The 40-Year-Old Version: ZX81's sleek plastic case shows no sign of middle-aged spread It has been 40 years since the launch of Sinclair's ZX81, a device that welcomed countless Brits to the delights of home computing at the dawn of the 1980s. Released on 5 March 1981, the ZX81 was the successor to 1980's ZX80 and, like its predecessor, was based around a Z80 CPU. Both machines also featured 1KB of memory and …
The QL had the potential to be a good machine, but the price with FDD was going to be too much so it was saddled with micro-drives, too little RAM and much of the potential of the M68k wasted. Once you made use of the excellent expansion potential (card inside the machine rather than a box hanging off the outside) to add FDDs and memory it was an OK machine - only by then it was too late and too expensive. I had two with a "network" between them just to see how it worked...
EDIT: Prior to the QL, I as a non-Sinclair user so probably should not be commenting on ZX81 articles, my early days were using a Sharp MZ-80K followed by an Oric1 and an Oric Atmos (friends were in the Acorn cam with Atoms, BBCs and Electrons). While I did have the odd dealing with ZX80, ZX81 and ZX Spectrum machines it was the firesale QLs which caught my imagination.
The Register's own 30th anniversary article for the QL touched upon this:-
[Chief Design Engineer David Karlin] chose the 68008 [which] was built for backwards compatibility, so it featured an external 8-bit data bus and 20-bit addressing. The 68000 [had] a 16-bit data bus and 24-bit addressing. [..]
“I don’t know how robust the decision was,” David admits now, “but it seemed fairly clear that the the 68000 series would be a great platform [..] The problem with the 68000 was basically a pin-count issue. Motorola was pricing it gigantically higher than the 68008, double or treble the price. It was sufficiently high I didn’t even argue about it.”
Rivals’ use of the full 68000 would later come as something of a surprise. If Sinclair couldn’t afford the 68000, how could they? Today, David blames Sinclair’s negotiating skills, not just for the CPU but for a whole variety of logic chips and add-ons: “I question how good we were at purchasing, because people like Amstrad, certainly the Japanese, certainly Apple, who did not have gigantically higher volumes than us at the time, got massively lower prices.”
During 1983, it has been claimed, Motorola cut the price of the 68000 to below what Sinclair had agreed to pay for the 68008. Renegotiating the purchase contract might not have been costly, but adding in the architecture the full 68000 required [..] would have been, so it was decided to stick with the 68008. [..] It’s easy to say Sinclair would have well been better off going with the 68000 after all, but only with the benefit of hindsight.
68000 might have made for better specs, but if they are going to be releasing a machine before they finished developing it then it wouldn't make any difference to the outcome.
I guess with the pace of change, the challenge is to get something to market before the tech becomes old hat.
IIRC, when the 68000 was launched, there was going to be a problem with support chips for it, as Motorola concentrated on getting the processor out of the door.
To compensate for this, they built in the ability to use support chips from the successful 6800 and 6809 families of their product set, which only supported an 8 bit data bus.
I'm sure that it was this ability that allowed Motorola to offer the 68008 at all. They just disabled the 16 bus hardware out (I'm pretty certain that the original 68000 only had a 16 bit data bus, the wider address bus parts came later) and stuck it in a smaller package.
It is quite possible that the silicon came off the same production line, but were in some way test failures, but this is just speculation on my part.
"I'm pretty certain that the original 68000 only had a 16 bit data bus"
That's correct; the original 68000 was 32-bit internally, but only had a 16 data bus. But it was generally only considered a 16-bit processor anyway (the Atari ST and earlier Amigas were all generally called "16-bit", and the 68000-based Mega Drive's badge said "16-BIT"). It wasn't until the 68020 that it got a full 32-bit data bus.
While we're on about the "8 or 32 bit" QL, it's worth remembering though that even the original IBM PC and PC XT *didn't* actually use the 16-bit 8086, they used the 8088 which was the cut-down version with (again) an 8-bit data bus. So though some people complained about Sinclair characterising the QL as a 32-bit machine, it was probably as legitimate as claiming the original PC was 16-bit.
(Some have also noted that the 68000- and presumably 68008- used two cycles for the 32-bit instructions, so one can start nitpicking this, but... yeah.)
Loading the 32 bit instructions would take 4 bus cycles (4x8 bits), so it would be much slower.
IIRC (and I could be wrong, and don't want to have to look it up), there were a mixture of 16 bit and 32 bit instructions in the 68000's instruction set, but I believe that it was a word-addressed processor, so all instructions had to be aligned to a 32 bit word. If there were 2 16 bit instructions together, they could be loaded in one 32 bit word, and fetched together, but this was moot on the 16 bit data bus systems, and doubly so on the 68008.
