My first 65xx PCB design way back in 1984

[pbj]

After seeing a photo of the Sontec board, I thought about my many 65xx boards, but the very first one was based on a R6511AQ (romless) where I merged and hacked the Rockwell Forth kernel and dev ROMs with a few extras. So this board was used for development purposes but ended up in some early POS terminals in 1984 until I then shrunk the CPU+ROM+RAM etc into a credit card sized module. I don't think I've posted this before but here it is.

[GARTHWILSON]

Very nice!  What did you use to lay it out?  It doesn't look like crepe tape!  Did you actually have access to some kind of CAD back then?  I'm not sure I had even seen a color monitor yet in '84, only TVs used as very low-res monitors.  What's the WDC IC?

[pbj]

The artwork was done 2:1 with red and blue tapes but within the year I was using MacDraw with a custom library to layout these PCB manually. I did have a MacPCB package but it had too many restrictions.

WDC is not Western Design Center - it is actually the people that make hard drives and SSDs etc - Western Digital Corporation. The WD2793 is a floppy controller as the IDC header gives that part away. I actually used reversible 3" hard jacket diskettes that Amstrad used but 3.5" became the defacto standard after Mac started using them.

[Martin_H]

Was the Rockwell Forth kernel derivative of another Forth, or was it entirely it's own thing?

[barrym95838]

Apparently it most closely resembles fig-FORTH.

https://github.com/glitchwrks/rsc_forth ... Manual.pdf

[Martin_H]

Thanks, that makes sense given the era and architecture.

[pbj]

What I liked about this implementation of Forth was the one extra Branch-on-Bit-Set instruction in the donext loop. I would get low-level interrupts to set this bit and cause it to interrupt Forth at a high-level. For the early POS application the keyboard scanner would set this bit when a key was pressed and so my keyboard processing could interrupt the foreground console/remote serial console. Most keyboard processing was done with 100ms or so anyway.

[pbj]

Ah, found a photo of the original artwork, Remember that these were done 2:1 so reduced photogs could be made for the top and bottom negs as they did back then.

[glitch]

Nice work! Presumably the FDC sits at 0x0100 so the built-in RSC-FORTH routines can talk to it?

[pbj]

glitch wrote:

Nice work! Presumably the FDC sits at 0x0100 so the built-in RSC-FORTH routines can talk to it?

That was so long ago it set me off on a rummage of my old files. This board was the only one I bothered to include an FDC on-board - mainly for development purposes. At that time my dev system was an old Compucolor II computer as it had an inbuilt floppy and I wrote the ROM decoder generator in Fortran which I managed to find a copy of. However I couldn't locate the actual ROM equation files I used, I think it's been lost to history! However I'm sure it used the same address as the original RSC-Forth did.

Now I've got to clean up the mess I made rummaging.....

btw, I later switched over to Macs in 84 which I used for dev work and documentation too.