RE: [pandaboard] Panda dead (worked fine in EU, doesn't in US)

[Joshi, Vikas]

Do any messages show on serial terminal?

 

Vikas

 

From: panda...@googlegroups.com [mailto:panda...@googlegroups.com] On Behalf Of Jan
Sent: Saturday, July 28, 2012 7:36 PM
To: panda...@googlegroups.com
Subject: [pandaboard] Panda dead (worked fine in EU, doesn't in US)

 

Dear knowledgeable list,

 

(Email submission to the list is currently broken?)

I've recently taken my Panda on a flight to the U.S.;
In good ol' Germany the board was booting merrily, but
now there's not even a flashing LED.

My first assumption was that the power supply I brought
along isn't getting enough wattage out of the US 110V grid
(before you ask: yes, it accepts both), so I got myself an
USB OTG Y-cable as suggested elsewhere.

Still no sign of life whatsoever (and Linux doesn't even
notice that there's something on its USB bus, either).
I've checked the board surface thoroughly, no obvious signs
of damage (and it was in my cabin luggage, not thrown around
with the rest - so it just got x-rayed, heh).

What would you suggest as next steps in reviving the poor
thing? Check if there's any voltage on the JTAG pins? Is
there an easy-to-reach spot where I can check if there's
any power going into the board and where it's magically
dissipating? Other obvious things I might have forgotten?


Thanks so much,

Jan

[Lioric]

Mine did something similar, but in my board it started to boot but
hang a few moments later, it did this for 5 times or so and then it
never went back on (don't even the power led)

Now it is my soldering practicing board, along with a similar non
working BeagleBoard
(and they say "find something cheap to practice" ;) both boards total
was 500 dollars after shipping and taxes)

Lioric

[Jan C. Nordholz]

Hi,

> Do any messages show on serial terminal?

no, nothing on the serial either. I'm working on a private
x-loader branch, but it's the same with a stock image.
I mean, even with no SD card in at all it should at least
flash an LED, right?


Jan

[Pranay Sharma]

Dear Jan,

I think you need to get a good DC poweresupply( it doesnt matter if its converted from 110 or 220V).

In my case the board initially used to work fine until a month later when i ran it with newer versions etc.

then symptoms like random restarts, no bootup etc started to appear. My initial hunch was overheating, so I did all sorts of things to solve the problem(which also included putting a big table fan near it :p, and ofcourse a decent heatsink )

Finally, as the problem still didnt get solved I checked the voltage being supplied and that turned out to be the problem.

As the rating of my powersupply was less, so when more amps were eaten up by the board, the voltage used to drop.

I suggest a good 5v(3amp+) powersupply. So try that first.

Best of luck

Br

Pranay

PS: I think even a USB y-cable is not a good source. Also while plugging your powersupply take care that the positive DC supply goes to the central pin

[Jan C. Nordholz]

Hi Pranay,

> Finally, as the problem still didnt get solved I checked the voltage
> being supplied and that turned out to be the problem.
> As the rating of my powersupply was less, so when more amps were eaten up
> by the board, the voltage used to drop.

can you tell me where you measured that (voltage eaten vs. voltage
supplied or amps flowing through)?

> I suggest a good 5v(3amp+) powersupply. So try that first.

Huh. I've got no peripherals connected which need to receive power
from the Panda, I'm not doing heavy computation... I wonder how
that could consume 3A.

Anyway, maybe it's really time I get myself a multimeter here in the
US and start poking around. Hoped it wouldn't come to that...


Thanks!

Jan

[Tom Mitchell]

excuse the sloppy snips below...
It was working and is now working --
Thus I suspect mechanical or electrostatic discharge damage.
Perhaps by an inspector. I would never package with anything
connected cables and SD cards present leverage that might
bust a solder connection.

Custom boot loader... is there a chance that a setup
went wrong. I need to double check on the pandaboad but
some ARM systems the power can be boosted too high.
Increasing the clock (over clocking) should be OK. Try
a known standard card.

Power: good USP power from a newer smart phone or tablet
should get things to light up with a standard loaded SD card with nothing
connected. If you have one that has booted and is configured
in the most generic way LEDs should be good indicators. i.e. nothing
loaded no keyboard, no display, no serial, no network to over tax the
power. In hindsight one would always make a second known good "boot" SD
card.

Can you inspect the SD card? It is possible that customs inspected it
looking for something and their virus tools crushed the boot loader/ MBR.

A blanked SD card... I never tried one.... I need to check if all the LEDs
are software controlled or if there is one PowerOK LED.





On Mon, Jul 30, 2012 at 8:12 PM, Pranay Sharma <mynamei...@gmail.com> wrote:
> Dear Jan,
......

> Finally, as the problem still didnt get solved I checked the voltage being
> supplied and that turned out to be the problem.
> As the rating of my powersupply was less, so when more amps were eaten up by
> the board, the voltage used to drop.
> I suggest a good 5v(3amp+) powersupply. So try that first.
> Best of luck
> Br
> Pranay

......

>> I've recently taken my Panda on a flight to the U.S.;
>> In good ol' Germany the board was booting merrily, but
>> now there's not even a flashing LED.

--
T o m M i t c h e l l

[Jan C. Nordholz]

Hey Tom,

thanks for your ideas!

> Thus I suspect mechanical or electrostatic discharge damage.
> Perhaps by an inspector. I would never package with anything
> connected cables and SD cards present leverage that might
> bust a solder connection.

The panda hasn't been inspected - as I said, it was in my cabin
luggage, so they just x-rayed it with the rest of my bag and
that's it. No opening, no funny questions...

Besides, I had nicely re-wrapped the board in bubble wrap in
advance, placed it in its original box etc. - the whole shebang.

> Custom boot loader... is there a chance that a setup
> went wrong. I need to double check on the pandaboad but
> some ARM systems the power can be boosted too high.
> Increasing the clock (over clocking) should be OK. Try
> a known standard card.

I'm making minor changes, we're using these boards to teach our
undergrads finding their way around an embedded system, so in
order to make it interesting we add a puzzle here, an encryption
layer there... in short: nothing that should overclock or in
any way "exhaust" the board beyond recovery.

> Power: good USP power from a newer smart phone or tablet
> should get things to light up with a standard loaded SD card with nothing
> connected. If you have one that has booted and is configured
> in the most generic way LEDs should be good indicators. i.e. nothing
> loaded no keyboard, no display, no serial, no network to over tax the
> power.

That's how it's set up. In fact I've never connected anything to
it besides the serial line and then and again a network cable.

> In hindsight one would always make a second known good "boot" SD
> card.

Yeah, I've got the original dd image right here, no problem dumping
that back onto a card... didn't change a thing.

> Can you inspect the SD card? It is possible that customs inspected it
> looking for something and their virus tools crushed the boot loader/ MBR.

The SD card was with me, too - no customs inspection involved. It
mounts just fine in my laptop, data looks ok

> A blanked SD card... I never tried one.... I need to check if all the
> LEDs are software controlled or if there is one PowerOK LED.

IIRC the Power LED is already controlled by the "stage 0" ROM bootloader.

Hmm, maybe it's something completely obscure like a jammed Reset button,
so it simply never actually boots...


Best regards,

Jan