Zapp98
Posts: 4
Joined: Mon Aug 13, 2018 1:49 pm

How to make a warrany claim?

Mon Aug 13, 2018 1:52 pm

Hi there.

I was wondering if you guys could assist me with my Pi. When i received the pi, it seems to have been functioning well for about a week. During this time i had shut it down from the software a few times, and a couple days ago, the pi was just in the middle of running and it froze up, and so i had to hard reboot by cutting power. When i tried to turn it back on, the device would not load, did not show any activity on the sd led, and power light is solid red.

I have already tried reinstalling the OS but unfortunately this did not help. I am sure the sd cards are still functional but it seems the device will no longer boot. I have contacted the company which i purchased from on ebay but they are unable to point me in a direction to have this replaced. Could you guys please assist?

User avatar
RaTTuS
Posts: 10108
Joined: Tue Nov 29, 2011 11:12 am
Location: North West UK

Re: How to make a warrany claim?

Mon Aug 13, 2018 2:11 pm

who you bought it from is who supplies the warranty
How To ask Questions :- http://www.catb.org/esr/faqs/smart-questions.html
WARNING - some parts of this post may be erroneous YMMV

1QC43qbL5FySu2Pi51vGqKqxy3UiJgukSX
Covfefe

hippy
Posts: 3913
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: How to make a warrany claim?

Mon Aug 13, 2018 2:35 pm

It depends on where you are, what jurisdiction you are in, what the laws, rules and regulations say, but it is usually the retailer or seller who is responsible for resolving disputes on product which may appear to be faulty; repairing, replacing or refunding as may be required.

That is because your contract is with the retailer or seller not anyone else.

What rights you have will likely vary depending on when and where you bought an item, how long ago. In some cases you may be entitled to a refund even if an item is not faulty. If you cannot resolve the issue with the retailer or seller you may be able to make a charge-back claim with your credit card provider if you paid by credit card. As you bought through eBay they may also have some dispute resolution or refund provisions.

Time to get reading up on Consumer Rights for wherever you are. Perhaps let us know where that is if you want more specific advice.

Out of interest: What model Pi is it which you have experienced your issues with ?

Zapp98
Posts: 4
Joined: Mon Aug 13, 2018 1:49 pm

Re: How to make a warrany claim?

Mon Aug 13, 2018 3:38 pm

It's a new Pi3 b+ Model.

Was running MotioneyeOS. Was working flawlessly for a while then stopped. I had also installed a fan to keep it cool.

Fun fact, the fan still spins up when pi is supplied power, must be just as its connected through constant 5v pin.

The device froze during use and i was required to hard reboot it, when powered back on, nothing... Device will not show any activity on Green LED. tried numerous SD cards to see if it was the problem... Is there some sort of button combo i can press to do a system check?

Thanks guys

EDIT:

Tried the "Secret method" to check if my Pi was Dead... When plugged in, it did not "Ding" on windows
If you have a PI model A, A+ or Zero there is a special trick which can be used to check if such a PI is "dead":
Take your PI, with nothing in any slot or socket (yes, no SD-card is needed or wanted to do this test!). Take a (special) USB-A to USB-A cable (or a more normal micro-USB to USB-A for the zero) & connect it to your PC, plugging the other end of the cable into the Pi's USB port. If the PI is alive, your Windows PC will go ding for the presence of new hardware & you should see "BCM2708 Boot" in Device Manager. Or on linux, with a "ID 0a5c:2763 Broadcom Corp" message from dmesg. If you see that, so far so good, you know the PI is not dead.

DirkS
Posts: 9248
Joined: Tue Jun 19, 2012 9:46 pm
Location: Essex, UK

Re: How to make a warrany claim?

Mon Aug 13, 2018 3:57 pm

Zapp98 wrote:
Mon Aug 13, 2018 3:38 pm
It's a new Pi3 b+ Model.
Tried the "Secret method" to check if my Pi was Dead... When plugged in, it did not "Ding" on windows
If you have a PI model A, A+ or Zero there is a special trick which can be used to check if such a PI is "dead":
You have a 3B+ so you cannot use that method.

