If you do this you probably want to include at least the 1V switch reg on each board and then provide a juicy 3.3v supply given that you don't know how many will be stacked!!Is it possible to have a base board supplying power to several XMOS processors? Or would this introduce problems?
(1) Is a nice cost/complexity saving idea but means they can't be used individually which severely limits your audience, thus a more XK1 like approach makes more sense.Which do you guys prefer:
1) A slightly larger base board with flash,programming, power,programming ,etc that acts as a master to smaller bare-minimum modules, deriving power from the master. Could possibly put an FTDI on the master so xtag isn't required.
2) Similar to the XK-1, everything can be it's own master. Everything has flash, programming, etc.
Obviously a master board can still provide XTAG (via 2232d?), usb connector, 5v to 3.3v regulation and alternate 5v power input connector.
the signals you want to pass from board to board are :
1) Xtag signals for debugging (and resets from power sequencing/monitoring)
2) 3.3v + GND
3) Channels 1 in 1 out using 4 wire mode
With (3) you hit topological issues (I have with similar designs), because you have input connectors and output connectors which must mate from top to bottom. This is OK for the XK1 because it use edge connectors for in and out. Obviously when boards are stacking (rather than side by side) you have to use some sort of surface mount top and bottom pin and socket headers. with signals (1) and (2) this isn't an issue because these use pass through and can hence use standard through hole headers with longer leads.
I would be interested to hear your thoughts on how to deal with some of those connectivity issues.
regards
Al