From my memory of the time, we used to refer to processors with a register size of 32 bits as 32 bits internally, and then call the data bus width the external size. So a 68008 would be called 32 bit internally, 8 bits externally, or written as 32/8 bit system.
Nat. Semi. coded this into the names of the processor, so the NS32016 (although it was originally launched as the 16032) was 32 bit internally, 16 bit externally, with the middle number used to represent the generation of the processor. There were 32008 and 32032 processors in the family, which followed the pattern.
Sorry, I should have been clearer; I meant it would (AFAIK) take 2 bus cycles if the data was already available internally, presumably held on the 32-bit data registers. I also notice the Wikipedia article says that the ALU is only sixteen bits (but it has another two used for addresses or something).
That said, you sound a lot more knowledgeable about such things in general than I am anyway(!)
FWIW, in light of what you said I also noticed that the Wikipedia article claims that "Motorola termed [the 68000] a 16/32-bit processor".
And the "ST" in the 68000-based "Atari ST" also supposedly stood for "Sixteen/Thirty-Two" (though some suspected it was actually "Sam Tramiel".) I'm guessing the less-well known "TT030" workstation is so-called after the "Thirty-Two" bit 68030 CPU...
We had a family BBC B which predated the school issue by a year. Our son took it to heart and so he and my wife became programmers of note which led to our boy being the class lead on computers.
I went for hardware, and came up with some grotesque builds, and some elegant ones now and then.
It was about then that my family and I stopped talking computers because we didn't know what the other side was talking about.
Looking about me, I fear that divide has permeated wider than I ever thought through the industry.
Agreed entirely; in principle the ZX81 isn't that far advanced over the ZX80 — it's a larger ROM, an extra counter and some corresponding logic to trigger WAIT and NMI appropriately — but in practice it's a huge redesign, shrinking all the discrete logic down into a single ULA. The entire machine is four chips: ROM, RAM, ULA, CPU. For 1981 that's pretty spectacular.
... and then rolling that on into the Spectrum just a year later is amazing.
This post has been deleted by its author
Indeed. As far as I'm aware, the hardware design of the ZX81 is almost the same as the ZX80's at a logical level (aside from the addition of WAIT/NMI). The biggest change is that much of it was re-implemented via a single ULA.
Even so, the similarities were such that you could (almost) convert a ZX80 to a ZX81 simply by replacing the original 4K BASIC/OS ROM chip with the same 8K ROM that the ZX81 used. Sinclair sold this as an official upgrade, complete with an updated keyboard overlay.
The only thing apparently missing was the non-flickery ("SLOW") display mode, which the ZX81's aforementioned WAIT/NMI hardware upgrade was needed to implement.
IIRC, a running program had no display at all on the ZX80. Or only when you reached an "INPUT" statement.
ZX81 paused running your software for long enough to display a TV screen picture from top to bottom, then executed in the gap until the screen needed to repaint - 25 or 50 times a second. And not the border around the (not) bitmapped display, I suppose. A bit like "Ceefax" being digital TV text in the form of binary dots in a few lines in the top of the picture. A copious hypertext magazine, almost too small to see.
FAST got things done quicker, but you couldn't watch, again.
Not sure about the Ceefax analogy, but that's otherwise broadly correct AFAIK.
Since the CPU was used to generate the display, the ZX80 could only output an image when things were otherwise idle or waiting for input (i.e. essentially the same as a ZX81 that could only operate in "FAST" mode).
The WAIT/NMI logic on the ZX81 gave the user the choice of a continuous display at the expense of speed (hence "SLOW" mode) by interrupting the processor whenever it was needed to generate the active parts of the display. As you say, you could always switch back to the full-speed non-continuous display via "FAST" mode.
"a single ULA"
Courtesy of Ferranti. Ferranti devised their ULA: a chip with a load of gates in the lower layers for which the customer-defined top layer would determine how the gates were interconnected. It was ideal for customers who could not afford to have custom chips made and was cost effective, even at relatively low volumes.
Which basically confirms the adage: wait for version 3. We had a ZX81 and that keyboard was truely awful. It's clear the company was using the cashflow to fund the next version and the Spectrum was a much better machine, but, of course, it was to be the Amiga that became the most successful.
The Planet story has echoes of the Sinclair computers with the Gemini being a nice idea with a few design flaws and crappy software support. I would hope the subsequent versions really do show improvements as my Gemini is now basically just a paperweight.
