Acer Predator Helios 300 - FH52M LA-J891P - MISSING +1.35VSDGPU!

[reessi]

Hello:

The GPU is recognised in windows but with an error, up on testing I found +1.35VSDGPU is missing for the Video Ram.

Coils (PLW1 & PLW2) read 78ohms on the video ram side when I lifted one side to isolate the video ram. I believe this is normal, let me know please?

Up on powering up no attempt is seen of any voltage on the coils.

The power supply (RT8816B) is receiving 2.6v on the enable pin.

There is 19v on the switching Mosfets (PQW1 & PQW2) via PLW3.

Power good output from RT8816GB remains at 0v which is expected since it does not turn on.

I swapped another RT8816B from another board that has a shorted PCH but no change. I cannot be 100% sure that IC works though.

What else can I check and what reasons may there be for this power supply to not come up.

There are no shorts on any of the pins on the power supply or the mosfets.

Can a problem with the video ram cause this, it doesn't seam shorted at 78 ohms.

I also powered up with one side of the coils lifted from the video ram side and no change.

Can I safely inject 1.35v at the video ram side if I lift the coils for testing? Is it possible to boot the laptop like this and see the GPU working if it is not faulty or does it need power good signal from the power supply for the laptop to use it?

Many thanks.

[mcplslg123]

You can try to boot it with 1.35V injected on the consumer side of VRAM coil(Coil removed). Check schematics to see if the powergood of this pwm regulator is used anywhere.

[reessi]

I'm happy to report I fixed it! Although I think I should have spotted it before posting!

I neglected to check the power supply to the chip (RT8816GB). After checking the datasheet I realised it should be 5v on pin 18 (PVCC) but it was at 1v!

I then noticed there was a burnt dot in the middle of the 2.2ohm resistor (PRW1) feeding it from 5v Always. There was 5v before it. It was partially open, I replaced it and it burnt again. After a bit more messing around I realised the capacitor (PCW17) connected to it was partial short at 200ohms.

As it wasn't full short when I measured all the pins on the chip I assumed 200ohms was the resistance of pin 18!

Removed the cap and replaced the 2.2ohm resistor and all is well!

See attached pic!

Hope this helps someone! Could have solved a lot quicker if I had checked for the input power to start with!


Prior to finding the problem I did test by injecting 1.35v and got these results:

Laptop off -> ~450mah,

Booting ~2.1a booting and 4.1a in windows but windows keeps crashing with random errors like watchdog.

What could be the reason it wouldn't work off the bench power supply?

Current limit was 5a but it never attempted to go above 4.1a!


I have attached where Power Good goes, I do not know what it is doing, it ends up on a line called PEX_CLKREQ#, which appears to be PCI Express! May be this is why it did not work...

Could it have worked if I had also injected a Power Good signal as well?

Just thinking for future repairs/testing say I was waiting for a new power supply IC.

[mcplslg123]

Whenever a 2.2 ohms resistor is burnt on input of a pwm ic,the IC should also be changed for stable repair.(The resistor got burnt due to excess current which hit the pwm Ic).
So i prefer changing the 2.2 ohms resistor,cap connected to it as well as the pwm ic itself.
Very likely it would've worked if powergood was also supplied voltage from external source.

[reessi]

Ok thanks.