Zapp98
Posts: 4
Joined: Mon Aug 13, 2018 1:49 pm

Re: How to make a warrany claim?

Mon Aug 13, 2018 4:01 pm

Regardless, it still is non functional. I am not sure what could be wrong with it but it seems to be beyond my scope of repair

hippy
Posts: 3913
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: How to make a warrany claim?

Mon Aug 13, 2018 4:02 pm

Zapp98 wrote:
Mon Aug 13, 2018 3:38 pm
It's a new Pi3 b+ Model.
No prizes for guessing what I was guessing.
Zapp98 wrote:
Mon Aug 13, 2018 3:38 pm
Tried the "Secret method" to check if my Pi was Dead... When plugged in, it did not "Ding" on windows
That won't work with a Pi 3B+ but if you have the skills and tools you can check to see if 3V3 is present on the GPIO power pins. If not, it seems it is another dead one.

Zapp98
Posts: 4
Joined: Mon Aug 13, 2018 1:49 pm

Re: How to make a warrany claim?

Tue Aug 14, 2018 2:43 am

hippy wrote:
Mon Aug 13, 2018 4:02 pm
That won't work with a Pi 3B+ but if you have the skills and tools you can check to see if 3V3 is present on the GPIO power pins. If not, it seems it is another dead one.
No voltage from 3v3 Pin, 5V pin is showing constant 5v but nothing from 3v3. Guess its dead :(

User avatar
RaTTuS
Posts: 10108
Joined: Tue Nov 29, 2011 11:12 am
Location: North West UK

Re: How to make a warrany claim?

Tue Aug 14, 2018 7:54 am

your warranty is still with the vendor
though adding a fan may have killed it in other ways

IMO it's possible the new power chip may have some issues in a batch - though warranty as before

any date marking on the board may be useful to the general issue


i'm not sure if you can supply 3.3v on that pin and make it work - I'm pretty certain that the SOC needs 1.8V and that may be also dead
How To ask Questions :- http://www.catb.org/esr/faqs/smart-questions.html
WARNING - some parts of this post may be erroneous YMMV

1QC43qbL5FySu2Pi51vGqKqxy3UiJgukSX
Covfefe

klricks
Posts: 5899
Joined: Sat Jan 12, 2013 3:01 am
Location: Grants Pass, OR, USA
Contact: Website

Re: How to make a warrany claim?

Tue Aug 14, 2018 12:49 pm

RaTTuS wrote:
Tue Aug 14, 2018 7:54 am
your warranty is still with the vendor
though adding a fan may have killed it in other ways

IMO it's possible the new power chip may have some issues in a batch - though warranty as before

any date marking on the board may be useful to the general issue


i'm not sure if you can supply 3.3v on that pin and make it work - I'm pretty certain that the SOC needs 1.8V and that may be also dead
I killed a 3B+ by accidentally shorting 3V3 pin 1 to the metal chip cover with RPi logo (GND). The RPi instantly went off.
The 3V3 reads 0V and also reads as as shorted to GND using Ohm meter. The PMIC is totally shut down so no 1V8 output either. Don't know if fault is internal or external to the PMIC?
Unless specified otherwise my response is based on the latest and fully updated Raspbian Stretch w/ Desktop OS.

hippy
Posts: 3913
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: How to make a warrany claim?

Tue Aug 14, 2018 1:53 pm

klricks wrote:
Tue Aug 14, 2018 12:49 pm
I killed a 3B+ by accidentally shorting 3V3 pin 1 to the metal chip cover with RPi logo (GND). The RPi instantly went off.
The 3V3 reads 0V and also reads as as shorted to GND using Ohm meter. The PMIC is totally shut down so no 1V8 output either. Don't know if fault is internal or external to the PMIC?
The XR77004 marked on the circuit diagram seems to be an MxL7704 in reality - https://www.raspberrypi.org/blog/pi-power-supply-chip

The failure mode of that chip does seem hard to determine. It is understandable that a short or adverse event could shut down the chip, but odd that it persists after power cycling.