Put Sailfish on it.
My Gemini is still going strong. Had another update in February, to the Sailfish which is becoming really very nice.
Wouldn’t want any of the subsequent Planet devices, and even if I thought they were any good I wouldn’t want to be stuck with Android or a half baked Debian attempt.
I also have a ZX81 which has had an update (ZXpand) more recently than the Android Gemini.
I might put Sailfish on it if the screen would stay on for more than a couple of seconds. I'll see if recovery mode is stable enough for this but if, as I suspect, it's a hardware issue (and I'll already checked the cables as well as I can), then there's little I can do.
I like the idea of the Sinclair computers but I bought a SOL with a Kansas City tape interface because it was S100 based and gave me the ability to build my own boards for it ... The SOL is a beautiful computer although the keyboards need a little maintenance every 10-20 years - it's still running.
It would be interesting to build a SOL look-alike based on a Raspberry Pi (LOL).
I'm toying with the idea of either getting a Pi-400, or the Pi-4 desktop kit. The Pi-400 looks nice, but appears to only be available with 4GB RAM, whilst the desktop kit is available with 8, and I'd like to be able to use the Pi for hardware hacking...
There have been rumors about the 400 getting an 8GB option, but it depends how popular the 4GB version is. I'd wait to see if they announce it in 9 days, as they sometimes have announcements for pi days (in the MM/DD calendar that's 03/14, March 14, and in the DD/MM calendar it's 22/07), and if they don't you can make your decision.
That said, depending on what kind of hacking you're planning, you might not need very much memory. Most of the hardware-control projects using Pis are light on resource usage. Back when Pis had only 512 MB, they were still used almost always for hardware control. The increased memory is useful if you want to use it as a desktop or server with a lot of caching, or if you have a lot of data coming in from the hardware which you need to process, but otherwise you probably won't need it.
A Raspberry Pi with an onboard RP2040 microcontroller is what I’m looking forward to. The Cortex A series MPU used by the Pi is far from ideal for control, but the RP2040 is short of compute horsepower. The two in combination should be a winning proposition.
I hadn't thought of that, but now you've mentioned it, I want it. Especially if the Pi can be shut down and restarted by the RP2040. The most limiting factor of the Pi for me is the power consumption, meaning I can't run them off batteries for a long-term project. Having a low-power coprocessor which could run some native code which could invoke the boot process and spin up Linux whenever I need something more complex could be really nice.
The Pico, which has the RP2040 on board, has some nice features, such as programmable state machines, IRQs, lots of GPIOs, I2C and SPI, and built-in PWM. I've not looked too closely at what the Pi has, but I reckon you could get the two talking to each other at a fair old rate, and have the Pico offload heavy tasks to the Pi, and the Pi offload complex control tasks to the Pico, in near real-time.
I was thinking of using the Pi as a development environment for the Pico. The toolchain can be made to run on Windows, but it's a bit clunky, and from the documentation, it's pretty clear that it's Linux first, Windows/Mac second. I could set up a dual-boot partition on my desktop machine, but it would be just as easy to keep it all separate and have a dev environment on the Pi instead. Experience tells me that such things run a lot more smoothly with plenty of RAM, and whilst I doubt I'll be trying to do anything fancy like virtualising hardware on a Pi, or running anything as heavyweight as Visual Studio is on Windows, I feel like it should have the headroom...
I should have been clearer, I meant to run the development tools to program the more modest Pico, so for instance, using the GPIO pins of the Pi to interface with the debug port on the Pico. I quite like the idea that you can use a second pico as a debug probe for this purpose. At a couple of quid each, sacrificing one for this purpose is a nice idea.
One does not simply cite an old home computing magazine in a forum, one has to post a link to archive.org as well.
... and now I think of it, I even have my old fx785p, a clamshell thing that was more progammable and even allowed you to work in (its) assembly language ... not sure why it didn't have an exciting large marketing number like the 3600p, given it had more features. And for that, the manual are even in a non-tatty condition ... especially the assembly language one, which saw little use.
The 702P was the best one I ever had - but some git nicked it when I was at Uni.
I tested my old FX-19 a few weeks ago - and it still works perfectly. I am utterly convinced that that thing changed my life!
When I was at school, I was in the 'top form', but at the time I had no idea that I was, or what it meant, and I often got in trouble. But when I started my O Levels, it was the very first time calculators were allowed, so my parents bought me the FX-19.
