Trying to disconnect shorted GPU on my Acer Predator 17 G9-793-73A6 (N15P4)

[Proxy64]

Hi, I am trying to bring back to life my Predator without the shorted GPU (bought in 2017).
For now I don't have schematics, my mainboard is far not similar to others I found here, all mosfet are 7658A part one RA10.
Each coils have 2 or 3 mosfets, so I removed coils, still shorted on GPU side 0.2ohms.
But still not enough, It begin to start, as soon I see hdd led and hear some noise from hdd it restart.
Here some picture of the board

https://youtu.be/tMUdAhtZk_g?si=0ph58aw6QR4JJKik

[Proxy64]

The other side

[Proxy64]

Oh looking with the microscope the down side of the board I noticed something weird !

[Sephir0th]

I have doubts about your diagnose to be honest, since 0.2 Ohms are nothing to worry about in the GPU Core section. These are normal values. But it also needs to understand about what GPU related power rails we are talking here in particular. the one with the 3 Phases is most likely the Core, but what is the one with the single inductor? Here you are possibly right and it most likely isn't okay anymore to measure 0.2 Ohms there. Unfortunately the schematics and the pdf boardview doesn't reveal it for some reason. The only thing I know now is that the coil got the designation PL9700 from the designers.

And it would be a big advantage when you tell something about the (previous) symptoms of the device, whether the CapsLock test was performed, things like that. Before you started to mess with the board.

[Sephir0th]

Oh looking with the microscope the down side of the board I noticed something weird !

And yes, this pin needs to be bent back, since it won't make proper contact anymore to the internal display.

[Proxy64]

I have doubts about your diagnose to be honest, since 0.2 Ohms are nothing to worry about in the GPU Core section. These are normal values. But it also needs to understand about what GPU related power rails we are talking here in particular. the one with the 3 Phases is most likely the Core, but what is the one with the single inductor? Here you are possibly right and it most likely isn't okay anymore to measure 0.2 Ohms there. Unfortunately the schematics and the pdf boardview doesn't reveal it for some reason. The only thing I know now is that the coil got the designation PL9700 from the designers.

And it would be a big advantage when you tell something about the (previous) symptoms of the device, whether the CapsLock test was performed, things like that. Before you started to mess with the board.

Humm really interesting thank you!
So I went back to take the measurement with more precision, my owon oscillates between 0.10-0.12 ohms
This laptop is back from Acer repair service, diagnosis gpu dead, the entire mainboard needs to be replaced.
So now I realize I made a mistake, I went straight in that direction!
At first, it didn't power up, just some flashing of the battery LED.
Later, I realized that by removing the battery connector, it goes further, I got the current state ( video on Youtube).
If you still think this value of 0.11 can be considered normal, I think it’s could be better I go back.

[Proxy64]

And yes, this pin needs to be bent back, since it won't make proper contact anymore to the internal display.

I tried but this pin seems to be broken (pin37).
In this schematic from another, more recent version, but it probably hasn't changed for this connector, there's an inconsistency between the schematic and the table on the right.
According to the table on the right, this would concern the camera, which isn't my priority at the moment, but in the schematic, it would be a +5VS_LGF, which could be more problematic.

[Sephir0th]

Only the schematics to the left applies. What you see on the right is imho how it connects to the panel after the cable.

[Sephir0th]

It's just needs to be visualized what connects to where. So instead of to measuring between ACN and ACP directly you can measure between these two points.

[Proxy64]

It's just needs to be visualized what connects to where. So instead of to measuring between ACN and ACP directly you can measure between these two points.

Sorry I don’t know what you’re talking about ?

[Proxy64]

I'm back to the beginning.
With the battery connected, the battery LED only flashes and nothing else (this laptop has been inactive for 2 years, so I think it's dead).
If I disconnect it, I have the sequence on video: 2-3 times, and it turns off.
I noted the voltages on each coil.
5V is the only one present as soon as the charger is connected.

[Proxy64]

Updated with processor, measured on capacitor, I think mosfet and coil are on the other side

[Sephir0th]

Sorry I don’t know what you’re talking about ?

Yep, "slipped" into the wrong thread. Sorry....

[Proxy64]

Sephir0th or of course anyone else.

I'm really surprised it refuses to start if the battery is dead.
Especially since it's not a removable battery!
At this point, do you think a disconnected battery could have an impact on the failure to start?

What do you think about voltages measured (last picture) ?
Thank you for your time

[Sephir0th]

It can't power up with the battery being disconnected? Sounds strange

However, the voltages you shared are looking fine. So it could be something simple. I would try removing the SSD and reseat the RAM, and try with only one RAM module installed at time. Tey excluding one of the module and one of the slots being faulty this way.

[Proxy64]

Fixed !
That was a long adventure, I want to give the shortest feedback as possible, maybe that can be useful for someone.
At first, this laptop went to Acer service in Thailand, they refused it because it was bought in Europe.
Then it went in a local repair service, after 3 months they gave it back, we cannot repaired it.

Then I finally found schematics for this board enzo_mb_n17_KBL_161011.pdf (posted here), this schematics include the missing page 87 !

Like Sephir0th said at the beginning I was in the wrong way.
Reminder of the symptoms: it was powering up then restarting after less than 1s. No image.

GPU NVIDIA N17E-G2-A1 resistance to ground 0.1 ohm ( that is OK )
CPU Skylake i7-6700HQ resistance to ground 2.4 ohms ( that’s OK )
I noticed to my great surprise that without heatsink the CPU temperature easily exceeds 100° after just a few seconds of powering it on! This requires working with a heatsink. Not practical at all.
At the beginning I didn’t check the charging stage, because it passes all the 20 stages of powering up, I was looking for what was causing it to restart, such as a processor temperature alert.
To my great luck, at one moment it no longer reacted to the power button, so had to restart from the beginning, USB charging IC p88.

And finally found, in the back to back mosfet stage, the first mosfet (ACFET) mounted was incompatible with this circuit, on the photo we can clearly see small solder balls which demonstrate that it has been replaced, but remplaced by a wrong type an AOE6932 dual mosfet !!!

After replacing it with a compatible MOSFET, the laptop wouldn't boot. In fact, it was booting with the wrong MOSFET because it had the effect of bridging the source drain.
Then I discovered a problem with the charger's ADP_ID, I removed this control (Q8810), then the laptop is FIXED.

BUT last issue
After all these journeys (and others, lol), I put the motherboard back in the laptop case.
And unfortunately, I noticed that the screen also has a problem (note: the image is perfect on an external monitor).
I'm afraid the screen needs to be replaced. I was surprised by the high temperature in the bottom center of the screen. What think of that ?

the wrong mosfet
screen issue
screen issue
screen thermal picture
screen without the frame thermal picture