Given it has an I2C interface, seems to be programmable, and presumably contains EEPROM or other non-volatile memory to hold what the power sequencing and generator settings should be before the controlling device is powered-up, it is possible that some 'don't ever power-up' or 'never allow PGOOD to be set' state could arise which would naturally be hard to clear when what could clear that needs to power-up first.

It seems odd that the PMIC would put itself into a 'bricked' state following a fault condition but perhaps it is expected to be used in a manner where the controller is able to clear that state once it has arisen.

If that is the case, or something like it, it would explain its apparent sensitivity to 3V3 shorts, and perhaps also explain random failures. If any voltage it is generating goes out of spec for some reason it may be reacting and 'bricking itself'. But that's only guessing.

There are four main possibilities I can see. The PMIC is being destroyed or rendered physically unable to function. Its internal configuration data is being set so it doesn't power things up. Internal configuration data is being corrupted or wiped. Or something external has been damaged which prevents the PMIC reaching a 'power good' state.

It could be enlightening if someone with a failed 3V3 on a Pi 3B+ could read the I2C registers of the PMIC. Comparing not-failed and failed readings could prove useful. Possibly not an easy task though without the skills and appropriate tools to achieve that.

jbudd
Posts: 716
Joined: Mon Dec 16, 2013 10:23 am

Re: How to make a warrany claim?

Tue Aug 14, 2018 10:16 pm

I too have a dead Pi3B+ and would be happy to donate it's body to anyone in UK who cares to investigate the cause.


My testing is pretty thorough ;)
Insert the SD card from another, working Pi.
Does the red LED come on? Yes
Does the green LED come on? No (Actually the green LED ceased to work a couple of weeks before the Pi itself died)
Does it connect to my wifi? No
Do any of the chips get warm? Only the MXL7704-R3 (power supply?) chip appears warm to the touch.

User avatar
Imperf3kt
Posts: 1380
Joined: Tue Jun 20, 2017 12:16 am
Location: Australia

Re: How to make a warrany claim?

Wed Aug 15, 2018 12:52 am

The cause is well known and stated; shorting 3v3 to GND causes a latchup inside the SoC, instantly and permanantly damaging your Pi beyond repair.
Stop plugging your fan directly into the GPIO 5v
https://www.electronics-tutorials.ws/power/transient-suppression.html

jbudd
Posts: 716
Joined: Mon Dec 16, 2013 10:23 am

Re: How to make a warrany claim?

Wed Aug 15, 2018 2:16 am

The cause is well known and stated; shorting 3v3 to GND causes a latchup inside the SoC
How do you know that the damage is inside the SoC rather than in the MXL7704-R3? How many dead Pis have been inspected? If the MXL7704-R3 is replaced (expert soldering job for sure) does the Pi stay dead?

Some people report that their Pi died during reboot after an update.
Do you think it's possible that they gripped the device with a sweaty finger touching the RPi logo and GPIO1 to unplug the power lead? Or did they forget about poking around with a screwdriver?

klricks
Posts: 5899
Joined: Sat Jan 12, 2013 3:01 am
Location: Grants Pass, OR, USA
Contact: Website

Re: How to make a warrany claim?

Wed Aug 15, 2018 3:33 am

Imperf3kt wrote:
Wed Aug 15, 2018 12:52 am
The cause is well known and stated; shorting 3v3 to GND causes a latchup inside the SoC, instantly and permanantly damaging your Pi beyond repair.
I was thinking the SoC latchup was due to applying 4V or more to a GPIO pin. The usual symptom is the SoC quickly overheats and then eventual catastrophic failure. Not necessarily instant.
Unless specified otherwise my response is based on the latest and fully updated Raspbian Stretch w/ Desktop OS.

User avatar
Imperf3kt
Posts: 1380
Joined: Tue Jun 20, 2017 12:16 am
Location: Australia

Re: How to make a warrany claim?

Wed Aug 15, 2018 4:21 am

jbudd wrote:
Wed Aug 15, 2018 2:16 am
The cause is well known and stated; shorting 3v3 to GND causes a latchup inside the SoC
How do you know that the damage is inside the SoC rather than in the MXL7704-R3? How many dead Pis have been inspected? If the MXL7704-R3 is replaced (expert soldering job for sure) does the Pi stay dead?