Suddenly, I started playing with numbers, and began to understand them - especially logarithms. I can remember the look on the maths teacher's face when overnight (almost) I went from being a right pain in the arse who never got anything right to being able to understand three-dimensional trigonometry. Even the top-of-the-class group were asking me how to solve trig problems.
God only knows how I'd have ended up if it hadn't been for that FX-19. But I am absolutely convinced it saved me from... well, I don't know what.
Yeah, sorry, I meant to upvote but fat fingers. Fixed it. Icon for me.
I guess it was a one of the chimp covers they did for Sinclair computers, right?
Weirdly this isn't Personal Computer World, either. This was a weekly one that later on used to have a comic strip from Automata Software on the back every week advertising their current wares using the Pi-Man and Groucho Marx. Alas, I haven't kept any of the later issues - or the PCWs either, many of which I had. Most of them went into a skip when we moved house. Given the prices on eBay these days, I probably chucked out enough magazines to buy a house with if I sold them now :(
I was working for a firm that had it's own in-house service dept and my £49 pre-bult ZX81 (bought from WH Smiths, along with a £19 16K RAMPAC) was handed over to one of the service guys to see if he could fix the "wobble".
He did - by removing the edge conenctor from the RAMPACK and soldering the ZX81 edge connector, directly to the RAMPACK (by way of many short wires).
And it then worked very well after that !!
Soldering the RAMpack is also, apparently, what my Dad did at the same time that he got a proper mechanical keyboard for it (something like this) and transferred the internals. Don't recall *ever* having a problem with RAMpack wobble after that!
I started saving my pocket money for the ZX80 and before I could afford it the ZX81 was launched. After begging the advance to get the ZX81 I didn't have enough for the RAM pack. But I well recall the advice to use a frozen milk carton - it put me off even wanting the RAM pack! Space Invaders in less than 1K - amazing, but you had to shoot the left and right half of the invaders separately!
My solution was to separate the keyboard from the rest of the system (well, add an external keyboard, I guess).
I took a Tandy membrane keyboard I bought in one of their 10% off per day sales, opened it, scratched and repainted the conductive paint traces to match the ZX81 keyboard, and extended them to include the space bar. I connected this to the ZX81 wlth a ribbon cable that ran through a slot I cut in the front of the case. I took a photo of the ZX81 and cut it up to put key legends on the keys.
Fitted into a case built from thick styrene sheet, with a switch for power, and it worked a treat. Once you isolated the keyboard from the rest of the system, everything became much more stable, especially if you had any additional expansion cards (I had the Quicksilver AY-3-8910 sound board, together with a second modulator to put sound through the TV). It was not as elegant as the system in the article, but it worked very well.
I never got round to putting Quicksilver's Hi. Res. graphics card in, but I did find a neat trick to implement programmable graphic characters with an extra 1K of static RAM on the ROM side of the R7-R14 data bus isolation resistors and by manipulating the I register.
"Although Sinclair wasn't the only one to play that game, he did rather have form in that area!"
On the other side of the world, the machines came in at something over 8 times the UK price (A BBC micro was something like 2 months of my father's salary as a senior primary school teacher, so understandably, we never got one - even if you had a ZX81 you were well off or a computing fanatic)
But at least you could buy them. It was illegal to sell/bill for products you didn't have in stock
And around the same time I built my Maplin 300 baud modem for my Atari - I was always an Atari user rather than a Sinclair.
I know it was naughty, but the electronics department at the company I worked for had an EPROM programmer, and for a while there were alternative OSes for Ataris that could be piggy-backed on the existing OS and switched as required. I burnt dozens of chips and built piggyback boards for Atari users using that programmer.
Happy days, indeed.
As I've mentioned here before, I did a similar thing with character EPROM on Kaypro 2X. It had scandinavian characters mapped to {}[]/| so any programming was a pain. So I read out the EPROM, edited the characters and burned to a fresh EPROM, soldered EPROMS on top of each other bar CS pin and then wired a toggle switch to CS pins. Worked a charm.
Mine came with the 'Learning Lab', a rather thick book and bunch of cassette tapes.
Seem to remember that Memotech did a rather nice range of modules for it.
Also remember hand entering the code for FIG-Forth; a fun language which used reverse polish notation.
Also it made a rather effective radio jammer!
The ZX81 was genuinely my first own computer, bought with the earning from a holiday job when I was 16. It kept me occupied until I bought a Spectrum at Uni. Mine is long gone now but I acquired one a couple of years back that a friend of a friend was throwing out.
