Replacing a burned mofset on a P8P67 Pro

Hello All,

(Skip down if you don't give a hoot about the story)

I'm by no means electrically savvy I have basic knowledge on how to use a multi meter... However here's how the story goes.

I was given two intel P3700 NVME drives as engineering samples to test and use, however I hit a snag. It seemed when I installed them on my p8p67 I could not post. After some lenghty torubleshooting it turns out that using the 4th DIMM slow will cause an incompatibility of that card is installed and ASUS gave the big middle finger even acknolwing this. Telling me that I can send the board in for a paid RMA to find out if there is anything wrong with the board.

Not satisfied with this answer, I started doing some searching and found a site (win-raid.com) which had posts about how to make virtually any UEFI bios boot to pci-e drives!! After hacking my bios to allow it to boot from this drive, I decided that I would try a few other things and possibly flashing my bios to a similar p8p67 board. After all if the flash went bad who cares, the hardware was so similar it may just not function or detect right??

WRONG...

Flashing to the p8p67-LE I failed to read that it actually had less phases for cpu voltage regulation. After flashing the motherboard was able to semi boot half way to windows before never posting again. I ordered a new bios chip and programmer and restored the original bios but it was too late what ever I had damaged it seemed there was no turning back.

I turned to craigslist for a new board and found a p8p67 PRO which was for sale for 15$.

I gambled and after setting it all up on a bench test, it failed desirability. I would turn on the power but the board would immediately turn off and the fan would move not even one tick more like a nudge. After removing the VRM and mosfet heatsinks and turning on the power again, I noticed that a puff of smoke. Narrowing down the area of smoke I tested a Mosfet with my multi meter only to find that there was no resistance.

Since I do not have a SMD rework station I was at a loss on how to remove this chip and replace it from my other p8p67 board which had the same chip. Then it dawned on me, and I took my heat gun which I normally use for car things, and put on a reducer and then further reduced that by making a tinfoil nozzel!!! HAHA

*** Tech starts here ****

I removed the bad mosfet from the p8p67 pro and obtained a doner chip from my p8p67 which used the same chips.

When I powered up the machine the machine will keep power to the CPU now and the fan will spin however, the machine does not post. The red CPU LED will stay lit meaning there is a problem there, and although there was one bent pin on the BGA, I'm pretty sure I bent it back in to place and it should be a problem.

I guess my question is after replacing the MOSFET, and now being able to power the motherboard and it stay powered, how can I determine if the BGA pin problem or still a motherboard problem?

Since this board does not have a led readout of codes I cannot see what could be happening.

So the MOSFET failed short? I was under the (amateur here) impression that they provide power by pulse width modulation. So, shorting it like that could mean the CPU got some serious voltage.

But that's a far cry from an actual answer. Anyhow, are you still running the LE bios that caused the problem?
You saw smoke. Could you take a magnifying glass and give a careful look at the traces leading to the troublesome power phase? One phase consists IIRC from two MOSFETS, a capacitor and inductor. No bubbles on the PCB? no warping? no darkening under strong light?

Do the other MOSFETS check out fine?
Purely guessing because noone else has responded yet, but if the failure was so severe that smoke was involved, then the nearby capacitors could have been damaged.

If you're stuck with that killer BIOS image, then I can see a socketed DIP8 package IC (near the SATA ports) - if you could do a macro shot of it we could determine if it's the BIOS chip.
If so, you might need to send it to your local friendly geeks for programming it with the older version.

Hey Mchalrz,

Thanks for the reply..

Let me clear up the story I have 2 p8p67's. The my the first p8p67 is a non pro which was working fine until I decided to flash the LE bios. I ordered a new bios chip from ebay, but the machine didn't boot still. I then ordered an SPI programmer, which allows you to reprogram the chip using a USB dongle type thing. I verified what was on the chip and what was in the bios file and I know that it's all good.

Since I couldn't get that working I ended up finding a p8p67 pro for sale on letgo for 15$. When I picked it up i saw one bent pin, and when I powered it up after trying my best to fix the bent pin, I saw that it would immediately turn off like there was a short. Taking off the VRM and MOFSET heatsinks revealed that when I powered on the machine there was one mofset smoking and heating up really fast as you saw the top of the chip change colors for a second. Also it burn off some solder and made a little ball beside the chip. Testing that with the multimeter using diod mode gave me different results from all the other chips so I assumed it was bad. I pillaged the MOFSET off my old p8p67 which was the same chip, and after installing it with my ghetto rigged heat gun, I was able to turn on the system and it would stay on, but still not post. The CPU LED would stay lit indicating that there is a problem with the CPU.

So i took the CPU and tested in an MSI z77 board I bought and it didn't post. To ensure that the MSI board worked I took the I2600k that I had in my good machine and tested it on that MSI as well. it worked. So I assume the board was dead. Being very scared p8p67 pro would fry another cpu (my i2600k) I dind't want to put it in, but I eventually did try it out but still still unable to post. The issue I have now is I don't know if it's the Pin's that are causing the problem (the one that i tried to bend back) or the board still has issues.

2 things can be happening here:

1. The FET driver chip is broken and has fried your second MOSFET. (Same cause as first time.)
2. Your air-gun solution (ingenious, I give you that) was so hot it raised the MOSFET's junction temperature to the point were it broke.

Check your second MOSFET to see if it is OK. Be aware that to be completely certain it is fine, you must take it of the board, putting you in danger of option 2 above again.

I hate to say it, but it is unlikely you will get that board running again.

Any idea what a FET driver chip looks like? Or where to find it on the p67?

Thanks for the reply btw...

I ended up buying a rework station anyways just for shits and giggles to have for later times, but I got to say the heat gun worked pretty damn well

You can also isolate nearby parts by building an alu foil shroud/shield around and over the components that don't need the treatment of excessive 200'C + heat from your air gun.

It would depengt on which mosfet failed. If your high side fet burned, the rest of that phase and even the CPU may be goners.

Have you tried tracking the VRM signal from the controller to the driver and cutting that, so the board doesn't use that phase?

Putting a lot of effort into a $15 motherboard!

I was a bench tech for years.

It was common to see multiple chips and discrete parts blown when you have a failure that causes shorts and high current.
Not likely you'll get that board running right again.

Sounds like you have a blown CPU now too. You put it in the working MSI board and it wouldn't boot?

As stated above, too much time, effort, money wasted on an old obsolete board.

.