large BIOS (128K on its own!) and not much room left over. I only have 32K SRams on hand, so instead of using 3 chips I ordered up some cheap 128K chips.

This is partly to free up some conventional on that particular machine, partly just the academic exercise.

Vintage computing and hardware in particular is just me reliving my youth by re-learning what I knew 25-30 years ago and expanding into what I couldn't understand back then.