The edge interfaces on both machines were vey useful, with assistance from a couple of hardware guys I had quite a few peripherals including a 4 fire button fully programmable* joystick for playing Halls of the Things on the Speccie.
*Programmable by attaching crocodile clips to a custom PCB.
I never had a ZX81 - at the time I was going from Commodore PET to a VIC20, but I did go on a school French exchange where the family I was staying with had one. I remember spending what seemed the best part of a day typing in the code for "that program" that had the tyrannosaurus rex advancing menacingly from a 3D maze (without the benefit of the muscle-memory ZX81 owners developed for which keys corresponded to which BASIC commands) only to have it totally crash when I tried to run it. No idea why I didn't save it first, I think I was just used to the Commodores where the worst you'd get if something was wrong with a BASIC program was a ?SYNTAX ERROR IN 270
My ZX81 came from WH Smiths back in 81/82. After the ZX81 came an Oric Atmos, it was a revalation to have a proper keyboard. At school we had BBC Acorn and a bit later IBM PC. Then I kinda lost interest for a good few years. Computers could not compete with girls, late teenage years, what do you expect.
My ZX81 self-build kit was a present from my parents when I was 14.
It had been packed in such a way that the pins of a couple of the ICs had punched through the membrane keyboard, rendering it useless - it took weeks, and many irate phone calls by my mother, to get a replacement.
Eventually it did all work though, and I spent many a happy/frustrated hour tinkering - I built it into an old Linotype Paul case with a full size keyboard, inverted video output (white text on black background) and a ram expansion on a ribbon cable (so no wobble).
40 years later and I'm still a serial tinkerer and coder.
A few years ago a school friend got me an autograph from Sir Clive with an apology for the damaged keyboard!
I still have a working one, it has a converted to digital output in place of the tuner. I also have an excellent emulation on iPad, by a Kevin Palser..
The ZX81 was life changing for me, no exaggeration. I can still remember the feeling of seeing the thing work for the first time. It was special to be around when this movement was born and took off.
Although it's in a "proper" (well sort of, rather clunky) full size keyboard case, we got sick of the membrane very early on. Plus the case makes the (non Sinclair) RAM pack much less likely to wobble.
Chess in 1K of RAM, 3D Monster Maze, Catacombs, etc etc. Even had a Dig Dug in Hi Res by the time it got updated (to an Acorn Electron). No idea how someone managed to get that to work. I remember it being a swine to load from the tape though.
Nerd attack!
To output graphics the ZX80 and ZX81 rely on the Z80 — it jumps to video memory and attempts to execute. The video logic (discrete or ULA, respectively) spots that based on the high bit of the address and from then on it pushes whatever opcode the CPU was going to fetch into a latch and forces a NOP to the CPU. It's a Z80 so it follows up the opcode fetch with a DRAM refresh cycle, during which the video logic forces the just fetched opcode plus a three-bit-counter on top of the low 9 bits of the refresh address and uses whatever comes back on the bus as the next 8 pixels of video output.
The internal RAM is static, so there's no actual refresh.
(and I've simplified that description a little by omitting HALT; there is some intelligence in there so that if the CPU reads a HALT then it actually gets a HALT, which is how the 1kb machines don't have to maintain a full 32x24 text buffer unless the user has actually filled the display with characters)
So, the two main options for high-res output are:
(1) find a suitable set of locations in ROM to cover a decent number of the possible range of pixel patterns, and use that as your target for the refresh cycles — on a Z80 you can set the top 9 bits of a refresh address, as DRAMs at the time had 7-bit address buses and therefore the Z80 increments across only the low 7 bits. This is how any high-resolution game you played that had odd graphics that weirdly gained or lost lumps as they moved worked.
(2) move the refresh address into RAM. Technically relies on a quirk of the bus, but works with the internal 1kb of RAM and with many expansion packs. But not all. And it's hard to get that much into the internal 1kb. So this approach is better in terms of fidelity, worse in terms of compatibility.
Source: I wrote an emulator, of the attempts-to-be-entirely-accurate variety.
I'd forgotten about more text on the screen = the thing runs slower and slower!
I think I remember a hack existing to display just a couple of lines of program code instead of a full screen listing to work on, very satisfactory...
And at least one "type it in" program listing, maybe hexadecimal, with very unfortunate long program lines - that slowed down input a lot, too.
Some of this may be on the ZX Spectrum, though.
I have some questions as someone far too young to have seen one of these. When you say "Chess in 1K of RAM", was that the computer displaying a board and letting you make moves on it, or was it actually playing against you?
