CH32V003 Makes For Dirt Cheap RISC-V Computer

These days, when most folks think of a computer they imagine a machine with multiple CPUs, several gigabytes of RAM,  and a few terabytes of non-volatile storage for good measure. With such modern expectations, it can be difficult to see something like a microcontroller as little more than a toy. But if said MCU has a keyboard, is hooked up to a display, and lets you run basic productivity and development software, doesn’t that qualify it as a computer? It certainly would have in the 1980s.

With that in mind, [Olimex] has teased the RVPC, which they’re calling the “world lowest cost Open Source Hardware All-in-one educational RISC-V computer” (say that three times fast). The tiny board features the SOIC-8 variant of the CH32V003 and…well, not a whole lot else. You’ve got a handful of passives, a buzzer, an LED, and the connectors for a PS/2 keyboard, a power supply, and a VGA display. The idea is to offer this as a beginner’s soldering kit in the future, so most most of the components are through-hole.

On the software side, the post references things like the ch32v003fun development stack, and the PicoRVD programmer as examples of open source tools that can get your CH32V computer up and running. There’s even a selection of retro-style games out there that would be playable on the platform. But what [Olimex] really has their eye on is a port of VMON, a RISC-V monitor program.

When paired with the 320×200 VGA text mode that they figure the hardware is capable of, you’ve got yourself the makings of an educational tool that would be great for learning assembly and playing around with bare metal programming.

It might not have the timeless style of the Voja4, but at least you can fit it in a normal sized pocket.

Thanks to [PPJ] for the tip.

36 thoughts on “CH32V003 Makes For Dirt Cheap RISC-V Computer

      1. If you use something like minicom, screen, or putty to it, then a tiny board like this, can easily scroll down your name, and do a lot more. Look at the Raspberry Pi pico, or most other tiny MCU boards. You can install something like micro python, basic, JavaScript, or some other language on those little boards.

  1. I do find such things interesting, but I miss the times when people had worked with S-100 boards or ECB boards.
    Real PCBs or through-hole boards, in short. With components in a reasonable sized PLCC form factor, for example. Like an 80186.
    These miniature SMD parts of today don’t look and feel any precious anymore.
    There’s no physical bond between the tinkerer and the hardware anymore, I mean.

    1. Indeed. Those DIP packages tended to pierce your fingers, or your buttocks when you accidentally left one on your chair. There is none of such intimacy with SMT stuff.

      1. Hello my son, please let yourself be enlightened by my wisdom: you can use DIP packages as SMDs, too.
        Just bend the little legs a bit sideways and you’ve got an SMD part somewhat superior to an normal SMD part. Amen. 🙏

    2. The person who has handbuilt a small sailboat might feel thus about the building of ships in a bottle one fine afternoon, while sailing in the local lake.

      Hardly would he consider that, some dark and stormy mindnight, the builder of ships in a bottle would feel thus about hand-built sailboats, while fiddling to get the rigging just… so.

      Different people engage in different parts of the hobby for different reasons, often with a great deal of overlap. I, for example, do agree that messing about with very-early-era solid-state components is quite special. I grew up with a nearly unlimited supply of surplus milspec junk, courtesy of an uncle who worked in the demil industry. By definition, anything available to me had last been “interesting” two to three decades earlier, but that just made things more hands on. I won’t say “easier to work with”, by any means, and when a component gets to the size where it can easily sever your fingers if it slips, just maaaybe it’s a little “too” hands on :)

      At the same time, working with the modern smt stuff is also quite fun. While there is a lot of overlap, they’re almost two different hobbies. I personally find both to be interesting (under different conditions), but each person is different, I guess.

    3. Just popped a SOT363 off a board. It wasn’t bad. Used a $30 hot air pencil. Judging by inflation that’d be a $10 soldering iron in the 1980s.

      Most of the surface mount chips are bigger than that for hobbyist stuff. And if you use flux with solder in it you can flow solder in a toaster oven, on a hot plate, or whatever. You can also drag solder and use a braid to clean it up.

      Don’t be afraid of SMD, move with the times. For every naysayer complaining that EV is killing hot rodding there is a guy making a 400HP Mustang or Audi re-powered by a junkyard Tesla motor and a Chevy Volt battery.

