[Arm-netbook] Improving the wiki, build and test processes for allwinner kernel and ubuntu images for the Mele A1000/A2000

[Matthew Boyd]

Henrik Nordström <henrik <at> henriknordstrom.net> writes:

> 
> sön 2012-06-03 klockan 21:01 +0000 skrev Matthew Boyd:
> 
> > I did make an SD card adaptor, though I think I trimmed it just a bit too 
much 
> > and managed to have it wiggle just enough to blow out one of the ports 
because 
> > now the tablet recognizes an uSD is installed, but does not tell me 
anything is 
> > on it.
> 
> Ouch.
> 
> > I can still get the tx side off, though all I have been able to see is 
> > when it starts booting the kernel, have not gotten a peep out of u-boot or 
> > bImage (depending on which version I have loaded up at the time).
> 
> Is your u-boot modified for SD-console?
> 
> Stock u-boot do not use the SD pins for serial.
> 
> > I was 
> > figuring on trying to tap straight off of the test points on the 
motherboard by 
> > the A10 that by schematics and location on the A10 should be the serial TX 
and  
> > RX, and JTAG ports.
> 
> Yes, if you can then it's hightly recommended to use those for serial.
> It's different pins from the SD-card slot, and is normally configured
> for serial & jtag.
> 
> Identifying the serial pads should be easy. First, order from bottom of
> the chip is serial, jtag, so serial TX/RX is the bottommost two pads
> unless they have done very strange wiring of the traces. But it's hard
> to say which is which. Connect GND & VCCio, then use the RX cable (the
> pin where your USB UART receives data) to look for the TX pad (where the
> A10 transmits data). The other one is RX.
> 
> > I was figuring on attaching them to a uSD to SD adaptor 
> > that I could mount into the device without being to obtrusive, though I can 
> > only get the pads narrowed down to a 50/50 shot on each of the group of 3.
> 
> Group of 3?
> 
> Serial is 2
> 
> JTAG is 4.
> 
> The order of the pins on the A10 these pads connect to is
>   | 
>   | MS  CK  (JTAG)
>   | DO  DI
>   |  -   -  (TWI, not wired to these pads)
>   | TX  RX  (SERIAL)
>   |
> 
> So bottom two should be the serial pads. But a bit of guesswork which is
> which.
> 
> Regards
> Henrik
> 
> _______________________________________________
> arm-netbook mailing list arm-netbook <at> lists.phcomp.co.uk
> http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook
> Send large attachments to arm-netbook <at> files.phcomp.co.uk

I meant to say the 3 groups of 2.  I spent a little time looking at it edge on 
with x10 magnification and I am fairly sure I know which of the pins goes to 
the A row and which goes to the B row.  Good enough to go for it anyway.  At 
bare minimum I will have it close to the way it was when it was plugged in to 
the USB (Just to keep the connections the same if I can, but not a big deal if 
they dont, as you said, I can always adapt later).  As yet I dont even have a 
JTAG box to connect up with so those dont matter as much.  Only bad thing 
though is I will probably have to use my other unit then to solder to since I 
will need one I can boot from the SD card to keep things consistent.  I might 
practice on the test unit first, just to make sure I get things right before I 
start messing with my good unit.
As for the u-boot, I did try to build a copy (dont remember which source I 
used) about 2 weeks ago.  I copied it into the NANDA boot partition to over 
write the original one and I got no boot or uSD console (again in this case I 
blame my utter newbieness on the failure), which is why I was working on a  
plan to use the test points as my console even before you had suggested it 
hoping that these pins might be active earlier in the boot process with the 
stock setup.
By the way do you have any suggestions on additional components to add in to 
try and protect the A10 since those test points go straight into the chip.  I 
was already figuring some small ohm resistors at minimum would help protect 
(although it didn't seem to in the case of the uSD slot since those have series 
resistors, pullup resistors, and ESD arrestors and it didn't seem to protect it 
completely in the end.)
Matt