Also, since this machine is the one that always gets brought up in the battle of who had the worst specs in their youth, I have a question that probably a lot of you could answer while having a similar competition. What was the most useful or full-featured program you got one of these to execute? For fairness, we'll have two categories: just the 1K of RAM or the full 16K. Most intriguing or complicated gets a prize once I've tracked you down using only a ZX81, which will probably take a while to get to read this page.
Chess in 1K - The computer did play against you.
So you had the graphical display, the rules of what moves are valid, and the decision making algorithm all wrapped up in 1K.
IIRC, the book "Mastering Machine Code on Your ZX81", took you through every step of the code in one of the later chapters.
Caveat - It wasn't pretty, it couldn't handle castling, and it wasn't even all that good at winning, so if anyone's up for a real coding challenge...?
I bought one, used it twice, gave up.
Getting it to talk to the TV was hit and miss (mainly miss), connecting to the tape recorder never worked, and the fake button keyboard membrane was a POS
Absolute waste of money
Its in my spare parts bin, I guess a museum will get it one day
...tapping in code from this month's magazine. Always wait for next month when the inevitable correction to the code is published.
Although troubleshooting those typos became laid the foundation for the career I have today, so I cannot complain too much. Cheers!
There was a word processor for the Atari computers called Speedscript.
I've still got the book for it. You had to type it all in via a special editor, and CHRIST it was tricky. The book is just page after page of hex code, and it's the size of a short novel. Fortunately, each line (or page) had a checksum, so - or 'but' - if you made a mistake (and you did) you had to try again for that page.
But once it was in, it was a mighty powerful WP for the time.
And yet... the MK14 (which I still have, in a box somewhere) was the machine with which I first learned hardware expansion logic... a video card, as I recall. I do remember waiting for a couple of 256x4 bit memories in a shop on Edgeware road, cost me half a week's wages. Mainly remembered for the fact that while I was waiting, the door slammed back, and a chap rolled in on a unicycle, made a very offensive racial remark (even for the late seventies!) and rolled out again. Strange days...
Ah, the memories. The ZX-81 was also my first encounter with a computer, my grandfather bought one soon after they first came out. He was a former codebreaker for the Admiralty during WW2, and university professor at Manchester where they had a machine he described as being as large as a room. I still feel particularly sad that he never got to see how modern computing turned out, as he died in 1987.
The one thing that sticks in my mind with the ZX-81 was typing in and running the code for something called "Life", which generated some really interesting (blocky) patterns that evolved over time. It left a huge impression on me, and helped make me into the geek I am today!
I think it was an implementation of Conway's Game of Life.
You can see an online version here for reasons of nostalgia.
Most interesting memory. Did your grandfather encourage your exploration of the game of life? Did he run it himself?
[Conway proved that GOL was Turing complete in 1982 according to wikipedia. Your grandfather could well have been aware of the implications of that if he was on the maths side and following the literature]
The Timex Sinclair version was the TS 1000.
My first computer. I can still remember the first time I wrote a program that used the entire 2k of RAM.
Unfortunately I trashed it many, many years ago. Found one a couple of years ago and a 16k RAM expansion pack, and spent a fun day getting it working on a modern TV.
Unfortunately discovered it really was only really good for a nostalgia hit.
Have a look at
https://medium.com/age-of-awareness/girl-computer-31ecd328bc53
An account of the impact of a TS1000 on a 7 year old... with some food for thought in the final paragraphs. Was on the orange and grey IT message board the other day and I thought it was interesting.
It was just "Sinclair" in the UK, although they were manufactured under contract by Timex. "Timex Sinclair" was the brand used for the otherwise near-identical North American version of the ZX81, known as the Timex Sinclair 1000 or TS-1000.
Interestingly, there was also the obscure Timex Sinclair 1500, which was also essentially a ZX81 but with a Spectrum-style silver case and rubber keyboard and 16K onboard.
(Apparently a complete flop- most likely because it didn't come out until mid-1983 and cost $80 at a time when numerous more advanced machines were available, and the US market had become so competitive that many of those were being driven below the $100 mark).
There was also a Timex Sinclair 2068 which was based on the ZX Spectrum, but with a number of improvements that made it incompatible with the latter. (And which didn't help it succeed in the aforementioned cutthroat US market, which Timex were forced out of shortly afterwards).
Timex's Portuguese arm also released the Timex Computer 2048.
Damn, my first computer is 40? How time flies.
I fondly remember glueing the damned extension and my father soldering the power cord. And I still have the notebooks with all kind of fun programs, including my very own drawing of an image of the Statue of Liberty pixel by pixel. POKE-ing was a lot of fun.
