[Arm-netbook] rev0 a10 eoma68 board bringup

[luke.leighton]

ok, i've received the PADS file from wits-tech, and ran the "design
verification" tools and it came up with several errors.  initially i
thought "argh!" but then later on in the 17-hour marathon experimental
redesign and learning experience with PADS i made several irrelevant
or cosmetic changes, and found that errors which had *NOT* been
detected previously were coming up *AFTER* irrelevant or cosmetic
changes had been made.

so the preliminary conclusion is that there's a bug in PADS which has
caused wits-tech to believe that the PCB layout they kindly did for us
was absolutely fine - according to the design verification checks -
when in fact it wasn't.

this leaves us with a rather hard choice: carry on trying to debug the
85x55mm rev0 samples and try to bring them up so that we can provide
them to people (including some rather large clients who have been
waiting like we have and everyone else has for several months for
them), or we ask wits-tech to focus on doing the [appx] 48x75mm
version.

as part of the 17-hour marathon i shrunk the board to 47.8 x 72.75 mm
(just to see if it was possible to do) and it turns out that yes it's
possible - there may even be enough room (2 or 3 mm can be added back
in - maybe) to not have an absolute dog's dinner rats nest for the 28
pins RGB/TTL interface to be equal length tracks [the autorouter put
them eeeeeverywhere].

the caveats: in order to get down to the smaller size (and to make my
life easier when learning PADS) i had to lose the following:

* all but one TSSOP-48 NAND Flash (on BOTTOM, leaving TOP as room for
the capacitors for the A10 and NAND)
* the FPC-45 *and* the DIL2-44 are both gone.
* the RTC battery, its 32mhz XTAL and associated components (this
isn't so much of a loss because the STM32F has an RTC)
* the HP Audio is definitely gone.

our thoughts at present are to focus on getting samples ready that can
go into the upcoming KDE Tablet (and developing a pass-through card to
go with that, so that it can at least be tested).

as a second phase - after getting successful samples up and running -
we thought about doing a version of the board which is square-ish in
size which has a stack of connectors - RJ45, USB2, 5V Power etc. etc.
- which would be suitable for engineering and educational purposes.

thoughts, advice and comments appreciated.

l