Calling any one computer the first hobby computer is fraught with peril. Most people think the MITS Altair 8800, first featured in Popular Electronics back in January 1975, was the first. Some might argue that others were first, but there is no doubt that the Altair started the hobby computer revolution from a practical standpoint. However, there was another computer that almost took the crown. It, too, appeared in a magazine — Radio and Electronics. But it was in the July 1974 issue. That computer was the Mark 8, and [Artem Kalinchuk] is building a replica that you can see started in the video below. This isn’t some Arduino work-alike. He has a pile of parts and some almost authentic-looking PCBs.
The Mark 8 used the 8008, not the 8080, so it was less powerful. [Artem] has been building a replica Altair, too. Check out his YouTube channel if you are interested in those.
Finding all the parts for such a project is a bit tricky. The schematics available have some errors, but others have created boards you can purchase, and they’ve been corrected. Some components will require modern replacements. However, he did find a CPU and some of the oddball logic and memory chips. We were curious about the total bill and if it was more or less than the original after adjusting for inflation.
The original article back in 1974 was just a teaser. You had to send in $5 for the complete plans. [Jon Titus], the designer, estimates that 7,500 plans were sold and about 400 sets of PCBs. If you could source the parts, build the boards (with no plated-through holes and, presumably, no solder mask), and troubleshoot the errors in the originals, you could have a working computer.
The project is just starting, but we are sure [Artem] will complete it, so stay tuned to see his progress. There are already videos about the backplane, the power supply, and the LED register display. The boards, by the way, are the ones we’ve seen from [Henk Verbeek]. If you want to use a 8008, but don’t have the stomach for the full build, try a clock.
The great thing about the Altair 8800 and IMSAI 8080 were their use of expansion slots (S-100 etc) and the use of the 8080/8085/Z80.
That way, they could be expanded into real Personal Computers running productivity software on a Control Program/Monitor.
The latter formed the basis for a whole ecosystem running same applications on different hardware. It was the first kind of true portability on a larger scale.
The Mark-8 beat the EDUC-8 into print by a month or so. But there was ‘Computer73’ which was published in a Dutch/Netherlands electronics magazine in 1973, beating them both. For the life of me right now I can’t find the pdf article scan which I downloaded from archive.org, not even the bookmark.
Yes it was in the dutch magazine Elektuur. But of this 73 design no pcb’s where made it was just a couple of articles with circuits diagrams and it never finished completly. I understood from the designer of the time it never really worked.
Interesting, thanks for the followup backstory about it.
It seems that PDF scans of old Elektuur issues can be found at https://archive.org/details/elektuur-1972-1975 but I could not find the articles you are referring to. Could you please elaborate on the title/author of these articles? It sound like they would be an interesting read.
You’re possibly mixing up things. Elektuur had some articles in 1974 about a “74 komputer”, a 16-bit TTL construction built around 74181 ALUs. The series ended abruptly after the sixth installment and was never mentioned again.
yes you are right. It was called the 74 computer . No mix up just the wrong name
Yes, thanks as well. I apologise for incorrectly remembering the name, which also explains why I could not find it. Thanks again.
Here is the link for the 74 computer information https://ia802505.us.archive.org/26/items/Computer74/computer-74.pdf
Thanks. May 74 is two months earlier than July 74, so this would seem to be the first home computer project detailed in a magazine, unless the Kenbak-1 had something similar?
“Never say Never or Always” might be the thing here.
Very cool! Pre-microprocessor homebrew computers are still the best things ever made. I was working on one myself, obsessively, but once the Z-80 came out, it was game over. Others might say it was once the 6502 came out, to which I say, fair enough. Except that once the 6809 came out, there was no reason to ever, ever again consider building an MSI-TTL-based computer, and the 68000 sealed the deal for any nonbelievers, while Intel was still futzing around with variations on the 8086. But in the early 1970s, the advancements in 16-bit minicomputers were so explosive, the early one-chip microprocessers were a giant step backward, where price was more important than performance, and it took a few years to get back to that level of sophistication again.
Thanks for the link to the original article. 😃
I love that 70s design. The PCBs back then looked so futuristic and fascinating, perhaps because they weren’t perfect and glossy.
