The purpose of this page is to provide information, documentation, and as a repository for all things Technosys and Aamber Pegasus. If you have anything to share, please contact me here

When I was a young boy living in New Zealand (and around the time I owned my ZX81), I was given an Aamber Pegasus computer by my next door neighbor who was involved with Technosys before it bellied up. The Aamber Pegasus I had even came in a fancy fake leather (actually cardboard!) suitcase style enclosure, with the keyboard exposed for easy access.

The hardware was designed by the late Stewart J Holmes, and the software by Paul Gillingwater. The machine is based around the Motorola MC6809C CPU. The basic system came with 4k RAM, and later versions had 64K RAM. There was an optional rotating switch that allowed you to select one of up to 6 language environments, from 12 different EPROMS - there was Extended BASIC, PASCAL, Tiny BASIC, MAD (Assembler) and FORTH, as well as games on EPROM such as Tanks, Invaders and Galaxy War. Other EPROMs included Monitor (which had to be present for the system to function), Disassembler and Word (which I assume was a word processing application)

The Pegasus was marketed on the strength of its support of multiple computer languages and low price. A network version of the Pegasus provided connectivity to a 6809-based server (SWTPC-6809). The Aamber Pegasus system (particularly the networking version) attempted to address the government's ‘computers in schools’ initiative, but never produced the hoped-for large orders. It is thought that Apple Computers introduction of the Apple II computer into the New Zealand market, and it’s subsequent heavy educational discounting was the final nail in the coffin for Technosys and the Aamber Pegasus Computer.

One of the most unusual aspects of the machine is that to save the cost of a CRTC (e.g. 6845), the processor set up some bits on the 6821 (PIA) to control the row being read out, then stepped through a series of NOPs so that the address lines of the CPU could act as a big counter. This counter drove the X address of the display RAM. On every row, the CPU updates the row number selected from the character ROM (and programmable character RAM) and every 16th row it  increments the Y address of the display RAM. At the end of the screen the output is blanked, and the CPU gets to do some “real” work until the FIRQ pin is pulsed by the 50 Hz line from the power supply. Essentially the Pegasus used the mains frequency to trigger vertical sync. Because of this, the CPU is ~90+ occupied as a counter, so in a non-real-time app you could disable the FIRQ (one bit in the 6809’s CC reg) and your computer was 10x faster – albeit with a blank screen.

On the left side of the Pegasus motherboard you can see a small blob of putty. This putty is hiding a series of diodes that act as a simplistic 8-bit ID. This 8-bit ID will only allow EPROMS encoded with a corresponding ID to work in any individual machine. For example, an EPROM from a machine numbered 2569 will not work in another Pegasus with a different ID. To defeat the ID check you could either alter the circuit under the putty, or alternatively, hack the main PROM and disable the check.

I have scoured the planet and have found some remaining New Zealand Pegasus users (a rare breed indeed). Paul Sowden, who is helped my fix my machine, and Paul Carter, who was one of the original software developers. Paul is presently trying to fire up his old SWTPc machine (which was used for development of the Pegasus, and is a rare beast in it’s own right), and backup the Pegasus EPROMs he has. Nigel Keam, another of the original Technosys team living in the US, and has some serious programming skills.

The Machine I have now (ID 2569) was kindly donated by to me by Michael Fincham who rescued it from certain destruction.

Please check out the blog for regular updates.

More to come. Please stay tuned...