Gigabyte RTX 2070 WindForce No RGB, USB-C and E43 after Short on PCIE_12v
Hey guys,
So i got this RTX2070 Super from Gigabyte for cheap and the reason i actually got it in first place is, because it initially had a full short on PCIE_12v, so my plan was to basically feed some juice in it on my test bench and see what gets hot using my thermal camera. This would have potentially been a quick repair, however things didn't actually work out quite as planned.
Whenever i hooked up one of my lab power supplies to my testboard to feed the 12v rail, well, even tho i had it current limited at 12v, the charged capacitors actually blew the 10A Fuse on the card right out.
Not a big deal, i was thinking, lets just short this fuse and try it again without actually having the power supply already set to 12v.
Unexpectedly enough, nothing unusual happened this time around, with 3v3 and 12v hooked up both to pcie as well as to the psu connectors of the card, the thing started right up and consumed its normal 1amp at 12v. So whenever the fuse blew, it unfortunately also cleared the short out, which is basically where we are now.
The card has fanspin, it has 5v, 3v3, 1008mV PEX, 1368mV Vmem, 812mV Vcore, so these are all there, but USB-C doesnt output anything, also the buck converter which is connected to the USB-C Port via a mosfet, isn't running. Added to that, the RGB at the card doesn't light up and, while the card has limited (800x600) output via hdmi, DP also doesn't output anything.
In Windows the card won't be initialized and obviously devmgmt reports error 43.
While i normally would look at this as a failed micron GDDR6 chip, since the card did actually have a full short initially and other things aren't working, i don't think that thats the issue with this particular one.
So as i was basically expecting this card to be a quick one, i haven't actually got the time to virtually reverse engineer the circuit since there are no schematics to be found for this one, so that i can figure out which issue we may have.
Anyways, that's why i‘m here now, does anyone actually have experience with this particular card (or 2070‘s im general, really) and can basically tell me what's normal and what's not? I think it is pretty suspicious that USB-C outputs nothing, but then again, it could be that the driver needs to be fully initialized in order to get usb-c going and enable its buck converter, if that one is even specifically just for usb-c. Does it even need a separate one aka does it even have PD? Same thing with the RGB, does that maybe also need the driver to be fully initialized?
Actually i wouldn't expect that for both, but who knows. I don't have anything to compare and unfortunately, i havent found any datasheet for the MP8858GQ which is the USB-C Buck IC, so i can't tell which one is the enable pin, or if it's actually broken.
I would suspect tho, that this is just another 5v rail, that powers usb-c, RGB and something GPU internal which makes the DP Signal. I might actually try feeding 5V into this bucks cap, just to see what happens.
Additionally:
I already scoped the voltage across all of the vrm and memory phases inductors using a diff probe and all the signals look nice and exactly the same.
I hope anyone has some clues for me.
Cheers!
So i got this RTX2070 Super from Gigabyte for cheap and the reason i actually got it in first place is, because it initially had a full short on PCIE_12v, so my plan was to basically feed some juice in it on my test bench and see what gets hot using my thermal camera. This would have potentially been a quick repair, however things didn't actually work out quite as planned.
Whenever i hooked up one of my lab power supplies to my testboard to feed the 12v rail, well, even tho i had it current limited at 12v, the charged capacitors actually blew the 10A Fuse on the card right out.
Not a big deal, i was thinking, lets just short this fuse and try it again without actually having the power supply already set to 12v.
Unexpectedly enough, nothing unusual happened this time around, with 3v3 and 12v hooked up both to pcie as well as to the psu connectors of the card, the thing started right up and consumed its normal 1amp at 12v. So whenever the fuse blew, it unfortunately also cleared the short out, which is basically where we are now.
The card has fanspin, it has 5v, 3v3, 1008mV PEX, 1368mV Vmem, 812mV Vcore, so these are all there, but USB-C doesnt output anything, also the buck converter which is connected to the USB-C Port via a mosfet, isn't running. Added to that, the RGB at the card doesn't light up and, while the card has limited (800x600) output via hdmi, DP also doesn't output anything.
In Windows the card won't be initialized and obviously devmgmt reports error 43.
While i normally would look at this as a failed micron GDDR6 chip, since the card did actually have a full short initially and other things aren't working, i don't think that thats the issue with this particular one.
So as i was basically expecting this card to be a quick one, i haven't actually got the time to virtually reverse engineer the circuit since there are no schematics to be found for this one, so that i can figure out which issue we may have.
Anyways, that's why i‘m here now, does anyone actually have experience with this particular card (or 2070‘s im general, really) and can basically tell me what's normal and what's not? I think it is pretty suspicious that USB-C outputs nothing, but then again, it could be that the driver needs to be fully initialized in order to get usb-c going and enable its buck converter, if that one is even specifically just for usb-c. Does it even need a separate one aka does it even have PD? Same thing with the RGB, does that maybe also need the driver to be fully initialized?
Actually i wouldn't expect that for both, but who knows. I don't have anything to compare and unfortunately, i havent found any datasheet for the MP8858GQ which is the USB-C Buck IC, so i can't tell which one is the enable pin, or if it's actually broken.
I would suspect tho, that this is just another 5v rail, that powers usb-c, RGB and something GPU internal which makes the DP Signal. I might actually try feeding 5V into this bucks cap, just to see what happens.
Additionally:
I already scoped the voltage across all of the vrm and memory phases inductors using a diff probe and all the signals look nice and exactly the same.
I hope anyone has some clues for me.
Cheers!
Attached Files
Comment