Thanks for photos!
Chip codes seems from 77...79. It is obvious, it is not mass producted (board type, also
serial no. of vector
board is 039). Maybe tailor made for some special CAD system?
The fact that 2 of the boards are wire-wrapped not PCBs suggests it was a low-production
unit, but equally it
is too well made to be a one-off kludge IMHO.
"Vector Gen" board seems some nice Signetics chips, like weird multiport
8x4-bit ram (82S112), maybe used as
some buffer? Top of that same card goes 12-bit bus to 8263 multiplexers (wild guess:
4-bit for each color?).
I suspect it's similar to the little state machine in the HP2623 terminal for line
drawing and these RAMs contain the
parameters for the state machine.
I doubt it's 4 bits per colour, I suspect lines are drawn one at a time, you specify
the colour, the state machine
does its thing and the colour setting simply determines which bits in the video RAM are
changed. So the
vector generator would not know anything about the colour
Also there is 32x8 PROMs (82S23), maybe used as glue
logic / sequencer, too small for look up table.
I think that's a state machine type of sequencer.
How many 16Kx16 + ALUs board there were? Three? Maybe there is one processor/memory for
each color
plane...
6 of them. It might be 2 boards per colour, but one thing I didn't notice until later
were the 74189 RAMs on
the video processor board (the one with the DACs on it). My guess is that these are a
colour look up table
forming a mapping between bits in video RAM and the DACs (maybe 3 off 5 bit DACs). This
suggests 4 bits
per pixel (to address the CLUT RAMs) at the video memory side, how that fits in with the 6
Pixel Store boards
I don't know.
You have to put that back to live, and make some reverse engineering. It seems to be
really exciting machine!
It is going to be a big job to sort it out, but one that will be worth doing. That is,
unless manuals for it turn up...
-tony