VoCore, The Tiny Internet Of Things Thing

vocoreWith tiny Linux boards popping up like dandelions, it was only a matter of time before someone came out with a really tiny Linux board. This is it. It’s a tiny board less than an inch on each side with an 802.11n System on Chip running OpenWrt on Linux. The best part? You can pick one up for $20 USD.

The VoCore isn’t so much as a cut down ARM dev board as it is a cut down router capable of running OpenWrt. It’s not a power house by any means with 8MB of Flash, 32MB of SDRAM, and a 360MHz CPU, but if you ever need something that’s less than an inch square, you probably don’t need that much power.

The VoCore features interfaces for 100M Ethernet, USB host and device, UART, SPI, I2C, I2S, and 20 GPIOs for blinking LEDs and listening to sensors. There’s also a dock that breaks out the Ethernet and USB ports, available as a kit or already assembled.

It’s a pretty cool device, and with low current draw (about 200mA) and being able to accept +5V power, we can easily see this tiny board popping up in a few projects.


40 thoughts on “VoCore, The Tiny Internet Of Things Thing

    1. It’s a Linux capable board so it requires some power.
      If all one needs is to flip some leds, drive steppers, read sensors and push data onto a network then this is NOT the best product as a smaller MCU board (+ network chip if needed) would do it fine drawing a fraction of that current.

      1. You would be surprised … back when I built my ego box ( http://www.electrobob.com/ego-box/ ) and used an ENC28j60 PHY I was surprised how much current it required: typical 250mA. So yes, this WiFi module will burn less power than an Ethernet PHY while providing more things (sans ADC).

        I just wish router manufacturers would include a nice 8-10GPIO header somewhere… especially for nice models like the TL-WR710N which costs about 25 EUR and comes with box and power supply.

  1. My 3 year old 1ghz dual core 2gb ram 8gb flash android tv box’s cpu module is 2″x2″, and could be smaller if they went with a wifi soc instead. So this is really underpowered compared to what it should be.

    1. You could cram a whole hell of a lot of power into that sort of size. However, it’s not cost effective, chews up many watts of power, and just doesn’t make sense for this purpose.

    2. Wake me up when someone will be able to ditch Android then install (natively, no VM tricks) a real Linux distro on those tv dongles/boxes. Android is too dumbed down and slow for nearly everything serious.

  2. This is an awesome little board – part of the excellent trend for genuinely super-useful subsystem parts at amazing prices, but it looks like it’s going to suffer a pet peeve of mine for minimized boards like this: No screw mounting holes at all!

    It’d be okay if the dock pcb had at least one – I hope the final version does. Two is nicer, then metal standoffs can be used to properly secure and earth the module.

    Note that RasPi also forgot for their first version too…

    Just one M3 mounting hole makes half-securing reasonably easy.
    Otherwise it becomes necessary to cut down nylon threaded standoffs, and superglue them in place, and that’s much more of a pain.
    I think this is a common mistake precisely because the designer considers the board to be the final product, and may suffer a little tunnel vision.
    For us of those who implement the possibilities these parts enable, it’s sometimes frustrating to hit little hickups like this.

    It’s like buying tools from a company that obviously doesn’t use their own products, even though they *could*. Stupid simple little annoyances that could have been fixed, if the designers had been able to really step back and take an absolutely holistic vision of their work. (An approach a certain computer engineering company is well known for…)

    Of course, the best part about crowdfunding these things, is that one can certainly point out these things before it’s too late. And so I think the open-sourced crowd-funded approach has superior efficiency to the behind-closed-doors development approach. ;)

    1. Use double-sided take or make clips out of plastic. I’ve lasercut little mounting boards out of MDF or acrylic.

      Next time take your stand-offs and cut a slot in the side the thickness of the board.

      In many cases they’re meant to be treated as modules (daughterboards), so they don’t need mounting holes.

      1. Exactly! I, personally, tend to just use headers to mount boards like this one, even when they’re not actually daughterboards– cheap Radio Shack mini perfboard works great for it, has mounting holes and everything. I just slap some e-tape over the exposed metal on the underside of the female headers and call it good most of the time.

  3. If you don’t need it quite this small then you can pick up one of the various mini routers using the same Ralink SoC such as the Hame MPR-A1, Hame MPR-A2 or one of the many clones. They’re all cheap and come a built in USB host port, ethernet magnetics + socket and a built in LiPo battery. The battery will power the router itself and the USB host port. Harder to get access to the GPIOs though.

  4. The HLK-RM05 uses the same SoC and is about the same size, same peripherals (with breakout board), though half the flash and ram. at about $13. Easily flashed to OpenWRT and I managed to get all the GPIO accessible/contrllable

    1. HLK-RM04 *. There are only 10 GPIO exposed on the pins, though. HiLink also has another version that is SMD like this the HLK-RM10 at a couple $ more. (and these prices include shipping to your door) If you need all that GPIO the VoCore may actually be worth it!

  5. Think you could do a mesh network with these guys? Like an 802.11s thing? If you could mesh these guys I could use them for a project. Need the bandwidth over something like Zigbee.

    1. If it’s running OpenWRT, then you could take a look at getting the same software stack that Freifunk uses. Their distro is based on wrt and runs on cheap routers, builds a free mesh network amongst nodes (and optionally a vpn link to a central server if the mesh networks are out of range from one another)


  6. I don’t get it Hack-A-Day? First it’s not an ARM CPU, it’s MIPS. Second Carambola 2 from 8devices has been around for years, has a faster identical core, double the RAM, DDR2 RAM instead of SDR, double the flash, two ethernet ports instead of one, a similar set of peripherals, only 3 x 8 mm larger, has a equally useful breakout, and you can hit buy now today for 20 euros (vs 20 dollars). What’s so special about this one?

  7. I’ve already got 10 neat uses in mind for this thing paired with a Ras-on-module, My favorite so far is a Pocket Sever the size of my old palm trio. I bet i could build it with enough battery life for 1 Day running with a full load. Runners up include the Obligitory emulator machine with netplay ability, and a full feature micro car computer with Phone Docking. bonus idea: easy wifi for an C64.

  8. Reminds me quite a bit of the tiny SBC, the Arietta G25 http://www.acmesystems.it/arietta
    Its a tiny ARM9 (Atmel) board with 128MB-256MB of RAM, SD Card and GPIO/SPI/I2C e.t.c. Used to sell for 9.99 Euro as a promo but is now about 18.99 Euros. You can get it with a WiFi chip as well much like this setup but a little bigger. The great thing about the Arietta is that the atmel ARM9 chip used in the board has excellent mainline Linux Kernel support…much better than any other SOC that I’ve seen so far (it’s also much older). The website also has some excellent tutorials in getting up and running

  9. “You can pick one up for $20 USD.” No, you can’t. You can pay $20 and HOPE that they send you one in August… no wait… October… no wait. My point is it’s vaporware at this time. Looks real, but may not be. Hopefully it will ship, it looks nice, etc… but… If you actually want to “pick one up” the currently shipping alternatives are:
    – Carambola by 8devices 19.00€ or $38 incl. shipping.
    – hacking transcend wifi sd ($50)

    1. It’s a kickstarter. Out of all the kickstarters I’ve backed, all have had delays.

      I understand this, my day job is research/development after all. In a balance between goodness, cost and time, time is the least easy to control. I’d rather have it come late than be shonky.

      So long as they get there, and don’t run out of support for themselves, I don’t mind if it takes more than twice as long as estimated. That’s totally normal.

      You want something delivered on time? Employ a company that specialises in making that *exact* thing. Oh, wait, there doesn’t exist such a company yet? Well, then expect delays.

      Yes, this project might have benefited from at least knowing about the others, but this happens. Information is often hard to come by.

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.