The board on the right used a dynamic RAM driver chip (can’t recall what it was, but probably Monolithic Memories). We did not use the Zilog MMU. The clock generator was from AMD (Zilog second source for the Z8000 for a while, and Zilog was a bit late with their own clock generator). We did use the Zilog clock drivers later, but they were not pin compatible with the AMD part. We also do not recall how much memory there was but it was most likely 512kBytes like the System 8000 (used newer chips).
After the static testing, several NOPS were programmed into EPROMs and the rest of the board was tested with an oscilloscope to check for valid timing. The board was never converted to a printed circuit board as there were few customers who were perpared to invest in a System 8000 for development. We were really happy with the Trump Card, but we were a bit too early with this card.
The Z8001 card had several peripherals that we wanted to test before laying out a board. The embedded cross-compiling with the System 8000 system was not a given and we had a lot to learn. It was also too late to change architectures on the masters degree (was Z80 based), so we did not have that much time to do a board layout. Wire wrapping was really fast.
The wire wrap side of the Z8001 board. The dynamic RAM array had power strips (from Rogers or similar name), and a capacitor per chip. The prototype boards were made from a layout program on a Tektronix computer running BASIC.
When wiring the board, the pin numbers are inverted as the board is “upside down”, but most of the board was regular arrays (memory). The schematic capture was on OrCAD as far as I can recall—MS-DOS based with plotting done on a dot-matrix printer. A couple of years later I would go half shares on an A0 HP plotter on a special from a friend who worked at HP in Johannesburg and bought the demo model.
© 2015 — Second Valley Software Pty Ltd. All Rights Reserved