Dont know why I didnt see this earlier, I have an A3 board, found
this on another topic
from one of the designers.
- Moved DVI_PUP pin to the TPS65950. No SW impact, however,
updated SW
can be used to turn of the DVI interface. *Electrical change
from A2*.
I verified that the DVI is not getting the proper power up signal on
the initialization of the A3 so it would seem this may be the problem
with the "no user screen" hang up.
I will add the line and get you a dump later this afternoon Im
afraid, I am at work and dont have the board here to try..
Ok, I checked the TPS410 LCD / DVI controller power up pin "/PD" pin
10 it is remaining low. On the A2 board the manual states this is on
GPIO170 in the table, but the schematics show it on page 3 Pin R2
GPIO129.
The neweer A3 boards which started shipping in Oct have a change ,
it was stated on another thread.. that the DVI power up has been moved
to the TPS65950 hub controller.
This would imply a software change is need for the A3 boards to
properly initialize the DVI ... it doesnt state which pin in involved
howerver as reported the shipped SD with angstrom has it correct so
the correct initialization / driver is available.. below is the post
i cited..
"" No there are not. Here is the list of changes.
- Slightly modified PCB layout (Rev B) to correct the following
- Changed silkscreen on L12 to R159 to reflect the usage of a
resistor
instead of an inductor. Resistor was used on Rev A2. *No
electrical
difference*.
- Changed routing on R66 and R68 to make them separate paths
instead
of parallel. *No electrical difference*.
- Added 33 ohm resistor R157 in series with MMC clock line. Not
used
on board, only for expansion. *No electrical difference.*
- Added R158 to allow isolation of drain pin on TPS2141. Loaded
with
zero ohm resistor. *No electrical difference.*
- Moved DVI_PUP pin to the TPS65950. No SW impact, however,
updated SW
can be used to turn of the DVI interface. *Electrical change
from A2*.
- Added R160 and R155 as a possible future option. Not populated
on
Rev A3. *No electrical difference.*
- Changed R120 to 0603 package to align with arts purchased.
*No
electrical difference.*
- Added R156 to remove the required lifting of U18 pin 4.
Resistor is
not loaded on Revision A3. *No electrical difference.*
There are no plans to move connectors or mounting hole locations.
Will there never be any changes? I cannot promise that.
Do I plan any changeses? No I do not and I will resist any such
changes
unless there is no other option to solve an issue.
If there are changes, will the documentation be available to allow
someone
to build the previous version? Yes there will be.
Problem verified, but not for the faint of heart or if you dont have
the tools, I know this is been reported to the proper people and
understood, so I expect a new driver will be available shortly when
the compiler gods have sufficiently smiled.
I have the new A3 release and a working 10.10 ubuntu boot. The
proverbial hang up isnt … the boot script turns off console boot
during start up however the DVI is not being powered up. I fixed this
temporarily to test my hypothesis by lifting pin 4 of U5, if you dont
have a hot air micro station, and spare 7404 dual DVB parts to replace
dont try this, it can cause problems if not exact… I placed 10um of
mylar tape under the slightly lifted pin. if you lift the pin high
enough to clear the pad, the pin will have a fatiqued bend and will
probably break if bent back down… so I can use the hot air station to
replace the .025 pitch device if it breaks… less risk.
Just a note. If you do this to the -xM board it will violate the warranty, so just be warned. The correct fix is to fix the Ubuntu SW which is currently being worked.
OK, I tried it with both of those and it appeared to get stuck in the
same place as it did before.
Just to make sure I did what you wanted me to: On the sd-card I
copied the image over to the boot partition, renamed the old uImage to
uImage.old1 and renamed these to uImage then try a reboot.
I would agree with Gerald, my board has the device replaced as
shipped, I only did this to verify the cause, not as a fix, it will
void any warranty and could cause problems as the DVI device is
powered up while the OMAP processor is still initializing which is
unpredicatable at best and not good for the board in the long run, I
also have acess to a $20K SMT work station with proper static control
ect so it was done strictly to verify, In the mean time Im waiting for
the relased fix.