Early Micros - some favourites
One of the earliest machines available to hobbyists, the KIM-1 is a single-board computer in the purest sense. In 1975, MOS Technology apparently had two ideas to push their newly developed 6502 processor into the market: sell them from big jars at the Wescon trade show to emphasize that their incredibly cheap CPU was really in full production (in truth, the bottom half of the jars was filled with chips from failed production runs). The other idea was to offer the KIM-1 as a demonstration board to third-party engineers. But much to their surprise, mostly hobbyists bought it. MOS had not realised there was such a hobbyist market out there. The KIM-1 is surprisingly usable, and Peter Jennings secured his place in history by cramming a chess playing program into its 1K of RAM. Actually, slightly less is available - the program used 924 bytes.
I looked 21 years for a good IMSAI before I finally got one... The IMSAI was a clone of the MITS Altair, which many people regard as the first commercially available microcomputer. The problem with the Altair, however, was that it, er, was severely challenged in the quality department. The IMSAI was really what you wanted back in 1975. A proper front panel, and delivered with the same 256 bytes of memory and 8080 processor as the Altair. The S-100 bus allowed the machine to grow with the microcomputer revolution for a good few years. My machine is equipped with two Altair 8" disk drives, a serial port to hook up a Beehive B-100 terminal and (yeah!) a boot EPROM so you do not actually have to toggle in a boot loader. And with 48K of RAM, it is the oldest machine in my collection to run CP/M.
MITS and IMSAI unwittingly established the first microcomputer standard with the Altair, the IMSAI and their S-100 bus. But it was Cromemco who first made the microcomputer a respectable business instrument for Serious Companies. At first, Cromemco (started by two Stanford PhD students and named after their dorm, Crothers Memorial) started with a rebadged IMSAI and complemented it with very high quality boards. Soon, Cromemco started building complete S-100 boxes, that were renowned for their quality - and the quality of the support Cromemco gave.
Cromemco System Three
The System Three is like no other micro I've ever used. Working on this machine, you really feel close to the pioneers of the late Seventies. It's big. It's heavy. It's not only bulletproof, it'd probably withstand a nuclear strike the way it is built. I bought this out of the estate of a university professor who had carried this 200 pound monster all through Europe. When I bought it, it was just a pile of boards. In the end, I reassembled the original Z80/64k machine... but there's tons of other S-100 boards, ranging from 1979 to 1985. There's enough to assemble a 68000 Cromix machine. If only the hard disk card worked.
Cromemco System Two
Same size as the Three. But because it uses 5.25" disks, it feels less ancient. But in reality, the Two actually predates the Three. Its disk drives are not nearly as fast as the exotic-but-unreliable 8" PerSci's of the System Three, making it a much more normal CP/M (actually, CDOS) computer in daily use. My S-100 rack holds a Z80 ZPU board, a TUART interface board, a 64KZ memory board, and a 16FDC disk controller/terminal card. I also have the DPU processor board, which adds a 68000 processor to the system. That's of very little use without the Cromix operating system alas, and I don't have the required non-standard hard disk.
A truly exotic computer from 1980 which holds the title of Least Documented Computer On The Internet (according to me). Made in Vienna, Austria by Ernst Steiner, this is a multi-user 6502 system. It boots up from a ROM board, and every user is assigned a memory board full of 2016 RAM chips to work on. It is representative for a whole class of custom-designed computers from the late 70s: no standards, sophisticated users knew how to make the things useful. My es65 has the luxury of a dual floppy disk drive so the users can actually save their work. Unfortunately, I have no documentation at all - if you have any information, please let me know!
Where the es65 represents the specialty niches of the late 70s, the Intertec Superbrain shows how standardisation set through in the CP/M and Z80 world. By 1979, CP/M was a ('the') standard and machines like the Superbrain were integrated boxes containing all you need. No separate terminal or S-100 expansion cards (although you could add one if you wanted to) were needed. Buy the box, switch it on, boot from the disk.
My Superbrain has a 5MB hard disk which worked up until 2012. The Superbrain let you know about this fancy peripheral: the noise was not unlike that of a Cessna in mid-air. But it is the best CP/M machine I've ever worked on: lightning fast, solid, and quite ergonomic despite the bizarre looks. Apparently, the machine's case is made of recycled garbage as the company had an environmentally-friendly idea. Or so I was told. I plan to replace the now-deceased hard drive with a Compact Flash card. But I'll keep the hard drive's motor running, because without its acoustics some of the charm will be lost...
Introduced in 1983, it was regarded as part of the Japanese Invasion in microcomputing. The machine has very American roots though. Its main claim to fame was Valdocs, which was a brilliant 'integrated office suite' idea at the time. But like many brilliant ideas, it arrived slightly before sufficiently capable hardware... In Europe, the machine was sold as a top-of-the-line CP/M (actually, TP/M-II, which is worth some Googling for CP/M enthusiasts) machine. Opening up the QX-10, you quickly realise that this machine must have had an unusual design history: despite all the fancy graphics capabilities the machine has thanks to its uPD7220 CRT chip, the rest of the machine is completely built up from simple 74LS logic chips! Not very Japanese at all. But the slick case design shows how Personal Computing was finding its final shape by this time.
This was the machine that started my collection back in 1988, and still my favourite. It represented the ultimate of what could be done in packaging computer hardware in 1984. It also marks, more or less, the end station for 8-bit CP/M business computers. Soon afterwards, much more normal MS-DOS machines arrived, even in laptop form. Owners of PX-8s are very sentimental about their machines - for a reason. This little machine makes you want to go back to 1984 and just - er - use it for Real Work, to experience the minimalist Zen-like exaltation it will bring you...
The PX-8 has a built-in tape drive for backups, two 32k EPROM 'drives' that you can hot-swap, and a RAM disk. Features like auto-off (and auto-on) and sleep mode were standard, and make this machine seem so strangely up to date 20 years on. It came with a custom WordStar, a good Basic and decent utilities for connecting it to the outside world. The most important peripheral is the 128k RAM disk, that sits as a wedge underneath the PX-8. Without that, the PX-8 is of limited use, unless you hook it up to a PC used as a disk drive emulator. Other peripherals: I have the TF-15 dual 5.25" disk drive, which is too bulky to fit the PX-8's purpose, and the PF-10 3.5" disk drive which is excellent. It took me 15 years to find a PF-10 though: they are very scarce indeed.
Looking back from the vantage point of 1984, it was clear that microcomputers had evolved into two separate strains. On the one hand, CP/M machines based on the Z80 and its brethren were geared towards mainstream business (word processing, early spreadsheets, but also mainstream software development). All the CP/M machines on this page ran pretty much the same software, and as a result, that software was smooth and its user base was large. Their role was taken over by MS-DOS machines and that is also why CP/M still feels very familiar to later computer users. Micros based on the 6502 or 6800, on the other hand, tended to be fitted with custom software. Yes, there were 'standard' operating systems for 6502/6800 machines, but these had tiny market shares and most machines had their own, incompatible, software setup. And therefore, non-Z80, non-CP/M machines were generally either specialist computers or remained in the hobbyist sphere.