It’s like traditional cartoon vs animated films, maybe. 🤷♂️
Those curvy, hand-drawn traces of the era look wonderful. You really see it’s made by a human being. The old, boxy aluminIum chassis of the day adds to this charming character, maybe. 🤗
70s tech is great, I think, even if it’s a modern replica. 🙂
PS: If the machine could handle ASCII and a serial terminal, it would make for a nice RTTY terminal
Btw, the modern version seems to use gray ribbon cables. That’s not necessary.
Rainbow colored ribbon cables are still available, like in the 70s.
The color coding helps at wiring. You’re less likely to confuse pins or pin order.
Again, 70s were great (and wise). They knew how to build stuff that lasts. Kudos to these 70s hackers!
“They knew how to build stuff that lasts”
Back then they really knew how to build stuff, but stuff that lasts – not so much, that’s more myth than truth. In the 1970s many technologies were in their infancy. Plastic DIPs were prone to failure due to humidity leaking in and chip passivation was not commonplace. This weakness killed a lot of equipment that was not constantly powered on within a couple of years. Also many ICs died in storage and on transport, device manufacturers had to do incoming inspection on every single part (that’s where the small colored dots on ICs in old equipment come from) to avoid excessive rework after soldering. Military and industrial customers paid a high premium for pre-aged ICs in hermetically sealed cases. It took another two decades to arrive at the quality levels we are used to today. And then there were other components, that didn’t age well, like aluminium electrolytic capacitors suffering from trace amounts of chlorine (fixed mid-1980s) and carbon composite resistors developing cracks (mostly replaced with carbon film by the end of the 1970s).
btw: the circuit of the Mark-8 is somewhat awkward with its multiplexer stuff, it’s a design published by Intel in 1972 and “translated” from Intel’s obscure Schottky TTLs to more common stuff. In the original circuit, Intel avoided TI parts wherever possible. Guess why. ;-)
I know what you mean and I do see some good and valid points, but I don’t agree with the conclusion. It’s too simplified, simply.
Both TTL and NMOS parts, the industrial/lab grade type, of course, were excellent. I remember there were ceramic ICs with gold plated pins. Diodes, too.
Theyaall weren’t very power efficient (they got warm), but the use of resistors protected them, on a side effect. True TTL parts had more “oomph”, too and could drive heavier things.
I have such parts here, so I definitely know it’s not false.
An NE555, 741 or LM386 from the 70s/90s (no China stuff) did often fux seemingly broken circuits here that didn’t work when sold with modern ICs.
The big exception are early DRAMs, early CMOS. Dynamic RAM as such was an inferior technology (Static RAM is superior), so it’s no wonder that early types did fail.
I always mentioned this detail for sake if fairness, also.
Anyway, this is just my opinion/my experience, of course. I can’t prove it and I don’t need to, either.
It’s a common fallacy to use the small number of remaining and still working devices as a reference for how reliable ancient gear was if you don’t know how many of them were actually produced and the percentage that went to trash because of material related failures. There are no useful statistics on that, because in the extremely fast-paced technical world of the 1970s-1990s, three years old electronic stuff was considered junk and replaced, broken or not. In my personal experience (repaired some 70s stuff in the 80s), reliability and quality wasn’t all that great.
“It’s a common fallacy to use the small number of remaining and still working devices as a reference for how reliable ancient gear was [..]”
Yeah.. But the flaw in that conclusion is that I’ve also got my father’s old parts boxes here (those radio shack boxes with little cupboards), which are largely consisting of said 70s/80s parts.
And they’re almost _all_ fine, I’ve once checked them when being bored with various digital multimeters (checked diodes, transistors) and those universal LRC measurement devices.
They result was very optimistic. With the exceptions of a few resistors who became brittle and fell apart. About 3/4 of the electrolytic caps are fine, too, albeit I haven’t checked their RF behavior.
In addition, several 1970s devices here are in perfectly working order, without any maintenance.
Examples: Yaesu FT-101, a bunch of CB radios (allunmodified, one is a big PACE base station) , Hameg oscilloscope, MZ-80K computer, a bunch of big RTTY decoders in metal chassises, small green monitors, early slow-scan TV converters (one is a Robot model with discrete RAM), various shortwave radios et cetera pp.
Okay, we could also say maybe I’m just lucky. But throughout the years, my personal experience was all the same here.
Things with a date code from the 70s simply rarely failed, simply.
