Both of these P-channel mosfets are suspects. A good mosfet should be hundreds of k ohms or higher (when measured off the board).

Start with #2 mosfet -> flux and remove it and then test the resistance again when it is cool and on your work bench. Still this low in resistance? If yes, replace it.

Repeat with #1 mosfet. Also confirm that the main power rail is no longer shorted with these parts off the board.

Dear mon2,

I have replaced mosfet #2 (values off the resistance were low when removed from the board, and cool). After measuring again mosfet #1, the values of resistance are good. So my 4 mosfets are good now.

Now, when I plug the board to my power supply, the board draws 17mA, and after one second there is a current peak and a really brief voltage fluctuation, and I can hear a small click from the board. It goes back to 19V and 17mA for a second, then another click, and this happens every second. When I activate the Over Current Protection of my power supply (set at 2.3A), it stops after the first peak, which I believe means that the current peak is more than 2.3A.

Also, there is still a short to ground on the main power rail (I have measured it at the shunt resistance). I believe I have to find the short now. Can I inject voltage after the shunt resistance, and if not, how can I know where to inject without boardview or schematics?

Regards,
mkdj

What is the exact resistance to ground for the short on the main rail?

Remove all power. Meter in resistance mode. Touch your meter probes together. Note the resistance.

Place one meter probe on the main rail. Other probe on the numerous inductors near the CPU. The CPU will always be a multiphase power supply with a mix of power supplies to share the heavy currents.

What is the resistance between these points of measurement? We are testing to see if there was a high side mosfet that has leaked. Then it is usually game over since the ~20v has hit the CPU rail.

Dear mon2,

Resistance to ground for the short on the main rail is 1 Ohm.

When I touch my meter probes together I read 2 Ohm.

Resistance between main rail and inductors are for instance 5, 8, 60 Ohm, on three different inductors.

How bad is this?

Regards,
mkdj

Dear mon2,

I have found the short on the main rail. That was a capacitor. I removed it, and I don't have a short on the main power rail anymore. I haven't replaced this capacitor yet. Also, I eventually had to change the second suspected mosfet. Now I have 19V on the main power rail (measured at the shunt resistance). I have checked voltages at some coils, and I have 3.3V at one of them. I have 0.22V on another, but no voltage on the others.

I have put the motherboard back, plugged the CPU fan, the screen and the keyboard for the power button. When I press the power button, I have no image, and the fan starts, but stops immediately. Current drawn by the board stays at 370mA, and nothing else happens.

I have checked resistance to ground of the inductors, and I have fairly low values on many of them, like 290 Ohm, 4 Ohm, another 4 Ohm, 21 Ohm, 55 Ohm.

EDIT: I had not put back the RAM... Now I have, and it works, I have picture!

A big big thank you mon2 for your help! I have one last question: do I need to replace the bad capacitor? What are the risks if I don't replace it?

Regards,
mkdj

Congrats on the successful repair!

While you can share the exact location of the cap for a review, the cap does not need to be replaced. The logic here is that the designers will majority of the time apply multiple caps on the same line for filtering the ripple of the power supply. Respectively, if this one cap is missing, will the power supply halt from operating correctly? No. You will be fine. If it keeps you up at night, certainly order a few spares of this part for a future replacement project.