C64 maze generator
Pak Pak Monster
Research Machines and 380Z history by Nick Hegenbarth
Research Machines (RM) had been trading as Sintel, a firm selling digital clock kits and the like in amateur electronics magazines by mail order. They once had a letter fom Intel's lawyers suggesting that they should change their name as they might pass themselves off as Intel by answering the telephone "Hello, it'S intel". Being an Oxford and a Cambridge graduate the pair of them, the Directors had enough sense and confidence to write back telling the lawyers where to go! And in those days it was enough, they said no more about it.
THE EARLY DAYS
I had one of their clock kits. I also was looking at the new Microcomputers you could buy and in 1978 I and a friend bought between us a 280Z kit. Well we ordered a kit, but they never shipped any. Thinking better of the support problems they supplied fully built 380Zs instead. In those days in a blue case. One revolutionary thing was that the bus was made of the new ribbon cable with crimp on connectors (now commonly used on the end of hard drive cables). This obviated seating problems with edge connectors. They also had plans for a multi user system with multiple processors in one box connected by a high speed serial link.
The interesting thing is, they invented something like USB long before USB was ever made. It appeared as
Piconet in the Nimbus.
Having visited the firm to see the 380Z they were aware I was looking for a job, and offered me one, which I took, taking over in May 1979 from the Managing Director as Technical Support.
It was the following winter when the notorious "dropped belt" problem arrived. I was very puzzled at first, but then we worked it out eventually. On a cold morning, and worse on Mondays, people would come in and switch on the machine, the motor would start in the 5.25" disc drive and the (cold and rigid) belt would fall off!.
The drives were made by the German firm BASF. It took a great deal of effort on the part of the MD Mike Fischer to persuade BASF that this was a real problem. I came up with a cunning workaround. We modified the ROM (now what we would call the BIOS) so that pressing "x" on the keyboard instead of "b" (for boot) would start the machine with the two disc drives swapped, thus booting from drive B. This solved the immediate crisis allowing people to put the belt back on at their leisure, or send it to us for us to fix.
PDP11 AND POLY 88 INFLUENCE
An interesting feature of the 380Z was the ROM operating system's "Software front panel". Meant to emulate a hardware front panel. Mike Fischer the MD had researched in neurophysiology and knew another chap at the National Hospital for Nervous Diseases, who always reminded me of Jonathan Miller in terms of his gentle nature, his astonishing intelligence, and his nose! His name was David Small. Without him they could never have made a computer. He knew how to program at low level, in Assembler code, and had some PDP 11's he used in his work, neurological signal processing.
So they gave him a third of the firm. He wrote the ROM (BIOS) which included the Software Front Panel. He included it as he was used to hardware panels with rows of switches like the PDP 11. He had seen adverts for the Poly 88 showing what he thought was a software front panel which he assumed was a static page, updated in-situ, and liked it.
He only later found that the thing scrolled like a glass teletype and was not as good as his, not a software front panel, more of a core and register dump. The same happened with the text editor he kicked off. It was an 'in place' editor, which we now take for granted. In those days editors generally scrolled and had arcane keyboard commands, if they were not actually on paper!
Subsequently we added an 80 column graphics board replacing the existing 40 column one. I had fun designing the 7x5 dot matrix character font for this. I subsequently did the extended set for the Nimbus, and the one for the later AX and VX series of PC clones. So I can now claim I have designed something people stared at for hours on end!
The first and most basic 380Zs used audio cassette tapes for data storage and unlike the others around used software to generate the waveforms on the tape, like an early DSP, using the knowledge they had of digital waveform analysis.
HIGH RES GRAPHICS
What we called High Resolution was in fact 640x480 in 4 colours (including black) or 320x192 in 16 colours. But in those days it was pretty high res! The thing had a Colour lookup table in hardware so you could map the 4 or 16 colours to any of 256 colours and flip them at will. This was cleverly driven, by David Small and his (by then) two programmers so as to be able to split the memory into multiple pages and switch on and off each page in hardware. This lead to some great animation possibilities, much to the annoyance of the makers of the BBC Micro, who hadn’t thought if that.
THE BBC MICRO
RM was sent an invitation to tender to make the BBC Microcomputer but the directors decided that any firm who made it would go bust as it was not practicable to include so many features at so small a cost. He was right, they did!
The BBC Micro sold heavily to schools in a Government Grant scheme, alongside the 380Z. RM sold fewer in volume terms but more in UK pound terms.
Of course, the BBC Micro also sold well elsewhere and as a games machine. I still have on in my loft I think.
RM, even in its original leaflet on the 380Z, always intended to make a multi user system for classroom use. The idea was to have a processor each, but all mounted in one box, to be connected internally by a serial link. They had realized that a serial link can be as fast as a parallel one due to timing problems with parallel links. A lesson that has only just caught up in the guise of USB/Firewire/Serial ATA/PCIe.
So I said "Why not put the processor card in the keyboard case and link them together by serial link with a 380Z as a shared file store. With this simple phrase I had conceived the LAN and the Server. Not that I had heard either term before, it just seemed a logical technical step. Mike Fischer immediately did just that, and we subsequently became involved with Zilog in what they called Z-net, which was an early lan hardware technology. There was a launch at London 's Café Royal and we were there along with a firm making an early electronic Hotel lock system and who wanted every lock to communicate.
We had been using CP/M. Luckily no idea is ever thought of by one person and they had come up with a network operating system where each station could boot off the central server and share the discs and printer - CP/NOS.
The BBC Micro, always our competitor in schools in those days, subsequently made Econet.
The stations on our network were discless as they could boot off the server (new word for us then) - this is what the
480Z was, a 380Z on a board with keyboard built in.
We knew we would make a lot of these (and we did) so we went for a plastic moulded case rather than our traditional metal. I was in charge of how it would look, but we had a tight deadline due to the lengthy process of making the huge moulds needed to press plastic cases (a meter cubed of solid steel, handmade) I foolishly omitted to ask the MD if he liked it (he was never at the meetings), and way to far down the process he took umbrage, thinking it looked to toy-like, he foolishly decided to redesign it. Meanwhile we had a folded metal case designed which I suggested should be strong enough to stand on as it was to be for school use. We made lots of these before the plastic case came along, looking rather like the metal one (ugly)! Buyers in many cases actually preferred the metal version and wanted to order them. A picture of John Bowden, the (large) Sales manager, standing on one became famous.