FemtoBeacon Is A Tiny ESP32 Coin-Shaped Dev Board

Our single board microcontroller platforms have become smaller over the years, from the relatively large classic Arduino and Beagleboard form factors of a decade ago to the postage stamp sized Feather and ESP boards of today. But just how small can they go? With current components, [Femtoduino] think they’ve cracked it, delivering an ESP32-based board with WiFi and Bluetooth, and an LDO regulator for 5 V operation in a circular footprint that’s only 9 mm in diameter.

There are some compromises from such a paucity of real-estate, of which perhaps the most obvious is a lack of space to make I/O lines available. It has SPI, a UART, and a couple of I/O lines, and aside from an onboard RGB LED that’s it. But SPI is versatile well beyond its number of lines, and even with so little there is much that can be done. Another potential compromise comes from the antenna, a Molex surface-mount component, which is an inevitable consequence of a 9 mm circular board.

There has to come a point at which a microcontroller platform becomes so small as to be unusable, but it’s clear that there is a little further for this envelope to be pushed. We’d love to see what other designers do in response to this board.

21 thoughts on “FemtoBeacon Is A Tiny ESP32 Coin-Shaped Dev Board

  1. For designer of this project or anybody who wants to immplement RF in their boards, I recommend to read the application note from Cypress “Antenna Design and RF Layout Guidelines”. It’s just an introduction, but It will teach you how important is a good ground plane and the feed lane.
    Just thowing a chip antenna on a design isn’t enough.

    1. Yes, as an application engineer of a wireless semiconductor company I see terrible RF designs almost every day.
      This particular antenna will not perform well on such a small board, so I expect wireless range of this board to be pretty poor.

  2. While designing something as small as possible is a fun exercise and a nice goal in itself, I’m strongly against the current trend of having development boards and tools like the Black Magic Probe shrunk down to the smallest size possible. Having 0.05″ headers with its flimsy cable connectors are a pain in the butt to use.

    I’d rather have a 50×50 mm board with usable connectors and all pins broken out and easy to connect/solder wires to than a 15×20 mm pcb that’s awkward and horrible to use.

    Remember that this is for *development* stuff – meant to be plugged in and disconnected repeatedly and probably used for many different project.

    Of course one could argue that a size constrained one-off project also needs development and then all those boards made for ants makes sense….. but…..

      1. +1 …. I deliberately made my weather station PCB:


        ….. bigger than necessary and used nice chunky 1206 components for most of the circuitry. Leaves plenty of room for adding extra circuits in the future. And since the price is only $2 for 5 (100 mm x 100 mm) boards from JLCPCB China, might as well make it max size for that price. By the way, boards arrived in 10 days, start to finish, thanks to their new budget air post service :)

        Ironically, I still had to use a super small watchdog circuit supervisor chip, but if my coding was better this might not be necessary.

    1. That’s the thing about open source design. – don’t like it, fork it.
      There are no one (other than yourself) stops you from making a breakout board with the proper spacing etc. There might even be a market for one.

      IMHO breakout boards are boring, actual projects are far more interesting.

      1. Well, duh… But why should I fork it when I just as easy can make one my own from scratch and then also have the satisfaction of making something by myself.

        Saying “don’t like it fork it” or “found a bug – fix it yourself” is all true, but t’s also a kind of silly wankery especially when it comes to stuff being sold. Most people that are using the code and/or piece of hardware don’t want to /are not capable of making those changes themselves. If I bought an open source carburetor for my car and it performed badly I wouldn’t have a clue of how to fix the problem. And if the firmware of the black magic probe had bad firmware I really wouldn’t like to learn the nitty-gritty details of the remote GDB protocol to fix it, I just want to be able to debug my microcontroller.

        But if a major portion of the “market” rather likes a bigger and more easy to use/handle version of the tools than the ant-sized for development wouldn’t it be. good idea if the creators of the devtools were made aware of that fact? It’s not like each creator makes two versions for the same device and analyzes the sale trends for them….

        Interesting or not is another issue. But if I give you a new exciting chip on a devboard – what would you rather have to play with? A20x20 mm pcb with half the pins broken out to 0.05″ headers or a 50×50 pcb with all the pins availabe on 0.1″ and jumpers to break/measure vcc and stuff properly?

        And what do you think about the Black Magic Probe (a lovely device otherwise) that is a JTAG probe meant to be repeatedly plugged/unplugged to your projects and it’s using a fsking flimsy SMD-soldered 0.05″ header. Every time I use it I expect my $100 (shipped) device to become a piece of dead FR4. Of course I could just fork it too and make my own, but why? If the guys that made it spent 50 cents extra for some more pcb real estate then it had been so much easier to use.

        On the other hand I don’t recall [Femto] ever saying that this particular board is a devboard – that label seems to have been given by HAD here.

          1. But if you make some holes in edges you could build a stack of these with extra up and do something like put in fuselage of model rockets with all kinds of sensors and maybe a tiny camera. Intriguing.

    2. There are plenty of larger boards, with 0.10 connectors. Very inexpensive as well. When you need something very small, that’s what this board is for.

      The other day I was thinking about putting an ESP32 inside light switches. A WROOM32 would not fit well. This would. I build and fly fly model aircraft, planes and quads (drones). A larger board is fine in an outdoor plane with a meter-long wingspan, for an indoor quad that has a 30mm frame, the big board won’t work.

      It’s all about the right tool for the job. Different tasks take different tools.

  3. Gosh, if only somebody made an ESP32 board that had standard 0.1″ headers. It’s totally reasonable to fuss at this engineering marvel because there’s no more-friendly alternative for people who want a different configuration. Ghastly that this is literally the only ESP32 board anyone has ever created.

  4. For all of y’all getting your panties in a bunch because it isn’t what you would do, you REALLY need to take a close look at your mindset and consider getting over yourself for your own sake. Please. This is a bespoke design that was intended to fit the particular requirements of a specific project, which he** then generously shared with the world for those who might have similar needs. While it undoubtedly could be improved (what can’t be?), it does what is required by him for his project, as can clearly be seen in the video.

    For all of y’all just celebrating his victory or at least willing to (gently) offer advice, bravo; y’all have healthier mindsets than average for internet comment sections.

    @Femtoduino, I salute you. o7

    **gender assumption based solely on the 10:1 ratio of male to female electronic/electrical engineers and the assumed similar ratio among those who are hobbyists. If femtoduino is a Jeri or Limor rather than a Ben or Steve, then that is something to be celebrated.

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.