André 's CS/A65 computer

•  
  2019-08-31 Overhauled the page and added more pictures to celebrate the 30th anniversary of the CS/A65 computer!

• Current state
• Transition to new tech
• The Beginnings...
• The Gecko board
• My configuration - in the 90s

This section shows what I currently have, in a way, as the pinnacle of the current design....

This is my current case for the CS/A computer

Here, more from the front, you can see the SCSI devices built into it, a CD drive and an 80 MB harddisk.

The keyboard can be taken away - the spiral cable shows it's the same as from 30 years ago - to expose the SCSI "plane" with a SCSI CD drive and two SCSI harddisks

Here you can see the two PC disk drives, 5.25 inch and 3.5 inch. The small boards on the back are a PetSD, and a video mixer to produce composite video.

Taking away the disk drives exposes the actual CS/A boards and the PC power supply.

This setup emulates a PET 4032 with CRTC, just with an MMU to address 1MB memory, and expandable with further cards...

In the beginning, I made my schematics by hand, and soldered everything on perfboard manually. I actually bought an etching kit for PCBs, and got one CPU board etched by someone, but that in the end was not successful, so I stayed with manual soldering.

Then, after university, when I had a bit more money to spend, I invested into an Eagle license and moved over from manual soldering to etched boards manufactured by a service on the internet. There was a learning curve, of course.

In the end I was even capable of soldering SMD parts on to such etched boards.

Here you see some examples from the transition

The NETUSB card shows that the board space to do voltage conversions and interfacing to the parallel bus has become relatively large compared to the actual functionality.

The Olimex-based Ethernet board was a short board, with two options that only basically did voltage conversion to 3.3V, and used a breakout board with the CS8900 as Ethernet interface.

The blitter board transfers blocks of data. It is the board with the most components on it (probably by far). The first try with SMD bypass caps.

Not really much of a difference on the component side...

... but the solder side definitely looks different. No wires on the etched board!

Etched board from an early try - no solder resist mask or even silk screen printing. Also decoupling caps are there - but supply line traces are still laughingly narrow like a signal trace...

Some parts of the schematics have been revised a few times as you can see from the rubber marks...

The solder side with many, manually soldered wires, and extra decoupling caps (that were not in the initial version... - talking of learning curve...

Below is a view of the computer itself. It consists of a passive backplane, and several card boards. From right to left you can see the CPU board, the BIOS board, the video board and the keyboard board. Then there are the IEC board and the shugart bus board to connect PC floppy disk drives.

I initially built all the boards using perfboard, meticulously soldering wire by wire.

These pictures must have been taken between around 1990 and 1992 where I lived in that student room where they have been taken. It shows an edge of the Atari ST I was using as mass storage device (using an IEEE488 interface I built and that I connected to the Atari's ROM port) and development platform. On this Atari I first wrote the xa65 cross-assembler for my work on GeckOS.

Another picture shows the Emulator board (the grey one in the back) that emulates the 6502 CPU of another computer (a CS/A65 CPU board without MMU, in the front - this very board later became the Gecko prototype board, see also below). The long cable is plugged into the socket where the CPU resides usually on the CPU board and on the special socket on the emulator board. Because the CS/A65 computer with MMU can handle 1 MByte of RAM, the 64 kByte of the emulated CPU are mapped into the memory of the CS/A65 system.

I have been able to test new ROMs for the VC1541 in RAM before burning them into EPROMs. I just replaced the old ROMs on the board with some RAM area from the CS/A computer with the MMU, loaded the new ROM into this RAM (from the CS/A computer) and started the VC1541 DOS.

The oldest picture of my configuration, where I already had CPU, BIOS, Video, Keyboard/RS232, IEEE488, Floppy, and some unidentified board (from right to left)

Here you can see how I built the boards originally. All sockets were soldered on perfboard, and then wired together using separate wires. Only occasionally I used a 2.54mm spaced parallel cable for busses.

The grey emulator board in the background allows emulating a 6502 CPU in another computer (for example in a Gecko prototype in front)

After so many people asked me to give some hardware diagrams to the public, I decided to design and build a simple non-MMU system especially for Web-publishing. So here are the photos of this one, that is now called Gecko:

You can see the CPU, RAM, ROM, a VIA and a 16550 RS232 chip, plus some glue around it.

The same computer from the bus side - as you can use this as CPU board on the CS/A backplane as well!

Between 1996 and 1999 I had a very special "tower" configuration of all my 8bit stuff. Everything was connected, and I used the PC as mass storage device and programming platform.

The PC was used as disk drive on the IEEE488 bus, that all machines (except the 128) could use.