Momentarily 3.3V pin shorted to ground, Pi 3B+ won't boot up

[ronie1988]

I accidentally shorted Raspberry pi 3B+ , 3.3V pin 1 to ground. Now the MXL7704 regulator's pin 25 or 3.3V pin seems internally shorted to ground. Is there any fix, how to reset the power management chip....or is my Pi fried ?

[Brandon92]

If it is internal shorted to ground. The only solution is to replace that part. So, you have to solder a new one.
And the other option is to buy a new Rpi.

[davidcoton]

I accidentally shorted Raspberry pi 3B+ , 3.3V pin 1 to ground. Now the MXL7704 regulator's pin 25 or 3.3V pin seems internally shorted to ground.

This is known to kill the regulator on a Pi3B+, it should not, and the reason is under investigation.
You could ask for a warranty replacement from your supplier.

Is there any fix, how to reset the power management chip

It is said to be possible to replace the regulator -- if your SMT soldering skills are good enough. In practice, no.

is my Pi fried ?

Yes.

[jamesh]

I accidentally shorted Raspberry pi 3B+ , 3.3V pin 1 to ground. Now the MXL7704 regulator's pin 25 or 3.3V pin seems internally shorted to ground.

This is known to kill the regulator on a Pi3B+, it should not, and the reason is under investigation.
You could ask for a warranty replacement from your supplier.

You could, but that woudl be dishonest. This was, unfortuately, user error.

[drgeoff]

@jamesh

Are you, and on behalf of RPF, disputing the "it should not" part of hippy's statement "This is known to kill the regulator on a Pi3B+, it should not, .."?

[rpdom]

The "it should not" part is an opinion.

There are no warnings that shorting the 3.3V rail to ground will kill the regulator, but it is certainly not something that you should do. Shorting the 5V to 3.3V or any GPIO (equally easy) on other Pi models will fry the SoC. I don't see anyone stating that is not "user error".

[Paeryn]

@jamesh

Are you, and on behalf of RPF, disputing the "it should not" part of hippy's statement "This is known to kill the regulator on a Pi3B+, it should not, .."?

I don't see how you construed jamesh's reply as disputing whether or not the power regulator should survive a short. AFAIK the RPF don't make any guarantees that shorting power lines won't render the RPi dead so whilst in an ideal world the power regulator may be able to survive, such behaviour is not to be taken as read, more that if you are lucky it may.

Since the OP said they shorted the power lines (albeit accidently) that should void any warranty unless the RPi was sold explicitly saying that shorting power would not cause failure.

[drgeoff]

The "it should not" part is an opinion.

The opinion depends on one's interpretation of "overcurrent protection" in the sentence "Fault protection features include input undervoltage
lockout, overcurrent protection, and thermal protection" in the data sheet https://www.exar.com/ds/mxl7704.pdf.

And Table 4 on page 18 of the same document has the line
ISC | Short Circuit Current Limit | 3V3LDO = 0V | 120 | 230 | 260 mA

[rpdom]

The "it should not" part is an opinion.

The opinion depends on one's interpretation of "overcurrent protection" in the sentence "Fault protection features include input undervoltage
lockout, overcurrent protection, and thermal protection" in the data sheet https://www.exar.com/ds/mxl7704.pdf.

Reading up on the OCP, it seem the chip will try shutdown in order to prevent damage. It doesn't say that it will always succeed.

And Table 4 on page 18 of the same document has the line
ISC | Short Circuit Current Limit | 3V3LDO = 0V | 120 | 230 | 260 mA

Which is to do with the 100mA LDO 3.3V output. Nothing to do with the 3.3V supply to the GPIOs.

[hippy]

@jamesh

Are you, and on behalf of RPF, disputing the "it should not" part of hippy's statement "This is known to kill the regulator on a Pi3B+, it should not, .."?

I'm not sure where I made that statement or if I used those words exactly, but it's a reasonable paraphrasing of my view even if I did not.

My belief, from my reading of its datasheet, is that the PMIC is intended to be short-circuit proof, protected against over-current. That is, a short-circuit drawing over-current should not damage the chip, render it permanently inoperable.

https://www.mouser.co.uk/pdfdocs/MxL7704-2.pdf

Fault protection features include input undervoltage lockout, overcurrent protection, and thermal protection.

It is reasonable to expect a chip advertising itself to be short-circuit proof to be short-circuit proof.

If something intended to be short-circuit proof isn't actually short-circuit proof it is reasonable to consider that to be a product defect.

If there is a product defect it would be reasonable to claim under warranty for a repair or replacement when that defect bites.

This view would seem to be confirmed and shared by a forum moderator here - viewtopic.php?f=28&t=226423#p1389238

The 3V3 supply from the PI3B+ doesn't seem short circuit proof under all conditions.
The best solution would be to return the PI3B from where you bought it, and to claim guarantee.

The problem is being investigated, as the chip maker claims the 3V3 supply should be short circuit proof.

And in the Boot Sticky - viewtopic.php?t=58151

There have been some reports of the 3V3 supply suddenly stopping working, often after shorting the 3V3 to GND, but in a few cases also spontaneously. The issue is under investigation.

If the 3V3 supply has disappeared, then return the PI to the reseller.

Those both appear to be indicating that if the 3V3 disappears, because of a short or otherwise; make a warranty claim.

If that is not so then the Foundation need to clarify what the situation is and reconcile what appears to be contradictory views being presented on on the matter.

[davidcoton]

@jamesh

Are you, and on behalf of RPF, disputing the "it should not" part of hippy's statement "This is known to kill the regulator on a Pi3B+, it should not, .."?

Actually my statement, not hippy's.

And to be clear, I am not advocating dishonesty. I share the analysis above that this is, at least potentially, a manufacturing (even design) fault -- although that is not yet proven, there is enough suspicion about it for the cause to be "under investigation". So the warranty claim should be made with full disclosure of the short, it is then up to the retailer/manufacturer to accept or reject the claim.

[jamesh]

I am not sure exactly what is being investigated, but as a company, I presume, we would expect a chip that should be short proof to actually be so under the majority of circumstances. When we design products, we rely on datasheets for all components and design to those. But I am a software guy, not hw, so not my field.