820-01598 with short on PP3V3_G3H

[2informaticos]

Still 5 ohm to ground there?
Is that ohm scale, or diode mode reading?

Not being pure short, is not easy to heat something; the injected current is low.
You may go until 4V, no more.
A thermal camera can help you; as it can detect small temp deviations.

 

[sekidata]

I went up to 4V injection and found U8200 warm. Removed U8200 and now the short to ground is 26 Ohms using resistance mode and 0.024 using diode mode.

 

[2informaticos]

Still bad; something more damaged there.
Diode mode reading should be approx 300-350...

 

[sekidata]

With applied 4V I can see several components getting hot, but I assume those are simply components in the network warming up from 7A flowing through PP3V3_G3H, right?

 

[2informaticos]

7A from 4V into 26 ohm, really?

 

[sekidata]

Yes, really. I know it breaks Ohm's law... maybe it's a sign to give up on this board.

 

[2informaticos]

This is only possible if some chip enters into thermal avalanche; but should be easy to detect it then.
If you mean that just set amp limit to 7A, doesn't mean the board sinks that current.
Set always amp limit to max; board will take the corresponding current to injected voltage and ohm value to ground.

 

[sekidata]

My current is set to max (10A). The power supply says it is putting out 6-7 A at 4 V. I measure a negative voltage on my multimeter during current injection... does that also indicate thermal avalanche?

 

[2informaticos]

That means 24-27W, something should go really hot.
At least a thermal cam must show something very clear.
You should get 4V indicated by the PSU.
If not,you are doing something wrong.

 

[sekidata]

Thanks for that comment, I indeed had a wiring issue with my PSU. Fixed now. I now do get 3.7 V when I apply 4 V from the PSU, with a current draw of 0.8 A. Resuming troubleshooting...

 

[2informaticos]

That is really low, approx 3W.
A thermal camera is required, and not sure if really helps.
Don't forget to remove CPU heatsink first.

 

[sekidata]

I found a 1cm x 1cm area of the board underside that slowly but consistently gets warm after several minutes. Unfortunately I think it corresponds to the footprint of U7800 (PMIC) on the board topside, under the heat shield.

 

[2informaticos]

Need to remove the shield and confirm it.

 

[sekidata]

I removed the shield and confirmed that U7800 is getting hot. This PMU is not commercially available as a spare part, correct?

 

[2informaticos]

You need scrap board to get it...