    4. Two things I really liked about the early Arduinos
      – Complete enough ecosystem that you could pick it up and use it, and learn the in-depth parts later if you wanted, but meanwhile you could make Real Stuff
      – Through-hole components, even for the -328 MCU itself (NXP did make one Cortex-M0 in DIP format.)

      When I was a kid I could solder stuff easily (lead solder helped with that); these days I wear reading glasses and would rather not have to buy a microscope to solder surface-mount stuff, but I’m perfectly happy to buy development boards that break things out to connectors.

  2. I’ve been looking at this chip recently because I’ve been working on a little cheap project. My first inclination was to lean into the CH552, because those are so cheap, but the CH32V003 series (and it’s bigger sibling, the V203) are also super cheap! The 003 can be had in an SOP8 for as little as $0.19 in qty 5.

    I feel like we’re about to see another explosion in inexpensive gadgets.

      1. Right, the 8052 was better anywayl
        8052-AH BASIC is so cool! 😎

        I’ve got the original MCU here, but the latest interpreter can be installed into, say, an AT89S52.

    1. And 10c in qty 50. That’s only a $5 investment! Or $7 for 50 of the 20 pin package.

      The 003 only has 2k RAM and 16k flash so it’s a bit bigger than a zx80 but smaller than an early Apple 2. Comparable to the ATMega328 in Arduino Uno (which cost about $3) as RISC-V code is denser. But faster than all of those as 32 bit not 8, single cycle execution, and higher clock speed (48 MHz).

      Several new models (002, 004, 005, 006) just came out with up to 4x the RAM and flash.

  3. Cool.

    But…. PS2 keyboard? Is there someone still making those or is it presumed that e-waste and ebay will supply them for the foreseeable future?

    I recall that back in the day, some keyboards supported both USB and PS2 (with USB/DIN adapter). Maybe that’s still the case?

    1. Hey, it’s the USB keyboards that are essentially e-waste!
      On PS/2 and DIN keyboards, you can press multiple keys no problem and have no polling-issues, whatever.
      They’re much better, protocol-wise.
      Good USB keyboards are dual-protocol still and can still work as PS/2 keyboard through a passive adapter.

      1. I once bought an active adapter to connect the good quality PS/2 keyboard I had to USB. (20+ years ago?) This did work, until I had to change some bios settings and found out the keyboard did not work without a running OS.

    1. Good point that it can digitally do anything a 555 can do analogue (certainly the original 50 kHz ones), plus much more, and at lower cost for the chip AND no external passive components needed (which could often cost as much as the 555, especially if low tolerance).

      But it is also more capable than a lot of early 6502 and z80 general purpose home PCs.

      1. I think you missed his point. It wasn’t the processing power, it was the I/O. With only 8 pins, minus 2 for power and ground, and possibly 3 more for VGA. How do you interface?

        The V203 he mentions is only 0.67 ea in QTY 20. Seems a reasonable price for an extra 12 pins. Thats less than 2 cents per pin. No matter what multiplexer or hub you might devise to make the V002 work, the V203 seems like a value 🤷🏼‍♂️

        1. I messed up my math, but I see there is one pin left after VGA and PS2. So it would have to pull double or triple duty if you wanted to communicate. I guess an I²C bus might fit on it. For getting sensor input etc. Then writing out to a non-volatile memory etc.

          I see now this is a “solder it yourself” kit for cheap. Very nice. And I get why they would want an 8 pin chip for that. Shame the PCB couldn’t have provisions on the reverse for the 20-pin. Choose your own difficulty so to speak. Or make the kit a £2 kit and include a 2nd PCB and 20pin chip. Passives and connectors all the same, but another header on the V302 PCB.

  4. I am somewhat incredulous that this eight pin microcontroller is driving a VGA display, and a keyboard, and also able to run some software at the same time. The writeup says VGA only needs three GPIO pins. So it’s bit banging monochrome VGA?

  5. OK, but when somebody starts talking about “productivity software” on a minimal computer I immediately find myself thinking of the Cambridge Z88 https://en.wikipedia.org/wiki/Cambridge_Z88

    Fine, so the Olimex board has a monitor: that’s great, and highlights something missing from most hackers’ education (understand a micro’s monitor program, and you’re most of the way to understanding a mainframe’s front panel).

    But without something like the Z88’s wordprocessor, spreadsheet and database (terms admittedly used loosely), can this really be called “productive”?

Leave a Reply

Your email address will not be published. Required fields are marked *

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.