The idea: Put a 5v 'mega' 1280 onto a 'breakout board' equipped with pins to fit, and operate in, a 'classic Arduino' uP socket.
Clearly, you wouldn't get all the extra I/O.
So why might it be interesting?
the extra code space
should be fairly cheap, the chip seems to be $10 or less
so, a simple, cheap upgrade for existing boards
and the ability to more cheaply change the large-code-space chip rather than scrapping an entire £50 Mega after a hardware screwup
and there's the possibility of offering some additional I/O (some extra analog, PWM and serial?) in a non-standard fashion for those that want or need it round the edge of the daughterboard (a bit like the Mini).
I've encountered such cpu upgrade daughterboards on other systems - would it work for Arduino and these chips? Is this a practical possibility? That's the first hurdle.
should be fairly cheap, the chip seems to be $10 or less
The problem is that by the time start with the $10 chip, design a PCB board, add pins to plug into the existing socket, and have the whole thing assembled, figure in the markups at least one level of distribution, you're looking at a pretty expensive proposition...
The 328 is a painless upgrade. I don't think the mega1280 would be (and ... how do you wire the conflicting pins like SPI? Arduino compatible? Or MEGA subset compatible? Issues, issues...)
The pins aren't expensive.
A 28-pin turned pin socket (maybe wirewrap version) does that job. My first search turned up 14-pin 0.3" wirewrap examples - two of those at about a dollar each in domestic quantity...
Through-hole and flow solder. Quick and easy assembly.
I don't have any idea about the cost/difficulty of the surface mounting of the 1280 and its tiny pins. I've been way away from assembly for plural dozens of years!
I also don't know the 'issues' involved with these chips - hence my question being asked.
If there's a handful of binary choices, it'd be cheap to design the board with bowties and solder-pads so that anyone that knew they wanted it different, could change it straightforwardly.
I'm presuming that there wouldn't need to be any support ICs on the daughterboard - is that wrong?
I'd have expected that such a daughterboard, even at current 1280 chip prices, could be sold profitably well below $30 - half the price of a mega.
Slightly surprised so many folk would read the topic and not comment. Whether "daft idea" or "I'd have one" - so thanks Westfw!
Just something to look into.. to use the mega bootloader, what pins have to be accessible, is it just the same as the real Arduino or do different pins need to be broken out? With the arrangement of the pins I could see them shifting to a dedicated UART for USB vs a shared one for USB and some other pins (given the number that the mega has). Just food for thought.
It sounds like a fun idea for a project but as a commerical proposition, I vote 'daft idea'.
As they other guy mentioned, I dont believe you'd have much in the way of margin to work with and in all honesty, why do you need both? Shield Backwards compatiblity would be out of the window. The code is already compatible...etc