You laugh.
I was at Data General when we started designing Ethernet boards (I designed the one for the microNova, codename "JOEY" -- "KANGAROO" was the one for the Nova).
I remember we had to type in the MAC addresses for some reason. 08-00-1B-E0-00-01. I still have our company prefix memorised after 40 years. The "E" in the most significant nibble of the second part was for "Engineering" boards. We figured that the production boards would take a while to get up that high.
I also built my own 6802 based boards. Off the shelf computers are fine for the masses, but I was cheap, we had used the 6802 in another project and there's no thrill like seeing a monitor prompt on a board you built yourself. They ran the ham radio repeater for the DG Amateur Radio club.
Yeah, I'm wicked old.
You still get sort of the same feeling even now when you build a new system and - when you hit the button that first time - you get the appropriate prompts.
It's a bummer when you don't, though. Like the last time, when I didn't remove (because I didn't notice) the film over the NVMe and the bloody thing wouldn't boot. There should be a law that says no shrink wrapping allowed on small components.
08-00-1B-E0-00-01. I still have our company prefix memorised after 40 years.
08-00-2B-xx-xx-xx was DEC, and I remember being told they had a nifty way of manufacturing the MAC address PROMs for the DEUNA ethernet interfaces (and some of the ones that followed) so that they all were unique. Later boards had the address in EEPROM. Swapping a DEUNA link board usually meant moving the MAC address PROM over to the new board; some 3rd party programs used it as a licensing identifier, although DEC warned against this.
Running DECnet IV the host MAC address changes from the hardware address in that PROM to AA-00-04 plus the inverted binary value of the DECnet address. Could cause a bit of fun with multiple network interfaces and switches that block that.
I managed to get an external bus board, that had a row of edge sockets on it. I then made a flexible cable to connect it to the back of the ZX81. The company that sold the bus board also sole 32K RAM Packs! Which therefore did not wobble. 32K was fairly useless, as the BASIC ROM could not use it. I used it as a frame buffer for outputting 256x512 graphics on the printer.
I also then built an I/O and SRAM board, that gave me 24 I/O pins, and another 8K of ROM (actually battery-backed RAM). More jiggery-pokery lead to white-on-black text, and user-defined characters in the SRAM chip.
All done when I should have been revising for my A-Levels....
In 1984 I have borrowed a ZX81 for a weekend. It came with one game, a LearJet flight simulator. Now, by that time I have already played the Star Wars arcade game, so I was expecting some pretty funky line art the very least. After half an hour of loading from the tape drive to the memory cartridge, the game was up and running, and lo and behold it was text based like Zork. “You are standing in a grassy meadow…” well, okay “you are parked on a taxiway.
Sadly, at the time English was yet to become a familiar language for me, so it was dictionary flicking time. I have managed to fire up the engines and make it all the way to the main runway, when I stood up to get a drink, and the memory cartridge shifted and cleared. Half an hour later it was loaded again, and shortly afterwards the memory cartridge shifted again. So, I have made an executive decision, hopped on a bicycle and went for a hoon. For fun. Never touched the ZX81 ever again.
This article was a good memory jog. Wow, has it really been 40 years?
I remember visiting my uncle back in 1980, who showed me his new ZX80. I decided there and then to start saving £80 for a kit. Being just 14 it took over a year to reach the £50 mark by which time the ZX81 was out; am I right in saying that you had to pay extra for a mains adaptor that was included with the pre-built version? I ordered one as soon as I could and although it didn’t arrive on time, I don’t think the delay was that long.
Once assembled I remember the proud moment of the great switching-on and the grey screen that greeted me: nothing could persuade it into life. After much disappointment I carefully checked that all the components were correctly placed and oriented, which they were, and then went over the board touching each soldered joint with the iron to remake the connection. Success! For the first time, I was greeted with the cursor that was a black square with a “K” in it.
I went on to buy a third-party, bag-of-components that became a 32K expansion pack (yes, double the Sinclair version). It had no case, just an exposed circuit-board but I suspect that the lack of weight is what kept it wobble-free. This all eventually ended up inside a keyboard kit that I had to butcher to get most of the memory board in. That kit was so cheap that the keys came with clip on, clear plastic tops under with was placed a legend that was cut out from the printed sheet of paper provided. One thing was for sure: when I told people that I had built my own computer, nobody ever doubted it.
