Sol-20 Integrated Computer Teardown

[Action Retro] came into an antique Sol-20 computer and argues that it was the first totally integrated computer aimed at consumers that didn’t require you to buy or build some kind of terminal. These are fairly rare, so we appreciated the peek inside that you can see in the video below.

Sure, the Sol-20 wasn’t the very first computer out there in the market. It was, however, one of the first ones that didn’t need anything more exotic than a monitor to have a functional system (and the monitor was included). There were alternatives such as a Xerox Alto or a Wang 2200, but those had price tags that didn’t land them in your home. Even Apple, which would become famous for a turnkey system, was only producing the Apple I at that time. As the video points out, it was complete as long as you could build your own power supply and knew how to interface a keyboard — keeping in mind that keyboards were all wildly different in those days.

While the Sol-20 was the first to the market, everyone “knew” this sort of form factor was necessary. The key was to get the price down. Even the Sol-20 sold for $2,100 with a monochrome monitor, BASIC, 8K of memory, and cassette tape mass storage. Remember, though, that $2,100 in 1976 is worth about $6,260 today. You could buy a motherboard kit ($575) or the base machine assembled ($745). Adding the case, keyboard, and power supply would add $300 to $500 to that price. Of course, unless you ponied up the whole $2,100, you had to source your own bits like cassette decks, memory boards, and the like.

The computer didn’t boot up immediately, but cycling the power a few times made it do a little better. The S-100 bus connectors take up quite a bit of real estate inside, and a huge linear power supply was pretty common for computers of the period. The keyboard was beyond help and required some major renovation and cleaning. It wasn’t totally fixed by the end of the video, but you can see he’s getting there.

The company had made inroads into the fledgling computer stores of the day and had promised improvements, including a floppy drive and a color video card. Vendor problems delayed the floppy disk release and then came similar and cheaper computers from Radio Shack, Apple, and Commodore. Processor Technology — the company that made the SOL — had legal troubles, too. The company they paid to write the SOL’s BASIC was selling the software to others, but the courts sided with the other company. The company also refused to move to CP/M, using a proprietary operating system that was reportedly buggy, resulting in a lawsuit against the company. The last straw was 32K and 64K dynamic memory boards that were sold and then failed frequently. The company went bankrupt in 1979, ending the three-year run of the SOL.

We suspect this machine will be totally function soon enough. If you want one of these but can’t afford or find one, you can do a reproduction. Or, emulate a computer from the same year — the Apple I.

19 thoughts on “Sol-20 Integrated Computer Teardown

  1. Assembled one of these from a kit back in the day. I was envious of S-100 machines that had blinkenlights, so I rigged mine up with LEDs for most of the bus signals. I cut some notches in the front panel, carefully folded the “Sol Terminal Computer” label out of the way, and added my own labels for the LEDs. Fun times.

    I’ve posted my original S-100 systems guide (full schematics & assembly info, etc) as well as a number of other Sol / Processor Tech info on the Internet Archive:”Processor+Technology”

    1. Unsure really – old age sucks. But my superboard2 was a heck of a lot more reliable than my neighbor’s SOL. The writer is correct – the SOL’s power switch was really called “engage prayer subroutine” around these parts.
      I second your motion: groovy vibes baby. Groovy.

    2. The OSI Superboard/C1P came in 1977. There were OSIs earlier, but I’m not sure how early.

      The Sphere-1 came in 1975, beating out the others. But it was never clear how many were sold.

      1. My first good computer was S-100 built using CCS California Computer Systems z80 Board, 64K RAM board and a two serial two parallel board, Hayes 300 baud modem, Floppy controller board and two 8 inch floppy drives. Later I added a Tandon HD. I recall having edit the CP/M bios to interface the HD controller board. I don’t recall the brand of HD controller but a do recall it used an 8X300 processor that ran so hot it would would immediately raise a blister if you touched it. This system was housed in huge aluminum chassis with an enormous linear power supply with two capacitors a little larger than beer cans . Threw that stuff away 25 years ago. Wish I still had it. Those were the days.

        1. The BIOS was of course in RAM loaded from the drive, for the kids not old enough.

          I have a 10,000uF 16V capacitor I bought surplus about 1974. Barely enough voltage for a 12v supply. As big as a Coke can. “Computer grade capacitor”.

          I have some now the same value and higher voltage the size of my thumb.

          1. Exactly. CP/M was delivered with the BDOS assembled and a stub BIOS frame work as ASM source. If you wanted drivers, you wrote them in 8080 assembler. The first driver I wrote was a Centronics parallel printer port driver for one of the first Epson MX-80 printers to hit Los Angeles. Good times…actually. it wasn’t that complicated. The MX-80 user manual provided the parallel port signaling information….those were the days 😀

  2. change c15 to 47 micro farad,R18?from 10k to 33k, schottky logic would be nice too, then the reset will be more reliable, the ol boysin design never gave much thought to POR in those days, lol [ “have you tried turning it off and on?” lol thanks “IT crowd” for the reminder! haha

  3. I just ran across a box of S100 cards and have it in the pile of stuff I am giving to a friend who is big into retro computing. The S100 was such a disaster. The bus was basically all the signals on the 8080 buffered and made into a bus. The truly obnoxious part about it (if you had to pick out just one thing) was that is had two 8 bit data paths (not bidirectional) one outgoing, the other incoming.

    Never did find out what the processor was in the SOL-20. I am betting on a good old Z80, but I couldn’t tell by skimming the video.

    1. It was an 8080. There was an article in Popular Electronics.

      Remember, the Altair came practically first. So who knew what the bus should be? It was very specific to the CPU, which is why it wasn’t used for other CPUs.

      I have a Processor Technology 16K dynamic ram board. In a box of junk. Cost a pretty penny new, now only valuable to an S-100 owner.

      1. That’s like the 64K static RAM card that I squirreled away – with 32 count’em Thirty-Two! 2116 2Kx8 chips on it. Crammed in from one edge to the other, hardly room for decoupling caps.

  4. I was developing a university microcomputer lab when the SOL-20 was the state of the art. I have two memories of the SOL-20:

    1: First demo of the SOL-20 in front of colleagues. The president of the local computer club and owner of the neighborhood computer store showed up, set the computer up, and plugged it in. Before loading the OS, the monitor showed an almost perfect screen from the Trek game that had been played on it the previous night in the store. Credit the fact that the SOL had static, not dynamic memory, and the video did not have dedicated memory. The contents of memory had started to decay, but most of it still held the contents from when it was powered off the night before.

    2. First demo of the SOL-20 in front of University administrators responsible for funding the microcomputer lab. Setup same as the previous demo, but when the SOL was plugged in it erupted in a shower of sparks worth of a Klingon direct hit on the Enterprise in an old Star Trek episode. Turns out the power supply in early SOLs was mounted under the main circuit board, and in numerous trips to demos the pins on the bottoms of the discrete ICs worked their way through the insulation of the transformer and into the windings causing the 120V short circuit that would inevitably happen. We just happened to be the lucky ones.

    In spite of the fireworks display – or maybe because of it? – we got our funding and started life with an IMSAI 8080, a North Star, and eight Apple ][s, all with significantly safer interior layouts. We checked.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.