MSI 1080 Gaming X keeps killing a particular MOSFET, gate driver problem?


May 8, 2022
Hello, I had my PC fail to boot and deduced that there was a short on the 12V rail of my GPU causing the PSU to shut down. I watched this video and found that I had a similar looking problem, measured 4.9K on Pin 1 (high side gate) of all the MOSFETs apart from one, which showed 0.3 Ohms just like when I measured at the PCIE power connector.

The MOSFETs on my card are these:

I replaced the part with this:

When I removed the original part the short disappeared, and it didn't return when I fitted the replacement, looking good, I thought.

Well, I plugged it back into my PC and it presented exactly the same failure that it did originally. I removed the card again, measured it, and the new MOSFET has failed exactly the same way the original one had.

I dropped a comment on the AHOC video for my card, and Buildzoid reckons the gate driver might be the culprit, which makes sense. Measurement agrees, I'm seeing 1.3 to 1.5 Meg on pin 1 of the rest of them, and 5K on the suspect one.

I can't read the part number on the gate drivers I have, but I have found forum posts from someone else who has read "DU F526" on theirs and is looking (unsuccesfully) for a replacement. Failing that, I'd need to figure out a suitable set of specs myself but I'm not quite sure where to start. I suppose I'll need a voltage rating comfortably higher than 12, an output voltage that's higher than the MOSFET threshold, and fairly short rise and fall times (how short matters?). They are 8 pin devices, square, about 2mm on each side.

So after all that, is there anything else I should check to minimise the chance that it just does the same failure again?


How are you replacing the part in question?

Also, have you ruled out a bad bootstrap capacitor?
How are you replacing the part in question?

Also, have you ruled out a bad bootstrap capacitor?
I'm using a hot air station with plenty of flux. I haven't got a lot of experience working with SMD but it went smoothly enough. I have done plenty of through hole.

I'm not familiar with the bootstrap capacitor in this circuit but I can certainly take measurements of it if you can tell me which it is. I've heard of bootstrapping MOSFETs to get an increased voltage output beyond the supply voltage, but I'm not sure if that's the same thing here.



  • front.jpg
    256.4 KB · Views: 1
You'll have to look at the phase driver IC datasheet to figure out which capacitor is the bootstrap cap. It will be connected at one end to the switch node and at the other to the phase driver IC's bootstrap pin. Sometimes called "boot" or "cboot" or something like that.

I would guess that your real problem, though, is that you're using too much heat with the hot air station when you remove and install the replacement FET package. There's so much copper in this part of the board that you end up having to really heat the crap out of it in order to get the solder to melt, and you end up overheating the FET packages in the process. I would try again using a preheater to get the board up to like 170 or 180 before using the hot air wand to get it the rest of the way. Edit: You may also need to replace the phase driver IC for this phase as well. It's possible that got damaged in the process too.
I faced a similar issue with my 980 Ti but it was actually TWO mosfets, not just one. Really strange how this problem is wide spread among the high end Nvidia GPUs... Hmm.....
Anyway here's the strangest part: I decided to run the card with both of those mosfets missing and it works perfectly fine at stock clocks! Its been almost a year now with no problems whatsoever! Amazing!
If you really want to play it safe you could try underclocking which greatly improves the card's efficiency.
I'm planning to finally get a new card later this year :)
Last edited: