Hi Hacker News!! The Daruino team here, working on making the world a smarter, more connected place....
In preparing to launch our Minumum Viable Product, we'd really like to know, which of these would you be most likely to use or to fund on kickstarter?
All of these are projects which I have found to be personally useful and I would definitely purchase and use if they were available on the market as described, but I am a "special use case" so I thought I would throw this out to the HN community and see what comes back!
Please answer either (I,II,III,or III+IV), and include any additional info that you want to! I Have all of these projects prototyped , but trying to decide what the market needs more...Thanks for checking this out!
I) A tiny, lightweight 14 pin DIP Package that costs <$8 with the following specs:
ARM Cortex M0 processor, 50MHZ
32KB Flash, 8KB SRAM
4 channel 10 bit ADC
10 GPIO lines
16&32 Bit timers
1 Uart, SPI
Pins optional for stand-alone integration
Programmable in C/++ using readily available open source tools over serial port
II) A tiny, lightweight Dip28 Package that costs <$15 with the following specs:
ARM Cortex M0 processor, 50MHZ
128KB Flash, 16KB SRAM
Pin compatible with ATmega328p / Arduino Uno, plugs into arduino board if desired
Drop in to "Arduino" circuits to give a big boost in capability
16&32 Bit timers
2x Uart, SPI, 2xI2C
Pins optional for stand alone integration
Programmable in C/++ using readily available open source tools over serial port, could probably be readily made compatble with the Arduino DUE IDE
III) A 4cmx6cm systemboard that costs <$25 with the following specs:
ARM Cortex M3 processor, 100(120)MHZ
512KB Flash, 64KB SRAM
8 channel 12 bit ADC
10 bit DAC
70 GPIO lines
16&32 Bit timers
4xUart, SPI, 3xI2C, I2S, CAN (tranciever not on board)
USB Host support (HOST connector not on board)
Ethernet MAC (PHY not onboard)
Pins optional for stand alone integration
III+IV) This unit would also optionally include a daughterboard (<$25) that provides Ethernet PHY w/ RJ45, USB Host connector, uSD slot, CAN transceiver, possibly other functions.
Programmable in C/++ using readily available open source tools over usb port
Out of the box Programmable in structured scripting language (BASIC-Like) with built in IDE via web browser, telnet, serial(USB), or composite video (or vga) / keyboard (stand alone retro computer action set capability!).
Note that III / III+IV are in the price range of the Raspberry Pi, which is -much- more capable computer, but as an application processor, not a microcontroller, it is not optimised for low power battery operation (months of operation from a few AAA's or lithium pack) and is not equipped with a similar range of analog peripherals, nor a similar quantity of GPIO.
Viva La Evolution!
I do a lot of microcontroller hobby projects using Arduino boards, ATMega chips, and STMicro ARM chips, so I'm a likely customer for your project.
I don't think I'd have a use for #1. I am not often performance-limited on the ATMega328, except for applications involving lots of high-resolution PWM, which wouldn't fit on such a device due to its limited array of IO pins.
#2 is interesting, though I wonder how long the Dip28 format will remain relevant. The LeafLabs people have tried doing the straight-line upgrade path from Arduino by designing a board with the same form factor, but your idea is even more elegant. I might use a handful of these for prototyping.
#3 sounds a lot like the Maple board. The crucial thing here is the growth path from protoboards up to custom circuits. The STM32Discovery board is cheap and capable, but its toolchain story is totally fucked (windows only proprietary uploader, undocumented protocol!), so it ends up being useless for prototyping. The Maple board is good, but it is hampered by the use of the Arduino IDE. It's really easy to outgrow the editor's constraints when you are developing for a Cortex-M3 chip, but it's not easy at all to break your project out into a gcc/makefile/upload type world. At least, I've had no luck making it work.
#3+#4 sounds like you're reinventing Arduino shields, no? Again, maple did basically the same thing and it seems to work out ok.
I don't think you need to worry about collision with Raspberry Pi, despite the price-point overlap. Not in the near term, at least, simply because of power consumption and complexity. It's a miniature computer, not a beefy microcontroller; I generally don't want an operating system getting in my way when I'm building embedded stuff.
Arduino took over because it offered a totally unbeatable development experience and didn't try to box you in. It's about as simple a process as one could possibly expect: you write code, pick a target board, push "upload", and it runs. It's all open standards under the hood, and they document everything you need to know about the board design, so there's no vendor lock-in anywhere. Mimic this strategy and you will go far. In fact I would say that your toolchain is more important than the capabilities of your board! Anyone who is capable of pushing through a dodgy toolchain to get their prototypes working is probably capable of building their own devboards, or of adapting some existing devboard to their needs. The real market opportunity comes from offering a seamless experience to beginners which still allows them a smooth upgrade path out of your environment. Beginners don't stay beginners forever, but the Arduino experience shows that ex-beginners will stick with your brand and buy your products as a convenience even once they've moved on and learned how to design their own, but only if you make it easy for them to outgrow you.