Dell inspiron 13 5000 - Doesn't turn on, overheating

[Nagyesz]

I have a Dell laptop with me, a DELL inspiron 13 5379. It doesn't turn on. After some investigation, I found that the charger is delivering 0-3V instead of 19.5V, and this voltage fluctuates constantly. Additionally, two ICs are overheating. I searched for these ICs, and it seems they are likely related to the issue of not having 19V.
Should I inject voltage somewhere, or should I replace the ICs right away?

So, PU4301 is overheating, and the surrounding area is hot, possibly just because of this IC.

[mon2]

Hi. No injection, yet. Both, PU4301 & PU4302 are mosfets. They are acting like solid state relays / switches to pass through the adapter voltage or not to power the logic board.

Remove all power. Meter in resistance mode.

Measure the resistance across each leg of the mosfet. Post each measurement.

Measure:

source (1-2-3) & drain (5-6-7-8)
source (1-2-3) & gate (4)
gate (4) & drain (5-6-7-8)

then measure the resistance to ground of the each group. Repeat for the next mosfet.

We are checking if a part on the rail they power is shorted. We are checking if the mosfet itself is shorted through resistance (ohms) measurements.

[Nagyesz]

PU4301
Source to Drain 5Mohm
Source to gate 2Mohm
Gate to drain 125kOhm

Ground to drain 114kOhm
Ground to source 125kOhm
Ground to gate 45kOhm

PU4302
Source to Drain 112kOhm
Source to Gate 100kOhm
Gate to Drain 225kOhm

Ground to drain 110 KOhm
Ground to source 3.5 Ohm (!!)
Ground to gate 100 KOhm

[mon2]

Ok. Remove all power. Meter in lowest resistance mode.

One meter probe on the source pin that appears to be showing a short.

Other meter probe on any inductor on the board. Check if there is a low resistance. Keep testing each and every inductor on the board.

One meter probe will always be on this source pin group of the mosfet.

List all suspected inductors. We are checking if there is a leaky high side mosfet. If for example, the cpu inductors are a low resistance in this checking, then the cpu is probably dead.

[Nagyesz]

Found it.



PL4701 PL4702

PL4701 to source pin is 0.3 Ohm
PL4701 to ground 4 ohm

PL4702 to source pin is 10 ohm
PL4702 to ground 6 ohm

also found this near to these inductors

[mon2]

See attached. This area looks like it has been reworked. The part with the yellow arrow is a suspect. Flux and carefully remove this capacitor. Do not lose the part.

Then check if you still have the very low resistance on the source pins of the mosfets.

[mon2]

PL4701 to source pin is 0.3 Ohm
PL4701 to ground 4 ohm

PL4702 to source pin is 10 ohm
PL4702 to ground 6 ohm

These measurements are of concern and usually imply that the high side mosfets that are internal to the regulators @ PU4701 and PU4702 are leaking and this will often kill the CPU.

That is, the high voltage from the adapter of ~19 volts would have travelled to these switching regulators. The high side mosfet turns on/off to regulate the voltage for the CPU to be ~1 volt.

If the high side mosfet (either one) is leaking, then the full ~19 volts would hit the CPU power rail.

For closure, would you have an adjustable power supply? If yes, dial it to 1 volt (only) and inject the voltage at the source pins of the mosfet and check who is heating up. Is it the CPU?

If you do not have a power supply, then flux and remove either PU4701 or PU4702. Then measure the resistance to ground again. If needed, remove the other VCORE regulator. Test the resistance again. Proceed only if you feel comfortable or test the process on a donor board.

[Nagyesz]

With the PSU, there was no heating when conducting 1V through the source; the machine drew 0.6-0.8A out of 2A. Now I will try the other method.

edit: Are you sure to remove PU4701 or PU4702, wouldn't be better to remove the inductors?

[mon2]

Some part on your main power rail is partially shorted. The resistance of 3R5 to ground is too low.

Try to lift one leg of the local inductor at PL4701.

Test the resistance again. Suspecting the short is inside one or both of the regulator chips.

[Nagyesz]

Unfortunately, after removing the inductors, the suspicion of a short circuit remained on the processor side. In both areas where I lifted the inductor legs, the processor side is around 2 ohms (VCC_CORE) , and from other side it's OL to ground (PWR_VCORE_SW1-2) .

Edit: I suppose if I plugged it into the charger now, the voltage wouldn't jump between 0 and 3, right?

[mon2]

Having a low resistance on a low voltage rail like the VCORE is ok. This is by design.

Can you check the source pins again at the mosfet to ground resistance? Do you still have 0.3 ohms to ground? Perhaps you can share a pic of the meter for a cross check of the measurement.

Do not plug in the power with these parts removed.


PS: Perhaps with some luck - the internal mosfets inside of these VCORE regulators stopped the flow of the high voltage to the VCORE pins of the CPU. My suspicion is still with these regulators and should be replaced. The key point is that the source pins should be with 0R3 to ground.

[Nagyesz]

Without the two inductors, the PU4302 Source is 2-3MOhm
VCC_CORE is 2.7 Ohms and 3.7 Ohms PWR_VCORE_SW2 is OL PWR_VCORE_SW1 is 170k Ohm
However, when any inductor is reattached, the PU4302 Source remains at 0 Ohms

Could I be lucky this time?

[mon2]

Do wait on feedback from other techs but in my opinion, these regulators have internal high side mosfet shorts. If you wish, replace both of these regulators and then check the resistance at the mosfet source pins again. If the source pins are at a high resistance after the regulator ICs have been replaced, then you may be lucky to try again. My concern is if the high voltage reached the CPU core voltage rails or not.

As a test, inject 1 volt at the source pins of the mosfet. Then monitor the voltage at the inductors that you have lifted. Do you see a voltage that is similar to what you are injecting?

[Nagyesz]

I tried it, with 1V injected precisely at the legs of the inductors, both inductor legs are fine, and there is no current draw, whereas when the inductors were connected, the machine drew 0.6-0.8A.

So, from what I understand so far, something is short-circuiting, because instead of 19V, the machine fluctuated between 0-3V. After removing the inductors, the MOSFET source leg was fixed. When I put the inductor back, there's a short circuit on mosfet source (0 ohms).

Unfortunately, I'm stuck here; could it be my luck, or is the processor the issue?

[Nagyesz]

Edit

I tried it, with 1V injected precisely at the legs of the inductors

I mean to the source pins of the mosfet, and same Volts could be measured from the inductors