Some people report that their Pi died during reboot after an update.
Do you think it's possible that they gripped the device with a sweaty finger touching the RPi logo and GPIO1 to unplug the power lead? Or did they forget about poking around with a screwdriver?
A quick google suggests: yes, it stays dead
viewtopic.php?t=68479

Admittedly, a much older model Pi
Stop plugging your fan directly into the GPIO 5v
https://www.electronics-tutorials.ws/power/transient-suppression.html

hippy
Posts: 3913
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: How to make a warrany claim?

Wed Aug 15, 2018 10:15 am

jbudd wrote:
Wed Aug 15, 2018 2:16 am
The cause is well known and stated; shorting 3v3 to GND causes a latchup inside the SoC
How do you know that the damage is inside the SoC rather than in the MXL7704-R3?
It seems unlikely to me, because it is hard to see how damage in the SoC would causes the MxL7704 to not produce 3V3.

It is possible that 3V3 becoming 0V while other voltages to the SoC are still applied could damage the SoC, however that does not seem to have been the case with earlier models of Pi. The electrical designs are however different so what applies to one version of a board may not apply to another.

As acknowledged, cascade damage may now be causing the MxL7704 to fail to reach a 'power good' state and shut down. But that's just one possibility. We are a long way off determining what the actual failure mode is.
Imperf3kt wrote:
Wed Aug 15, 2018 4:21 am
A quick google suggests: yes, it stays dead
viewtopic.php?t=68479
That seems to be an issue with using a greater than 5V supply. Applying higher voltages than the Pi is rated for can cause latch-ups in the SoC and cause failure but that is a different failure mode.

I am at a loss though to know how shorting the external supply 7V2 to 0V caused damage to the Pi as described in that link. Damage to the battery I would expect, not damage to the Pi nor cause the 3V3 regulator to overheat then fail. I think that can be set aside as a different cascade failure to what we are seeing with Pi 3B+ boards.
jbudd wrote:
Wed Aug 15, 2018 2:16 am
If the MXL7704-R3 is replaced (expert soldering job for sure) does the Pi stay dead?
I think there is every likelihood a Pi would be restored to normal operation with the caveat that any replacement MxL7704 must be programmed before installation. There can be no definitive answer though until someone tries that, or removes the on-board MxL7704 and supplies the required voltages from an external source with correct sequencing.

A possible alternative would be to remove the SoC and everything else on the board which uses the voltages the MxL7704 generates. Without anything external preventing the MxL7704 reaching a 'power good' state it should do so. If it doesn't it would suggest the MxL7704 itself is not doing what is expected.

Without full circuit diagrams, and not having silk-screened component identifiers, figuring out what to remove and what not to could prove challenging.

It may be easier though to just scope all the voltages generated to see if the MxL7704 attempts to bring up power in any manner or form. That's worth a try anyway before doing something more drastic.

An alternative approach would be to take an MxL7704 Evaluation Board or replicate the MxL7704 circuit, program it as it would be for a Pi, short 3V3 to 0V and see what happens.

If there is an issue I am sure it will eventually be identified and hopefully resolved. The good news is that it's in both RPT's and MaxLinear's interests to figure out what is going on. I wouldn't be surprised if they are already investigating.

hippy
Posts: 3913
Joined: Fri Sep 09, 2011 10:34 pm
Location: UK

Re: How to make a warrany claim?

Thu Aug 16, 2018 3:07 pm

jbudd wrote:
Wed Aug 15, 2018 2:16 am
How do you know that the damage is inside the SoC rather than in the MXL7704-R3? How many dead Pis have been inspected? If the MXL7704-R3 is replaced (expert soldering job for sure) does the Pi stay dead?
One post here from a Pi 3B+ owner who reports shorting 3V3 to 5V, replacing the MxL7704, and getting their Pi working again ...

https://www.raspberrypi.org/forums/view ... 4#p1319930

That's just one report but it is encouraging because it suggests whatever affect a 3V3 short has, to 5V and possibly 0V, it may be limited to the MxL7704, may be rectifiable, may not always be terminal.

Return to “Troubleshooting”