Since the late 60s, Moore’s law has predicted with precision that the number of semiconductors that will fit on a chip about doubles every two years. While this means more and more powerful computers, every year, it also means that old computers can be built on smaller and cheaper hardware. This project from [Bjoern] shows just how small, too, as he squeezes a PET 2001 onto the STM32 Blue Pill.
While the PET 2001 was an interesting computer built by Commodore this project wasn’t meant to be a faithful recreation, but rather to test the video output of the Blue Pill, with the PET emulation a secondary goal. It outputs a composite video signal which takes up a good bit of processing power, but the PET emulation still works, although it is slightly slow and isn’t optimized perfectly. [Bjoern] also wired up a working keyboard matrix as well although missed a few wire placements and made up for it in the software.
With his own home-brew software running on the $2 board, he has something interesting to display over his composite video output. While we can’t say we’d emulate an entire PC just to get experience with composite video, we’re happy to see someone did. If you’d like to see a more faithful recreation of this quirky piece of computing history, we’ve got that covered as well.
And it has a crappy keyboard, for that extra bit of retro authenticity!
And yet I really wanted a PET in 1979.
Me too. I still think about buying one and starting a “Museum of Crappy Keyboards”. It would have a PET, an Atari 400, a Sinclair 1000, an IBM PC jr, the original Radio Shack CoCo, and the original TI 99/4 (which I have *never* seen for sale on eBay – maybe none exist).
I didn’t think the Atari 400[XL?] keyboard was bad, back in the day.
But then, I’d just come off of a Timex-Sinclair 1000!
The Atari XLs had decent keyboards. The original 400 didn’t.
Don’t forget nearly every modern smartphone. I’d say relying on a touch interface that takes up 1/3 to 1/2 of an already limited display is certainly crappy!
Yeah, more phones need slide out keyboards.
+1
I kind of feel that PS/2 keyboard is easier to drive and cheaper than a bunch of momentary switches. Some of those old computers had pretty miserable keyboards. I was a bit spoiled, my first machine was a TI-99/4A, and for the price it had a pretty decent keyboard. It wasn’t as nice as a good electric typewriter, but it was quite usable.
I thought he improved on the keyboard!
This may be a fun project for some, but the “USD2” is quite a misnomer.
Each of the following items probably costs more:
* Empty matrix board.
* The buttons.
* Cabling and connectors.
* Power supply.
* Housing.
But the main thing:
You can get a USD1.5 “Blue Pil”, but I’ve stopped buying them. They do not have ST parts but can be any of about 8 different manurefacturers. with different incompatibilites, and often the top is sanded and STM32F103 is printed on it, regardless of the actual hardware in the chip, which makes diagnosing more difficult. It’s just Yuck!
There are still some sellers of “Blue Pill” boards left which ship with a real STM part. These boards tend to cost around USD3, which is also quite affordable. Cheaping out on these boards really is not worth wasting the time and the headackes of which variant your suplier put on the board last month. (Especially considering the cost of the rest of the project).
I’ve had pretty good luck with the fake blue pills: they seem to be sufficient for my projects. The worst I got were two that had half of the correct amount of SRAM, but the remaining was enough for the project, and the seller gave me a 50% refund, so the project just became cheaper.
You’re satisfied because you got a 70ct refund on a faulty product?
That’s just plain silly.
Once prices for uC boards drop below USD5 or so, the price of the board itself becomes completely irrelevant.
For simple projects, they probably all work. All of them can blink a LED. But once your project becomes more involved you start bumping into the incompatibities.
Some of the “alternate” chips have no built-in debug capabilities. The GD variant needs a lot of boot time because it loads everything in RAM from a separate FLASH die. I’ve had trouble with GRBL, which did not save settings to emulated “EEPROM”, you have to fiddle with product ID’s to get the programmer to work, and there are many other differences.
I’m not saying the differences or “alternatives” are necessarily bad, one of them even runs on 5V, which can be a bonus for some projects. Apparently USB and CAN can be used at the same time (Not for the “original” STM part) The trouble is they sand of the type numbers and print something else on the chip, and that is what makes them fake, and the price for these boards is so small that it really becomes irrelevant, but the risks of trouble and a lot of wasted time is very real.
The Reference manual for the STM32F103 is 1128 pages, complemented with 127 pages for the datasheet itself. Multiplying that with the other chips is sheer madness. (Documenation of the other chips vary, some are around 400 to 600 pages, but I have not looked closely at them all).
One or two of the variants apparently only have a datasheet in Chinese.
The difference between $3 and $1 is rather significant to me. I not infrequently find myself down to $1 or 2 in the “discretionary fun” part of the family monthly budget, and then I can order a $1 or 2 part, but if I want $3 part, I need to wait for next month’s budget or negotiate a temporary budget overrun, which is not something I want to get in the habit of doing. Plus not infrequently I want to be saving up the fun budget for some bigger purchase.
Plus some of my projects just aren’t worth that much to me. Last spring I made a dual USB paddle game controller for pong. I think my cost was about $0.90 for the fake stm, $1 for two pots, and maybe $0.20 for two buttons, and maybe $0.75 for filament. For around $3 it was worth making (or buying if I could get it that cheap, which I couldn’t). But if the microcontroller were $4, that would make the project $6, and I don’t know that I would have spent $6 on something like that. At least I wouldn’t have been happy spending it.
That’s fine for a throw-away project where your time and effort are considered worthless. But for other people, it’s better to spend a bit more money and save themselves a headache by getting board that isn’t a knockoff.
I really wish people would stop calling it Moore’s Law. It is best a postulate.
BTW, my school had a PET and we had fun trying to save and load programs on the cassette.
“the number of semiconductors that will fit on a chip about doubles every two year”
Just wondering what others think? I always though a chip / integrated circuit / silicon die was *one* semiconductor, with many transistors fabricated in it by doping to get P and N type regions and junctions that make up transistors?
In any case, Moore’s law predicts the number of *transistors* on a chip will double every two years, not the number of semiconductors (which stays the same every two years according to Dave’s law, as exactly one :)
The board is outputting plain old VIDEO signal and not composite.
I assume terms like VBS (Video Baseband Signal or Video/Blanking/Sync) or CVBS (Colour VBS, basically Composite) are nolonger as common as they used to be. But even on old electronics, I’ve sometimes seen a “Composite” label, even though the video signal was only the luma (brightness) part along with sync/blanking.
https://en.wikipedia.org/wiki/Composite_video
It’s written in Rust.
The most interesting thing about this project and there isn’t a single mention in the post or the comments!