Core i7 short to ground, possible cause?

[hgg]

Hi guys,

I am trying to troubleshoot a used laptop motherboard I bought which refuses
to power on. Before the problem, I've installed a mSATA SSD and installed
Windows without any problems. At the end I powered off the laptop and I
wanted to install an antivirus as well. After the shutdown it never powered on
again.

I removed the motherboard and what I found was a short to ground on the
5V rail. After a lot of searching I could not find the faulty component and
I injected 3V @ 3A in the 5V rail in order to see what component will get hot.

This was the result :


It looks like that there is a short somewhere on the i7 CPU!
You can actually see the temperature rising pretty fast the moment you
connect the power supply :



The hotspot is on the lower left side of the CPU die. The back side is cool.



Is this motherboard toast ?
How common is it to have a short to ground on the CPU?
Is there anything else that I can check ?

Thank you for your help!
George.

[caspian]

Re: Core i7 short to ground, possible cause?

probably multiple points are shorted.
now check all coils to find another shorted coil.
If the CPU is soldered and is shorted, probably this board does not worth repairing.

[hgg]

Re: Core i7 short to ground, possible cause?

probably multiple points are shorted.

On the CPU you mean?

[caspian]

Re: Core i7 short to ground, possible cause?

no, you have to search for short all over the mainboard where there is a coil.

[khaahk]

Re: Core i7 short to ground, possible cause?

What mainboard? How did you locate the 5V rail? Do you have schematic? How many ohms is short to ground? Usually it is not necessary to inject over 1V to find what is heating.
CPU-s die too, thats correct.

[hgg]

Re: Core i7 short to ground, possible cause?

@khaahk Its a Lenovo laptop and I have the schematics.
Yes 3V was not necessary. It was a 5V rail though.

@caspian I measured all the inductors for shorts to ground:

L26 (5V) : 0.8 Ohms
L27 (3V) : M Ohms
L24 (Charger Out) : 20 KOhms
L25 (5V) : 0.8 Ohms
L34 (VCCSA) : 22 Ohms
L31 (VCC1R05AMT) : 78 Ohms
L29 (VCCGFXCORE) : 6.8 Ohms
L30 (VCC1R05_VTT) : 0.8 Ohms
L32 (VCC1R5A_VX) : MOhms

5V rail is shorted for sure.
Now L30 looks like a short as well and VCC1R05 might be the CPU voltage.
Its connected on the output of the VT357FCX-ADJ (DC-DC Converter)

I cannot find any detailed data on the VT357FCX-ADJ.
It looks like its the DC-DC converter that takes 5V in and converts it to 1V.
Is it possible then that this is the component that shorts the 5V and the 1V rails?



The thing is that the motherboard heats up only on the CPU and nowhere
else and the CPU does not use 5V afaik.

What do you think?

Thanks.

[caspian]

Re: Core i7 short to ground, possible cause?

upload the schematics file to have better understanding.
is the CPU soldered to the board?

Is it possible then that this is the component that shorts the 5V and the 1V rails?
yes.

Now look for short on the components connected to VT357FC and L30 and L26 and L25.
capacitors and Mosfets have got the highest probability of being shorted.

[hgg]

Re: Core i7 short to ground, possible cause?

I have already found the components that have a short because
thats where I started from. Two MOSFETs on the 5V rail present
a short, but that does not mean they are actually shorted. Anything
on the 5V rail could be the cause. For example all the capacitors
on the same rail present a short. That is why I used high current to
find out what was getting hot.

The thing is that if its the VT357FCX that shorts the rails, then why
it does not get any hot ???

You can find the schematics online : https://goo.gl/XZ1n1v

[hgg]

Re: Core i7 short to ground, possible cause?

Just checked and there is a direct short (0.35 Ohms) between VDD (input) & VXD (output)
of the VT357FCX chip.

[caspian]

Re: Core i7 short to ground, possible cause?

now you have to replace suspected components by new ones until no short exists on the mainboard. To save money, only replace suspected capacitors and Mosfets at the first step. and let controller ICs stay.

The thing is that if its the VT357FCX that shorts the rails, then why it does not get any hot ???
If it is not getting hot, then it is not a serious suspect of being shorted. So leave it alone for start. After replacing suspected capacitors and Mosfets, you may conclude that VT357FCX is either faulty or healthy.

If the CPU is soldered to the board, you may need to replace the CPU as well.

[hgg]

Re: Core i7 short to ground, possible cause?

The CPU is soldered to the board, so replacing it its out of the question because
I do not have anything close to a BGA rig.

When you apply current to the shorted rail isn't logical to assume that only the
shorted components will get hot? The only thing that gets hot is one side of
the CPU die.

[caspian]

Re: Core i7 short to ground, possible cause?

only one of shorted components gets hot and other shorted components stay either cold or a bit warm.

[hgg]

Re: Core i7 short to ground, possible cause?

So, maybe its more convenient to start removing inductors and see which part of
the circuit remains shorted in order to close on the faulty component.

What do you think?

[caspian]

Re: Core i7 short to ground, possible cause?

yes, this is a good idea.

[hgg]

Re: Core i7 short to ground, possible cause?

Great, I think this is what I will try.
I will let you know.
Thank you very much for your help!

[hgg]

Re: Core i7 short to ground, possible cause?

The 1.5V rail inductor L30 has been removed and the 5V rail cleared
of any shorts. MOSFETs and capacitors are all ok now!

If you have a look at the diagram some posts above, pin 1 of R819 (VDD) and
pin 1 of inductor L30 (VXD) of the DC-DC VT357FCX chip still have a short.

Does this mean that the IC is definitely fried ?

The other thing is that on the VT357FCX pin VCC_SENSE_VTT is also
shorted to ground and a series of capacitors along the CPU VCC1R05
rail are all shorted to ground (C113, C115 etc)

How can we find out if the DC-DC vs the CPU are causing the short now.
The DC-DC is a tiny BGA chip and it would be quite difficult to resolder
once removed.

[hgg]

Re: Core i7 short to ground, possible cause?

Ok, with L30 removed, I applied 0.7V/1.5A to pin 2 of L30 and I got the same
thermal signature. Lower right side of the CPU in the photo getting hot again.

Although the DC-DC chip shorts its input with its output please have a look at
the following component schematic:



All the blue pins are interconnected with the remaining shorted rail.
The marked capacitors on the lower left are all shorted to ground as well.

As you can see, this line goes to the DC-DC chip VT357 as well. (lower right)
Is it possible that the short is been created by this chip and that the CPU
is actually fine?

Thank you.

[caspian]

Re: Core i7 short to ground, possible cause?

all the components which you mentioned cannot be shorted. Usually, one or two components are shorted. you have to test them one by one to discover the shorted ones. or you can apply an external voltage that has the risk of burning the CPU or PCH.

[hgg]

Re: Core i7 short to ground, possible cause?

I have tested the capacitors on the lower left and they are all sorted to ground.

[caspian]

Re: Core i7 short to ground, possible cause?

they cannot be all shorted.
if one of them is shorted, the multimeter shows the two lines shorted for all the parallel capacitors.