XPS 15-9530 faulty board LA-L663P

[eastcoastcomputing]

Hi there,

Sorry for the double post: I also posted in the Schematics only but couldn't find how to delete the post and I'm apparently not allowed to edit it... My apologies

I'm trying to get the schematics and BV for a Dell XPS 15 9530 (motherboard: LA-L663P)... Can anyone help?

The board powers up for a couple of seconds (the on/off light on the side of the board comes on, no display, no fans rev, no Dell LED code) then shuts down (the on/off light on the side of the board goes off), no liquid damage, no obvious damage to individual components...

Any suggestions or advice welcome.

Thank you

Update: Actually, I think I have a shorted component. This guy (circled in red) is the only one heating up on the board and gets to temps over 40° within seconds (I have a video on the thermal cam but I can't seem to be allowed to upload it). Can anyone help me identify the chip so I can source a replacement? The cap next to it has a 485 Kilo Ohms resistance.. Is it the culprit?

Thank you!

[saiiko]

Hello,

Can you post a full board photo?

Also if you thermal monitor when you connect the battery, this power splitter maybe dumping the 20v to ground as a protection mechanism. Are the CPU / GPU mosfets shorted? What Ω do you get from CPU / GPU coils to ground?

I have the same board, same fault

[saiiko]

Check this area specifically - may be a heat issue (4060/4070/4080) middle Atmel mosfet likely overheat.

Check resistance from 3 GPU coils to ground - do you see 3 OHMS?

[eastcoastcomputing]

Hi Saiko,

Thank you for your help.

The GPU coils aren't shorted (nor any other coils for that matter) and I didn't find any issues with the mosfets either. The only component heating up on the thermal cam is the one I circled on the picture. Unfortunately the customers decided against repairing the board and bought a new computer instead so I can't send you a full board picture anymore.

[Sephir0th]

The picture from #1 looks like a small power supply to me. Don't do the mistake to blindly replace it, it could be that it simply tried to pull current into a short circuit. In this case to replace it would change nothing.

[eastcoastcomputing]

The picture from #1 looks like a small power supply to me. Don't do the mistake to blindly replace it, it could be that it simply tried to pull current into a short circuit. In this case to replace it would change nothing.

Thank you for the advice. This is one eventuality I was contemplating...

[saiiko]

Just for future diagnostics, this is the LDO "linear voltage regulator" which is supposed to be warmer than surrounding chips.

This means, the 20v USB-C inputs and PD controllers are working (negotiating 20v instead of the default 5v USB) along with the entry MOSFETs carry the power to this chip.
From here, it creates 3v which provides constant power to the EC chip (other side of board).

In these situations, where board is taking the usual 20mA standby current (EC Chip being on/waiting for power button keypress) then the following may have happened:

- Faulty EC firmware (happened due to one of numerous windows forced bios updates)
- Shorted component (due to overheat, means EC detects over-current and enables protection mode to limit more power, thus saving further damage)

Replacing the LDO would be like replacing the nozzle on a faulty tap - the problem is always upstream and not due to the surface

To diagnose:
- Fully remove board and waterproofing plastic covers
- When first powering the board - 12v, 5v / 3v and 1v power rails should momentarily appear (1-2s) before protection mode disables them, this would start to pinpoint which rail is down (i.e. 1v CPU/GPU circuits, 12v charging circuit etc).
- Connect USB-C, and then plug battery: you should see power spikes and if your thermal camera has a fast enough refresh rate, thermal spikes where the short is
- Perform visual inspection to look cracked caps / burnt or blown
- A dead short is the same Ω as connecting your leads together i.e 0.3-0.5Ω. Anything more 1-3Ω GPU, 2-5Ω CPU for example
- Targeted voltage injection (1v max incase of CPU/GPU dual-MOSFET failure (they're the hottest components in intense use)

- Be mindful if troubleshooting BIOS, these newer machines first load EC which daisychains BIOS later after a successful read. You may think the BIOS isn't receiving 1.8 / 3v to power it, but this is normal.

TLDR; I think since On light appears, LDO and EC are good, likely a shorted component

Hope this can help the community, apologies if very basic but may help/encourage new members to ask and help us see and identify common failure modes

[eastcoastcomputing]

Thank you very much, this is very helpful.