C2: no serial output at power up

GoogleGroups
1

I have (also) a Beagleboard C2.

Long ago I downgraded to some very old version of uBoot which supported console on USB other than RS232 (I did it because I mounted a RS232 GPS unit and didn’t want the GPS NMEA strings to stop uBoot from booting; the GPS unit shared the same 5V power source).

So it always booted without outputting anything on the RS232, silently loading its “uImage” (and rest of Linux Angstrom) from that old 1Gb SD-card… at least until yesterday, when I removed the GPS and tried to upgrade uBoot (as of www.code.google.com/p/beagleboard/wiki/BeagleboardRevC3Validation) but without success.

Now that C2 powers up (all LEDs come on) and nothing else happens (not even a byte on the RS232, as before). I’m sure something changed (because it doesn’t anymore boot Linux from that verified SD-card).

I double-checked adapters and cables, SD-card files (FAT16, exact order, with/without User button, etc) and so on, without luck. I don’t remember doing anything that could have damaged the board.

Any hints?

Thank you.

2

Hi Alfonso,

Try this procedure (for BB C2/C3):

http://code.google.com/p/beagleboard/wiki/BeagleboardRevCValidationv2

Cheers,
Max.

3 
Try this procedure (for BB C2/C3):
[http://code.google.com/p/beagleboard/wiki/BeagleboardRevCValidationv2](http://code.google.com/p/beagleboard/wiki/BeagleboardRevCValidationv2)

I got at least one result: I see the full-screen Beagleboard logo on DVI output, but nothing more than that.

I still see no data on serial port (not even a “40T”, not even a spurious byte). I tried two different computers, two different usb-to-rs232 cables, two different homemade and verified “9pin to IDC adapters” (and also tried inverting TX/RX pins), but I still get no serial output.

I even tried using a boot.scr containing setenv stdin serial (and stdout stderr as well and a saveenv following them) but no results followed. The partition is a small (20Mb) FAT16-formatted one. Seeing the Beagleboard logo should mean that the 205kb uBoot file was loaded and executed, so I don’t think that the issue is the SD-card.

Suggestions? Did someone ever tried booting from an USB pen-drive?

4 
Try this procedure (for BB C2/C3):
[http://code.google.com/p/beagleboard/wiki/BeagleboardRevCValidationv2](http://code.google.com/p/beagleboard/wiki/BeagleboardRevCValidationv2)

Finally I was able to recover my C2 from “no serial output at power up”.

I had these problems:

a) the board did not show any messages after “RAM: 256Mb” – in fact, the board had “usbtty” instead of “serial” saved (!) in the environment (I did it to use the RS232 for GPS input; boot messages disturbed the GPS, so I just “saveenv’ed” with stdin/stdout/stderr setenv commands set to usbtty).
I had to wipe the environment without entering the environment (ouch! see below).

b) then it just stopped doing serial output at boot.
I had both SD-card and serial cable problems (see below).

Someone should update the wiki adding some troubleshooting hints:

  1. “Don’t always rely on pin labels”: I built my RS232-to-Beagleboard interface using a Sparkfun breakout board (Sparkfun product 8552). Well, the pin names printed on it assume you are using a male DE-9 adapter; if you use a female DE-9, the pin list is inverted (5-4-3-2-1 instead of 1-2-3-4-5: thus Beagleboard RS232 pin 3 goes to “DTR”-labelled pin, pin 2 to “TX” and pin 5 to “CD”).

  2. I have a microSD card with a microSD-to-SD adapter. Funny: it is readable/writable everywhere but the Beagleboard. “If you can read on your multimedia reader, it does not mean that it will be read by the Beagleboard”. Ouch. The same adapter, when using the microSD coming from Beagleboard xM, just works.

  3. I was not anymore able to enter uBoot console because its environment was saved with “usbtty” instead of “serial”. When I figured out a way to enter Linux using the User button (as suggested in the link above), I wiped out the environment using “nandtest /dev/mtd2”.

Note: I tried to manually edit the environment while in Angstrom Linux prompt, but:

  • nanddump /dev/mtd2 gave me a 135000+ bytes file (0x21000)
  • nandwrit’ing it did not write anything
  • nandtest without the “-k” option erased it (“nanddump” shows garbage)
    After erasing the environment, I rebooted and MLO/uBoot went back to default, letting me to edit again.