[Solved] RX 460 Blown Fuse

Just out of curiosity what hot air station have you been for this project?

I've been trying to talk myself into picking one up.
 
From what I found, 81161 is decreasing resistance of this resistor (without it resistance is 10K, just like after desoldering this resistor from the board). But on this phase even with a 81161 from the other phase it still has that 8.7K resistance.
Once again it was just one pin on 81161 that caused this. It wasn't soldered properly, because I really should use solder in paste or at least clean the old solder and not just reflow the solder that was there. I'm still learning hot air soldering. But anyway after making sure that every pin is soldered correctly that suspicious resistor has now 8K - which is I guess good. At least close enough to 7.8K-7.7K of other resistors.

So let's summarize what we learned now. It's not only a mosfet, not NCP81161 (that time gate to ground voltage was low I think I just killed the 81161 by running into the mosfet failure few times), nor the bootstrap capacitor. Also on my last try to run the card after fixing that resistor issue I once again checked the gate to ground voltage and it was the same as on the other phases, but it was hitting 13v more frequently. Which I think is normal, because this phase seems to be the main phase that is always used the most. Anyway card died when I was redoing my measurements - problem is still there.

In my free time I checked NCP81022 - VRM controller. And one pin made me suspicious. It's called CSREF and I have there like 2 ohms to ground. Pin description is in the image below. Should it be like that?

csref.png
 
Hello, it's a bit old here, but I just found that you have super photos of this graphics card. I'm trying to repair a rx560 and IC I'm looking for looks very similar. So if someone could send better photo what IC is it specifically I would be really grateful.
1610660789771.png
1610660860198.png
 
Just out of curiosity what hot air station have you been for this project?

I've been trying to talk myself into picking one up.
FYI I got an 878DD it's the big brother and it's a great station (solder and hot air). Paid for itself already many times over fixing LG v10s lol.

Pretty heavy actually and looks skookum built for a cheap unit.
 
FYI I got an 878DD it's the big brother and it's a great station (solder and hot air). Paid for itself already many times over fixing LG v10s lol.

Pretty heavy actually and looks skookum built for a cheap unit.

In life I've found that AIO unit's rarely do as well as a good singular use unit, plus I got a sore dick deal on a Hakko iron and it's pretty hard to compete with it at the pricepoint I paid.

Appreciate the follow up, I had to backburner my hot air station quest due to other priorities but at some point I'll get back around to it.
 
Another card with a damaged MOSFET on the same phase. After removing the short circuit, the card works and its performance checked in the FurMark program.
 

Attachments

  • Gigabyte RX 460 FurMark PCI-e 2.jpg
    Gigabyte RX 460 FurMark PCI-e 2.jpg
    43.9 KB · Views: 0
  • Gigabyte RX 460 GPU PCI-e 2.jpg
    Gigabyte RX 460 GPU PCI-e 2.jpg
    110.2 KB · Views: 0
  • Gigabyte RX 460 WINDFORCE OC 4Gb MOSFET.jpg
    Gigabyte RX 460 WINDFORCE OC 4Gb MOSFET.jpg
    596.8 KB · Views: 0
Last edited:
  • Like
Reactions: Kzoak
like this
Hi. I just fixed a card with the exact same problem. First card I ever fixed in my life. I was lucky that the mosfet was visibly blown so I knew immediately what the problem was. My question is, instead of replacing the same part number, can I install a uprated mosfet? I have no idea how to compare mosfet datasheets, what would be compatible and what to look for when trying to get a better one. I am also thinking of sticking in some tiny vrm radiators to help with heat dissipation. Any kind of help would be appreciated. Thank you
 

Attachments

  • 20211031_121839.jpg
    20211031_121839.jpg
    507.9 KB · Views: 0
Once again it was just one pin on 81161 that caused this. It wasn't soldered properly, because I really should use solder in paste or at least clean the old solder and not just reflow the solder that was there. I'm still learning hot air soldering. But anyway after making sure that every pin is soldered correctly that suspicious resistor has now 8K - which is I guess good. At least close enough to 7.8K-7.7K of other resistors.

So let's summarize what we learned now. It's not only a mosfet, not NCP81161 (that time gate to ground voltage was low I think I just killed the 81161 by running into the mosfet failure few times), nor the bootstrap capacitor. Also on my last try to run the card after fixing that resistor issue I once again checked the gate to ground voltage and it was the same as on the other phases, but it was hitting 13v more frequently. Which I think is normal, because this phase seems to be the main phase that is always used the most. Anyway card died when I was redoing my measurements - problem is still there.

In my free time I checked NCP81022 - VRM controller. And one pin made me suspicious. It's called CSREF and I have there like 2 ohms to ground. Pin description is in the image below. Should it be like that?

View attachment 266980

Got this card from someone who said it didn't work anymore after PSU crash and replacement. Fuse blown, checked 3rd MOSFET it was shorted. In such a state there are resistances on the phases resistors equivalent to R510:
Phase1 - 8.03K
Phase2 - 7.96K
Phase3 - 0.15K
Phase4 - 7.83K

It appears on each next phase resistor there should be ~65ohm less than on previous one. Removed the MOSFET, now on phase3 there is 7.80K, and it should be 7.90K (maybe it would be OK if new MOSFET is soldered). I think this is the reason there are failures soon after MOSFET is replaced, probably something else on the failed phase is damaged too, maybe low sides, driver or main PWM

I also checked CSREF pin - about 4ohm to GND on this card. Looks like the card failure actually took out the PSU.
 
  • Like
Reactions: Kzoak
like this
Back
Top