Page 1 of 1

Fried HW?!

Posted: Sat Nov 10, 2012 10:39 am
by max1972
Hello! After using device with USB power supply, I decided to solder wires to use a Lab PS. Because I'm lazy, I didn't double checked the schematics... and I powered (5V) for 20-:-30sec thru P6, which is dedicated to reset. When I understood the error, I reverted to USB... Now, the Rasperry show stable RED light and nothing else... Do you think I destroied my board?
Thanks for your help!

Re: Fried HW?!

Posted: Sat Nov 10, 2012 10:58 am
by redhawk
Can you power the PI using via the GPIO?? - http://elinux.org/RPi_Low-level_periphe ... .28GPIO.29

Also check TP1/TP2 voltage, what readings are you getting??

Richard S.

Re: Fried HW?!

Posted: Sat Nov 10, 2012 11:36 am
by dknute
There are protection diodes on pin 1 of P6 but the design is somewhat lacking - well, it was probably meant to protect the board from low energy transients rather than permanent short-circuit.

What exactly happened to your RPi depends largely on the behaviour of the 3V3 regulator when it sees reverse current. Since the protection diode is connected to the 3V3 rail the best case scenario would be for the regulator to do it's duty. Then the board would be powered and the RUN input would see 3.3V + Vdrop of the diode, so it might have survived this. Now, if the regulator didn't handle the reverse current, or the D18 diode blew due to too much current... then you got 5V instead of 3V3 everywhere, and most importantly also on the RUN input of the CPU itself. This is pretty much death sentence to the RPi. Did you limit the current on you lab PSU to a small value first? If it was set to output large current (>= 2A) then I suppose it's not going to be a happy end...

The red LED is connected directly to the 3V3 rail so if it lits then the regulator still works. And that's actually bad news because it would mean something else blew and it might be the CPU... Start with measuring the voltages. There are 3 regulators: 3V3, 2V5 and 1V8, see if everthing checks out.

Re: Fried HW?!

Posted: Sat Nov 10, 2012 1:47 pm
by max1972
Hi, tks for the reply.

I measured 5, 3V3 and 1V8: they are OK.

I checked P6 voltage: below 0.5V... It seems the forward voltage of the diode... Something like D18 (lower side) is destroied... and this is my hope!

Question: can someone measure P6 voltage (GND is the round pad, square pad is the +V) and let me know?

Thanks for help!

Re: Fried HW?!

Posted: Sat Nov 10, 2012 3:15 pm
by dknute
According to schematics, there is a pull-up resistor R15 that should keep the P6 voltage at 3V3. To be sure I just measured it and it is indeed 3,265V on my device.

You could try desoldering D18 to see if it helps, it's not a critical component. If this doesn't help I would also desolder C97 - it's very unlikely to have failed but as a last ditch effort you could test it as well.

Re: Fried HW?!

Posted: Sat Nov 10, 2012 5:48 pm
by obcd
C97 likely can handle 5V. So, why should it be broken?
D18 could have a short to ground. The Pi should work without it.
It's just an esd protection diode for the Soc RUN input.
If you don't measure a proper voltage between P6 pin 1 and 2 after removal of D18, it's time to order a new Pi. (you should measure something > than 3V)

Re: Fried HW?!

Posted: Mon Nov 12, 2012 9:14 am
by max1972
Hi...
I removed D18 & C97... Voltage @ P6 is 0,4V. I think my PI has gone...
Before ordering a new PI... Do you have some tips for me? TKS

Re: Fried HW?!

Posted: Mon Nov 12, 2012 10:20 am
by dknute
Sorry to hear that. There's probably some internal damage inside the silicon structrure so there's very little you can do about that. I do have an idea to try though.

Power up your RPi and then connect RUN (using P6 pin) to 3V3 via a 100 ohm resistor. Then measure the voltage on the pin. The point is to see if the voltage will raise from the ~0.4V or not. The way I see it, the whole SoC might be fried or just the RUN input. If the latter then maybe you could still somehow "overpower" it to register as high, enabling your Pi to start. At this point there's not harm in trying.
If the voltatge doesn't rise above some 2V then could even try connecting RUN directly to 3V3. Worst case scenario you'll see some smoke :) Keep in mind that R15/C97 are likely a part of some power-on reset circuit so without it the Pi might not be very stable at start. Try a few times.

By the way, did you test D18 diode after desoldering it? Did it fail in any way? Also, did the SD card survive (those too are 3V3 only)?

Re: Fried HW?!

Posted: Mon Nov 12, 2012 3:13 pm
by max1972
Thanks for the tips... I think the RUN pin has melt down... because it drains more than 100mA (limited by PS). D18 is fine (!) and also the SD card is OK... So the only way is to ask to Santa a new PI!
Thanks for the efforts.