Wanted: Recommendation for cheap (linux) single board computer

[micro571]

I'm going to be working on a project that requires network connectivity (via ethernet). What I'd like is a cheap, single board linux computer that has some controllable data buses (SPI, I2C, or even UART, but preferable SPI or I2C). I'd prefer if it was running linux as that is what I'm most familiar with.

I'll be designing a custom board to attach to the SPI or I2C bus that will control some leds.

As long as it has ethernet, the board can be pretty slow (like 100MHz, but I'll take whatever is cheap). It'll just be used for reading and writing data over I2C/SPI.

When googling I've found a number of boards, but they're quite over overpowered and expensive for my needs.

Any suggestions?

 

[cr0sh]

A prospect might be a "plug computer" - I've seen these as low as $100.00 in single quantities; the only problem is, I have yet to find just the bare board versions (instead of the full "plug") - however, if you need is stationary, and can run off of regular AC plug current - it might be perfect...

Your only other option might be an older PC/104 system; but getting such a thing up and running (ie, finding all the peripheral cards/boards) won't be cheap or easy.

You've probably already found the BeagleBoard...

Have you thought about maybe a Nano-ITX board and USB to an ATMega/ATTiny (and from there I2C/SPI)?

None of these options are inexpensive (of course, I am guessing here, too - what is "expensive" to you?)...

 

[Sceadwian]

Expensive and cheap don't equate to dollar values =) How much are you willing to spend. And if you answer "as little as possible" I swear I'll explode.

 

[3v0]

I would suggest an old PC motherboard. If you want get one with a built in ethernet and UART. Price should be zero.

Although I have not looked for one I expect you can get a PCI parallel IO card and bit bang either SPI or I2C. I do not know that you could use the printer port as you need both input and output bits at the same time. Possibly.

 

[arunb]

I would suggest an old PC motherboard. If you want get one with a built in ethernet and UART. Price should be zero.

a computer motherboard sounds like a very good idea, as these are cheap and easily available. The motherboard could be linked to other modules (say a relay card/ analogue card etc) via RS 485 or serial RS232

 

[Pommie]

Googleing "sbc linux" found this page.

Mike.

 

[micro571]

Thanks for the replies.

I should have been more specific, I'm looking for something fairly small, up to 10"x10" or so, so the full motherboard solution won't be an option.

How much am I looking to spend? Hopefully less than $100. Considering I can buy a Linksys WRT54GL and put OpenWRT on it for about $60, I was hoping to find a similarly priced option that would be more geared towards development.

The beagle board would be perfect if they had a cheaper option (<$149). I don't need a 600 MHz processor.

I'll investigate the plug-in option(s) and follow up on that Linux Devices link. It looks pretty interesting.

 

[Sceadwian]

Micro571, apparently you didn't do your research. The Micro ATX form factor is 5.9" X 5.9" That's considered a full motherboard solution for the Micro ATX form factor, there are multiple hundreds of products out there, they're often called shoebox PC's (for vertically mounted add on cards). For under 100 dollars, you're pushing the boundaries of sanity for a full fledged computing device. You can find smaller and better for a LOT more, but that's because they're niche markets, the economy of scale isn't there. For cheaper you have to go much lower power, so far much lower that you should re-think your criteria and design goals, to double your budget you'll increase the power by an order of magnitude or two... The cost/benefit ratio is so staggeringly in favor of avoiding dirt cheap it's not even funny.

Just think what the cost/benefit will be when you get the system you set your design goals on and find out that you want to add a feature that the system doesn't posses. You'll have saved 50-100 dollars at the gate! For a product that will fail to reach your goals, a brick, or it'll do what it's always done just fine and you'll have to go buy a whole new product to meet your new goals. You have to think ahead.