LA-6901P +VGA_CORE shorted

[Sojke]

Hello, I have an Packard Bell Easy Note TS13HR-197GE with motherboard LA-6901P that is shorted at +VGA_CORE and because of this the laptop turns on for a fracture of a second and than turns off.
Now, I am pretty sure that its the GPU chip that is shorted because I removed some components I suspected , such as PQ44 and PQ45 , and they check out OK, so I removed also the controller PU10. But now, all of a sudden, B+ is shorted to ground as well, and laptop adapter goes into fail safe mode where it shuts it self off.
After removing PL15 that connects B+ to +VGA_CORE the laptop started normally and even displayed picture (intel HD graphics).

What I am wondering is, why is B+ connected to +VGA_CORE via PQ42 and PQ43? What is the point of this? And is it possible that these transistors failure caused N13P-GL2-A1 to die?

(The scheme is edited, most lines are erased take a look at the pdf scheme provided)

[piernov]

Re: LA-6901P +VGA_CORE shorted

What is resistance to ground on +VGA_CORE without PL16?

IF PQ42 or PQ43 are shorted, this will cause B+ to short with +VGA_CORE: that is *not* good at all, this means sending 19V to the GPU, so it'll most likely die. Do not try to plug in the power supply with this short, you must first ensure that B+ and VGA_CORE are not shorted anymore before trying to apply power. Measure resistance between B+ and VGA_CORE.

[Sojke]

Re: LA-6901P +VGA_CORE shorted

Hi, it is definitely shorted. Power supply shuts down when connected and laptop is powered on. Resistance is very low. Diode mode test reports in one way 002 in other way 016.
There is a black spot at PQ44 and PQ45, but they report ~460 in diode test mode in one way (When removed from board).

What is the purpose of connecting B+ to +VGA_CORE? Does the PU10 manipulate them to get 1.5V?

[piernov]

Re: LA-6901P +VGA_CORE shorted

Please report the resistance in ohms, not the voltage drop, both between B+/+VGA_CORE and +VGA_CORE/Ground. This is a better reading (IMHO) for power lines (especially high-power low-voltage as +VGA_CORE is).
Please post picture of PQ44/PQ45 area if there's anything looking funny.

This circuit (with the most important parts: the controller PU10, the set of 4 switching MOSFETs, the coil PL16 and the output caps PC171/PC108) is a DC to DC converter, in the synchronous buck topology. You can read more on the internet about this but basically it converts a higher voltage line (B+ at 19V) to a lower voltage line (+VGA_CORE at ~0.9V) with a high output current capability.

[Sojke]

Re: LA-6901P +VGA_CORE shorted

PQ42 has blown and is the culprit of the situation. It is an MagnaChip MDU2657.
I will have to replace it including the graphics chip. i just hope all the 2.5V caps did not die from this.

Any one know of any common alternatives to MDU2657?

[Edit]

Was writing when you posted.

Resistance of PQ42 is 12ohm.

[piernov]

Re: LA-6901P +VGA_CORE shorted

Ok so PQ42 measures 12 ohm (between source and drain I guess) out of circuit? You should replace it with exactly the same part. If you cannot find it, find a N-channel MOSFET with very similar Vds, Id, Rdson characteristics, and replace both PQ42/PQ43 with it. I've got some 4921N which may be suitable, not sure.

Now resistance between +VGA_CORE and ground?

On your picture it looks like discoloration due to heat.

[Sojke]

Re: LA-6901P +VGA_CORE shorted

In both ways, the other way around is 14ohm. There is still a big short between +VGA_CORE and GROUND.
I can only guess either the chip it self or some of the smaller capacitors. But Ive check them under an microscope (Not a very good one) and I didn't see burns or cracks, which only leaves the chip as the suspect, but the chip onboard (on top of carrier) capacitors are not in short from what ive measured.

[piernov]

Re: LA-6901P +VGA_CORE shorted

Low resistance between +VGA_CORE and ground doesn't mean the GPU is shorted, it's normal behaviour. GPU and CPU (and PCH, but a bit higher) always show low resistance to ground, that's perfectly normal. It doesn't mean the chip is good, but a resistance of a few ohms mean it's not completely shorted at least. If the chip is shorted, you'll see a resistance lower than 1 ohm.

I would try to replace PQ42, PQ43, maybe PU10 and see what happens.