I eventually sold it to help finance a ZX Spectrum. Looking back, I’d love to have it now to show to my children. Throw in the long-gone Sanyo cassette-player and I could teach them Z80 on the assember I wrote.
a spectrum.... I think my brother had the ZX81 when my father died, no idea why dad kept both of them and quite frankly I'm surprised either survived the hammering a couple of teenagers could give them.
But then without them, I would never have got into Z80 machine code (mainly 'cos the BASIC was so flipping slow compared to my friends BBCs (bloody rich basts) ) and without the Z80, learning to program multi axis machine tools and robot controllers would have been damn near impossible (insert quick whinge about how kids today have it easy with CAD/CAM programs)
But the thing is... kids could pick it up, learn to do and to make it do simple things, then learn the more complex stuff, such as getting a black blob scrolling across the screen to hit another black blob, then figuring out howto do hit detection.
And that readers is the basis of every video game ever.
When the ZX80 came out the electronics company where I worked acquired one to evaluate. We also got a circuit diagram by undisclosed means. It was an engineering marvel that we poured over for some time. They had performed amazing tricks to get the part count down. I recall being surprised to see the memory bus lines going to the keyboard. Soon we realized they were using the upper memory address lines to scan the keyboard. The software driving this must have been bizarre.
With the ZX81 our analysis was quick as from a hardware circuit perspective they had just rolled up the relatively few discrete logic ships into one which would have made it all that bit cheaper so more profitable.
If anyone's interested, I have a video where I live-code a simple game (all games were simple on it) for the ZX81 here: https://www.youtube.com/watch?v=WqVfyyZbIvo
One thing that's often missed is that it came with a simple but reasonable IDE, which made programming a surprisingly decent experience. The terrible keyboard was a good match to the keyword-at-a-time input; pressing keys was difficult, but you got a lot of value for each keypress. It also had a lot of features to both cope gracefully with running out of memory (inevitable with 1kB of RAM) and to let you squeeze functionality into every byte. The Basic dialect was dreadful, however. No if...then...else!
Still have a ZX81 and wobbly memory expansion pack in the loft. The memory pack is an independent one (i.e. not Sinclair's) that was stabilised with Velcro. I also had the Memotech keyboard for it (with real key switches) - gave that to a nephew when I moved to bigger things. The bigger thing being an Acorn Electron (also in the loft, along with a a 3.5" floppy expansion, Plus 1 expansion, sideways RAM pack, co-pro and a 5.25" floppy disk unit; not sure if the display is still up there). That served as my home computer for several years, until it was superseded by an Amstrad 640. I was originally going to get the Archimedes, but I needed PC compatibility as well (it was becoming the standard for work) and Acorn discontinued their adapter in favour of hardware emulation.
Those were the days...
they code things differently there.
I got a ZX80 in retrospect, but my first computer was the 16k Spectrum. I didn't want it, I wanted a record player and a guitar but my dad wanted me to have a career in computing.
I thought at least I could buy some games for it to compete with my richer friends who had Ataris, but my dad wouldn't allow me any games until he could open the manual randomly and ask me to recite and explain the page. It was a huge flipping manual, including mathematical functions years ahead of my school course work.
I was writing my own games before I was allowed to buy one. I never knew there was a machine code book so my games weren't impressive. I found out how difficult the Turing Test is, I "developed" a diary/organiser, I did my first hardware upgrade. Nothing as impressive as most of your reminisces, but enough to trap me into 'the career'.
I still think I would have made a better rock star. (Ironically decades later I turned down a testing job at RockStar).
I sent off for the ZX80 kit and an assembled ZX81 was delivered. A year later I attended a computer show in London, where Haven Hardware (father & son) were selling a colour add-on for the ZX81. Shortly after the Spectrum appeared. It's hard to explain the thrill of that early computing and the magic of colour. I've forgotten the Poke number you could enter at switch-on, to extend the keyboard beep to give much needed feedback for 'touch-typing'. Such larks Pip old friend, such larks. Gavin.
I still have a ZX81, somewhere in the attic. It still worked when I last brought it out to show someone, about ten years ago. I also have a Casio fx201P - a chunky programmable calculator from the mid 70s with a gas discharge 7 segmant display. That also works and I have the original manual, with a load of example programs.
I built mine form the kit, thereby saving £20 to buy the Ram pack, which I fixed with blu tak.
It didnt run on initial start up but you could return it to Sinclair for testing.
They reported no dry joints and that the Z80 CPU was faulty, so provided a replacement.
They also sent me two spare keyboards as they had detected n error with that too.
Never had any issues after that, £D monster maze worked a treat.
It still powers on but the screen output is going now.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
