Pictures of the GPU

What's with lazyness of "bad firmware, let's try replacing that and hope for the best" ? 🤣
This isn't Maxwell or Polaris generation that have unlocked vBIOS for users to do whatever they want to it you know ?

If it's not being detected (under Windows/Linux), you are screwed because full test setup will be required to MAYBE make it work again [a.k.a. hot plate + soldering station with reballing and other bits and bobs].

As for basic test procedure :
1) Check if major rails are not shorted to ground (3.3V, 5V, 12V, 1.8V, Vgpu, PLL, and other minor rails [different depending on GPU generation]), all should be marked on particular card's schematic.
2) Check if fuses/shunts are OK (if present),
3) If all of above is checked and OK, you can try connecting card to test setup and try to poke all voltages again (this time to see if they are present).
4) If they all are, and you still don't get the signal - your GPU is probably cooked (or something hidden like broken trace or bad component prevents card from posting). It is at this point you should give up.
Watch some pro GPU repair guys do their magic so that you can tell how much work this may actually require to diagnose, and no - your setup isn't even close to being enough for this [putting your know-how aside for now].

I didn't put "0 step", which is to visually confirm nothing blown up/smells like burning plastic, before doing anything (which should be obvious).

agent_x007 wrote on 2026-07-22, 19:26:

What's with lazyness of "bad firmware, let's try replacing that and hope for the best" ? LOL This isn't Maxwell or Polaris gene […]
Show full quote

What's with lazyness of "bad firmware, let's try replacing that and hope for the best" ? 🤣
This isn't Maxwell or Polaris generation that have unlocked vBIOS for users to do whatever they want to it you know ?

If it's not being detected (under Windows/Linux), you are screwed because full test setup will be required to MAYBE make it work again [a.k.a. hot plate + soldering station with reballing and other bits and bobs].

As for basic test procedure :
1) Check if major rails are not shorted to ground (3.3V, 5V, 12V, 1.8V, Vgpu, PLL, and other minor rails [different depending on GPU generation]), all should be marked on particular card's schematic.
2) Check if fuses/shunts are OK (if present),
3) If all of above is checked and OK, you can try connecting card to test setup and try to poke all voltages again (this time to see if they are present).
4) If they all are, and you still don't get the signal - your GPU is probably cooked (or something hidden like broken trace or bad component prevents card from posting). It is at this point you should give up.
Watch some pro GPU repair guys do their magic so that you can tell how much work this may actually require to diagnose, and no - your setup isn't even close to being enough for this [putting your know-how aside for now].

I didn't put "0 step", which is to visually confirm nothing blown up/smells like burning plastic, before doing anything (which should be obvious).

Just uploaded pictures

To your first point, I've bought parts (including a 1080ti) from this place before, so although I knew of the gamble, they normally test things before they sell them. With this GPU, they hadn't gotten to it yet, so I was hoping that worst case, it would just need new thermal paste or the fan was unplugged if it didn't work.

I purchased the BIOS EEPROM device out of curiosity and because the price was low enough that if that wasn't it, I didn't mind eating the cost. It wasn't my first thought to do that nor did I have any idea that the GPU wasn't working or else I wouldn't have purchased it. The deal and the fact that it was what it was (my newest GPU happens to be a 2060, so to me it was worth it) is why I purchased it.

To your second point, I'll try those things in the hope that one of those steps fixes it.

Look at u614 on the back. Almost looks like it has been worked on before.

Also check the voltage regulator area above U25 (near card slot). Can't tell if it is just dirt or if it looks like that voltage regulator has let out the magic smoke.

Also look at Q601 - near bottom left of second picture. Kind of looks like it has a piece blown out near the middle leg.

You don't have a dead short anywhere if the PC still powers on with it installed.

If not detected at all, then either the BIOS or something in the power circuit is possibly dead.

cyclone3d wrote on 2026-07-23, 01:42:

Look at u614 on the back. Almost looks like it has been worked on before. […]
Show full quote

Look at u614 on the back. Almost looks like it has been worked on before.

Also check the voltage regulator area above U25 (near card slot). Can't tell if it is just dirt or if it looks like that voltage regulator has let out the magic smoke.

Also look at Q601 - near bottom left of second picture. Kind of looks like it has a piece blown out near the middle leg.

You don't have a dead short anywhere if the PC still powers on with it installed.

If not detected at all, then either the BIOS or something in the power circuit is possibly dead.

For U25, because of the thermal pads, there's dirt caught on it.
For U614, I know that I had been trying to use the EEPROM clamps on that since that seems to be the BIOS chip.
For Q601 (or any of the others really), if that's the case, what can or should be done or is nothing the answer?

My computer still powers on with it installed. If I have another GPU, it will post and get to the OS desktop. If it's by itself, it won't post due to not detecting the GPU.

How would I check the Power Circuit and would that be where I plug in the 6/8-pin connectors?