Here the board, nothing seem to be burned or damaged.

What ever i do now with the jumper to set the voltage is now ineffective and +5V go directly to the socket... 🙁

Alfox wrote on 2025-07-08, 19:37:

i was testing without the cpu...

It's hard to guess. How are you measuring things, with no CPU? It could be fine...

Re: Epox EP-58MVP3C-M - 3.54v Vcore regardless of jumpers

nickles rust wrote on 2025-07-09, 01:18:

Alfox wrote on 2025-07-08, 19:37:

i was testing without the cpu...

It's hard to guess. How are you measuring things, with no CPU? It could be fine...

Re: Epox EP-58MVP3C-M - 3.54v Vcore regardless of jumpers

That's also what I thought. Install a CPU. Try a regular K6-2 in case it goes bust (way easier to replace than a K6-2+).

Keep this in mind:

It is easily possible to accidentally damage the VRM on this board by jumpering it to the P55C setting (dual voltage plane) with two mismatched voltages for Vcore and L2 cache if inserting a single voltage plane processor like the original Pentium, AMD K5 and Cyrix models.

Source:
https://theretroweb.com/motherboards/s/pcchips-m577

It is very easy to mesure cpu tension without cpu.

Before my mistake, i was mesuring 2.2v at the mofset... without a cpu too.

And no way i'm putting a cpu when i read 5v at he same placee when i was reading 2.2v before 😁

I' don't have a lot of K6-2 and i don't want to fry one just to see.

Thank for the Epox i'll read that to see if some things are similar to my case.

In any case, is 5V will kill a Pentium MMX who normaly run @ 2.8v ? I can try one, but it ts a 233Mhz it'll hurt if i kill it...

I repaired this one some time ago: PCChips M575 - repair help - (done)

Looks similar

Thank i'll look at it! I have already place an order on a n mofset and a skotty diode replacement. But it is a blind guess and i don't know if it 'll work

Q8, 17 and 18, what's written on them?
I guess them to be connected somehow to deliver proper voltage.

Q8 is D2583LM

Q17 is PBYR1025D

Q18 is PHD45N03

Alfox wrote on 2025-07-09, 18:11:

Q8 is D2583LM

https://www.datasheetmeta.com/pdf.php?q=D2583

Alfox wrote on 2025-07-09, 18:11:

Q17 is PBYR1025D

https://www.alldatasheet.com/datasheet-pdf/vi … /PBYR1025D.html

Alfox wrote on 2025-07-09, 18:11:

Q18 is PHD45N03

https://www.alldatasheet.com/datasheet-pdf/vi … HD45N03LTA.html

Too dead to put a sentence together but if it helps for tests or check stuff.

First one I'm not sure if it is correct, maybe someone else can confirm or correct.

Are you sure that you are measuring the core voltage and not just the input voltage of the switching converter? The correct measurement point is the terminal of the inductor coil that's facing the socket, this path goes directly into the CPU.

Measure the voltage from Q18 (the MOSFET) drain (pin 2) to the source (pin 3). If it reads close to 0V, that would mean it's shorted and passes the full 5V to the socket.

Something i notice, when power is OFF, no shortcut, but when i power the m/b, +5V and +12V are shorted to GND (the PS doesn't go in protect mode!) ???

Are you measuring the resistance from the power rails to ground with the board powered? Don't do that, you could potentially damage your multimeter, the board and the PSU. To detect a short circuit while it's running, measure the voltage across the suspicious points (low voltage=short). But even then, only do that if you know that the short is in a low-current part of the circuit. If it's a power delivery circuit, don't turn it on at all, try to diagnose it offline.

asdf53 wrote on 2025-07-09, 20:43:

Measure the voltage from Q18 (the MOSFET) drain (pin 2) to the source (pin 3).

I would suggest checking resistance between those two pins (with the power off) first. It's possible that the FET is still good, but something in the control circuit failed and is holding the FET on permanently.

asdf53 wrote on 2025-07-09, 20:43:

Are you sure that you are measuring the core voltage and not just the input voltage of the switching converter? The correct meas […]
Show full quote

Are you sure that you are measuring the core voltage and not just the input voltage of the switching converter? The correct measurement point is the terminal of the inductor coil that's facing the socket, this path goes directly into the CPU.

Measure the voltage from Q18 (the MOSFET) drain (pin 2) to the source (pin 3). If it reads close to 0V, that would mean it's shorted and passes the full 5V to the socket.

Something i notice, when power is OFF, no shortcut, but when i power the m/b, +5V and +12V are shorted to GND (the PS doesn't go in protect mode!) ???

Are you measuring the resistance from the power rails to ground with the board powered? Don't do that, you could potentially damage your multimeter, the board and the PSU. To detect a short circuit while it's running, measure the voltage across the suspicious points (low voltage=short). But even then, only do that if you know that the short is in a low-current part of the circuit. If it's a power delivery circuit, don't turn it on at all, try to diagnose it offline.

Yep, i have 4.85V at the socket on VCC2 and 3.52 on VCC3 (i little high but fair enough)

On Q18, i mesured 0.04V from drain to source...

And yes i was checking for short while power on, i not good ? Oups, i'll don't do that again!

lti wrote on 2025-07-10, 04:01:

asdf53 wrote on 2025-07-09, 20:43:

Measure the voltage from Q18 (the MOSFET) drain (pin 2) to the source (pin 3).

I would suggest checking resistance between those two pins (with the power off) first. It's possible that the FET is still good, but something in the control circuit failed and is holding the FET on permanently.

i have 97 ohms between these two pins

97 ohms means it's not shorted. Can you also test the MOSFET's gate voltage (pin 1)? If it's high (anywhere from 5-15V), that would indeed suggest that the MOSFET is good, but it's held on permanently because something in the control circuit has failed. You can also test the MOSFET's switching frequency (use multimeter in Hz mode and measure pin 1 to ground). There's a chip KA34063A below the MOSFET, this could be the switching controller. It's supposed to turn the MOSFET on and off quickly to create a lower output voltage from the 5V input voltage.

https://www.mouser.de/datasheet/2/308/1/KA340 … 3_D-1810306.pdf

It has a pin 5 "Comparator inverting input", this could be the feedback pin to sense the CPU voltage. What voltage is at the pin?

Yes i saw this chip, it is my next guess.

Pin5 is at 2.89V

Please also test the MOSFET as I suggested above (gate voltage and switching). And I must ask again - are you sure that you measured a voltage of 5V at the inductor terminal that leads to the CPU socket, and it wasn't just the input voltage of the switching MOSFET (pin 2)?

V in Q18 is 4.92v (at Drain)
V out Q18 is 4.88V (at Source)

And so is the voltage at the coil, 4.88V

I'have dont have acces to an osciloscope to see the switching. I'm might buy one.

Another solution is, just below Q8 is another KA34063A chip. I'm tempted to desolder this to try te replace the one driving Q18.