Image Map Image Map
View RSS Feed

Gary C

Further adventures of Multibus

Rate this Entry
Finally got my 286/10 processor running in my multibus rack.

I wire wrapped a iLBX bus between the CPU and memory cards but it wasn't starting on power up. In the end it wasn't getting a clean reset so I have now wired up a toggle switch to the Multibus master reset line and now it starts with a memory check (i think, it gives a row of 0 0 0) then the intel monitor program welcome message).

Next issue is software. Not sure if I will program it myself with a simple OS, or try to get one of the vintage systems running on it. The big problem with vintage stuff is they all seem to want a disk drive and interface cards are rare and expensive. So my idea first is to get the system loading and saving to a TU58 tape drive. We use (in emulated form) this system at work so the code exists within the boot roms we use, however these boot roms try to load from a comms card and would need an IO card to enable the interface panel that allows the use of the tape drive. So I need to strip out the code from the rom.
My intel monitor roms are originals from intel and come on 27128 eproms, but our boot roms come on 27256 roms, putting these onto my board resulted in overlap between the two roms.
Quick look at the hardware reference manual, changed the link settings to match the 27256 rom for sockets U41/U76, then used primary/secondary memory option 2/2 to select 32K for U40/U75 and 32K for U41/U76 now means that the intel roms in U40/U75 exist from F0000 to FFFFF but is repeated in two 16K blocks and the ones with the TU58 code in U41/U76 run from E0000 to EFFFF as one 32K block.

Handily, the ROM when it starts up gives entry vectors to certain functions such as tape load and tape save to be run directly from the monitor such as GF000:80 to load from a tape. Disassembling the rom at E000:80 shows an instruction to load the SI register followed by a JMP instruction somewhere else, so I need to unthread the code and find where the actual code is held and this means learning 286 machine code. Was ok at 6502 so should be able to get the hang of it, but need to get a decent assembler and get down to it.

The sharp's PSU has been repaired but sits forlorn waiting to be re-assembled.

Submit "Further adventures of Multibus" to Digg Submit "Further adventures of Multibus" to Submit "Further adventures of Multibus" to StumbleUpon Submit "Further adventures of Multibus" to Google

Tags: multibus Add / Edit Tags