Their power supplies weren’t cheaply made, either. They had big, underpowered types, maybe, by today’s standards. But on the bright side these also had proper cooling blocks, proper filtering and all the protective diodes (thinking of the LM7805 and other linear voltage regulators).
The only thing that I have to complain about are certain US productions who were cheap on fuses, for some reasons. Thus I’ve added some extra fuses on the DC side where it seemed useful (internally, in the chassis). A diode against reverse-polarity on things with a DC input, too.
“There are no useful statistics on that, because in the extremely fast-paced technical world of the 1970s-1990s, three years old electronic stuff was considered junk and replaced, broken or not. In my personal experience (repaired some 70s stuff in the 80s), reliability and quality wasn’t all that great. ”
I believe you, it maybe could also be that certain parts/devices are more prone to fail.
As a negative example, I’ve encounterd various radio alarm clocks from the 70s that had defects. Either related to the caps (buzzing noise in radio mode) or defects on the seven segment displays.
What’s also possible is that we got different results because of different locations we’re living in.
I’m from Europe/Germany and many vintage parts we have are from Siemens, Philips, National Semiconductors etc.
It could be that these have a different quality standard.
On the other hand, E-Germany and Russia weren’t far away, either. They had produced (cloned) various semiconductors at the time, too. The E-Germany parts were not all bad, either, though. Still better than some of the lower end Chinese productions of today..
Another explanation could be that my father was never short on money. He always bought from known good sources. He didn’t waste money, but he wasn’t doing it the American way certainly (using always cheapest parts for maximum profits). He simply wanted things to last, rather, to make sure he wasn’t being bothered with fixing stuff soon again.
Another reason for the failures you encountered could be related to early CMOS parts, maybe.
The 4000 series wasn’t exactly mature at the time. Those parts needed careful treatment and stable, ripple-free power.
By contrast, the 74xx series of TTL chips was way less sensitive, I think. So it’s certainly true that not all 70s parts were rugged. But the then-new standard parts (741, 555, LM386, 7805/7809/7812, 2N2222 etc) weren’t bad, I think.
The NMOS versions of CPUs like the Z80 or the 80286 were power-hungry, maybe, but very rugged, too. NMOS used internal resistors, which wasted heat, but could limit current, also.
@Joshua:
Practically all defects were of mechanical origin, plastic DIP resin detaching from the lead frame and providing a path for moisture to creep in. Sometimes even the pins could be pulled out. Often experienced with Motorola and NSC parts, but never with TI and Fairchild. The latter manufacturers probably had better chemistry and screening methods. I suspect that the defects occurred due to thermal and mechanical stress during wave soldering and surfaced only after the devices had been decommissioned and put in storage for some time. If you compare reliability data published then and now, you’ll see that things got better by orders of magnitude in the past 40 years, mostly driven by the switch to SMD and lead-free reflow soldering.
btw: I’m located in Austria and the devices I’m talking about were made in USA.
Jonathan A “Jon” Titus wrote for EDN for decades. His articles were always worth reading – clear explanations of technical topics.
He did use the 8008 as the processor for the Mark-8, which obviously predated the 8080.
It will be forever argued what was the first ‘personal computer’, but for me, the Mark-8 was it.
The Mark-8 prototype ran successfully in 1973. I used two pieces of Vero board–the type with long strips of copper perforated on 0.1-in. centers. Everything except a small memory board used point-to-point solid color-coded telephone wire. The American Computer and Robotics Museum (Bozeman, MT) has the prototype memory board in its collection.
At the time I was a grad student in Virginia Tech’s chemistry department aiming for a Ph.D. My work used a DEC PDP-8/L minicomputer to control measurement equipment. After I programmed assembly-language routines I wanted my own computer. As soon as Intel announced the 8008 I asked for the chip’s details and for some samples, which Intel sent in the summer of ’73. As a grad student I had no extra money to buy edge connectors and create a motherboard. Thus, the board-to-board wiring and minimalist physical design. The case for the Mark-8 was one I had on hand from a consulting job.
I apologize for any errors in the schematic diagrams in the Radio-Electronics booklet. It was my job to proofread them and obviously I must have missed a few connections. The Mark-8 shown on the cover of R-E magazine was the only one I built. It now belongs to the Smithsonian Institution in Washington, DC.
It’s interesting to find so many people interested in “retro” computing.
I think it’s even more interesting to read a comment from the designer and builder on Hackaday 5 decades later!
Hear hear!
Kudos to the Mark-8 designer Jonathan Titus!
Live long and prosper!
That’s probably true, but there’s one thing to consider. That designer nolonger is same person! His body rebuild itself multiple times on a cellular level. His mindset has changed, too, maybe. So aside from his neural net, his brain wave patterns, he’s nolonger the person that buikt the computer. He’s a replica of the original he used to be.
Articles like yours in Radio Electronics helped me pick my career, and now I getting ready to retire after a successful career in engineering. Thanks!
People don’t realize how expensive this was.
1974 the 8008 IC cost $120 from Intel (from July 1974 RE story) or $750 in today’s dollars.
It was way too expensive to build then. A year’s wage was $11,100 on average.
I saw Godbout had them $50 June 1974 and something about a tariff and allocation.
By 1976 the 8008 down to $19, 8080 $38. PMOS is hard to work with.
Money, money, money.. The small people’s one and all. 🙄😂
A tip or trick back then was to ask for an “engineering sample” or just “sample”
(“Musterexemplar” in German).
If you had your own company, or your own business (even a one-man company), you could ask for it.
A big electronic manufacturer wouldn’t make a big fuss about it and give you a free sample worth a few hundred bucks for testing purposes, evaluating purposes etc. A oromotional gift, in short.
It’s most likely that the 8008 MCU samples Intel sent me were out of spec in some way–perhaps they didn’t operate over the specified temperature or voltage range.
8008 clock speed was initially 200-500kHz, then 800kHz for the 8008-1. So possibly samples were clearing out of the lower speed parts?
Intel did not actually design the 8008 but were given a spec from Computer Terminal Corporation (CTC) to make the IC. Unoriginal Intel lol.
Also found US Pat. 4093223 in 1976 for a football game using the 8008.
Not everybody was rich enough to get into computing in the early days. It was extremely expensive. I think this is why some machines such as the Mark-8 are extra rare. And the 8080, the Altair 8800 was out in full force if you had the money and that IC alone was $150 in ’75 ($860 today) plus all the required support ICs. Don’t forget your monrthly paycheque was $900.
There were no IC samples given out of exclusive, expensive parts. Even today, call up Intel or AMD today asking for a sample and tell me how that goes for you lol.
Byte magazine Oct. 1975 “Add a Stack to your 8008”
https://archive.org/details/byte-magazine-1975-10/page/n53/mode/2up
In December of 1972 I ordered and built a “Commodore Pet” of absolutely no relation whatsoever to the British company or computer. It was supposed to run on an Intel 4004 but to my surprise had been redesigned to work with an ‘engineering prototype’ of the 8008. The PCBs were riddled with errors but eventually me and my father, with the help of the designer and other buyers via snail mail and long distance phone calls, got it working. It looked much like an Altair and I’ve often wondered if the Altair designers ripped it off from this Commodore Pet. We ordered it from the back of an electronics magazine although I do not remember which one. He supposedly sold over a 100 sets of boards but only about a dozen people ever got it to work.
That sounds very interesting. Do you have any further info? Do you recall if it was a US or UK magazine?
HLK88 In the early days of computing, not everyone had the means to get into it. It was very costly. This, in my opinion, explains why some devices, like the Mark-8, are so rare. And if you had the cash, the 8080, the Altair 8800, was available in full force. The cost of that IC alone in 1975 was $150 (about $860 now), plus all the necessary support ICs. Remember that your monthly paycheck was $900?
HLK88 I am about to retire from engineering after a great career that was aided by articles such as yours in Radio Electronics. Regards!
Layargaming Yes, and once again, thanks. I apologize for misremembering the name, which also clarifies the reason I was unable to locate it. Once again, thank you.
Layargaming It appears that grey ribbon cables are used in the current version. That is not required.
There are still rainbow-colored ribbon cables available, just like in the 1970s.
When wiring, the colour labelling is helpful. There’s less chance of pin confusion or pin order.
I am Layargaming about to retire from engineering after a great career that was aided by articles such as yours in Radio Electronics. Regards!
At the time I was a grad student in Virginia Tech’s chemistry department aiming for a Ph.D. My work used a DEC PDP-8/L minicomputer to control measurement equipment. After I programmed assembly-language routines I wanted my own computer.