[SOLVED] 1080 Ti fried but PC running fine

[prateek320i]

Hi all.
3 months ago, I bought a new case Thermaltake Core P5 (upgraded from NZXT S340 Elite) and installed my GPU vertically using a Riser Cable.
PC ran fine and played some games. The CPU wasn’t overclocked, but I was running GPU overclocked with stock software (Aorus Engine), since 2018.
Next day in the morning I powered up my PC. It started with a smoke and spark coming out from the GPU. I rushed to the MAINS and turned off everything. Later I opened and inspected that one of the VRM has burnt and I could clearly see copper on the PCB (melted the insulation).
Unfortunately my card was out of warranty, I couldn’t RMA it. All I did is checked my PSU with multimeter.
I checked every female port for voltages and found all the voltage levels (+12VDC, -12VDC, +5VDC, +3.3VDC, +5VSB) excellent and within tolerance.
Thus I decided to run the PC without GPU, trusting the PSU with eyes closed. It booted and ran fine.
I am now using this PC since 3 months and I didn’t face any issue like rebooting or sudden shut down.

Now, I’ve bought and installed a new ASUS TUF GAMING RTX 3070 (3060 Ti is out of stock in India). And attached the card directly to the board (Horizontal mount) and PC booted fine with working graphics card. I installed the driver and shut down the PC after that.

But I’m in fear about my card. Will it work properly without burning? Honestly, I’m afraid of my new GPU.
I want to ask the community Is my PSU okay? Should I keep using the PC with RTX 3070 installed on it?

Key components-
Asus ROG Strix Z370E
i7 8700k
Gigabyte GeForce GTX 1080 Ti Gaming OC 11GB
Corsair HX750
Thermaltake Core P5 (using included Riser cable)
GSkill Trident 3200 MHz
Deepcool castle 360 rgb v2

 

[Dj0gany]

Likely your vertical gpu mount pcie cable/slot was bad.
If your psu was bad, you wouldnt get a post after that or something would've happened alongside of frying the gpu.

 

[prateek320i]

The new 3070, which I have now, is mounted on the same PCIe slot where I mounted the GPU RISER CABLE, earlier.
The new GPU worked fine (mounted directly on the PCIe Slot without any riser cable) for the time I used the PC. Now after reading your answer, I’ve started having doubts on the riser cable.

 

[Dj0gany]

Exactly.
I would get rid of that riser cable if i were you.

 

[prateek320i]

The cable came with the case. I would have sued Thermaltake if I was MP or MLA in India.
Should I RMA my case with Thermaltake?

 

[Eximo]

The cable came with the case. I would have sued Thermaltake if I was MP or MLA in India.
Should I RMA my case with Thermaltake?

No, you can just buy a new riser cable. You should do that anyway if you wanted to use one. RTX 3070 supports PCIe 4.0, your CPU doesn't, but having the higher quality riser cable will be handy in the future, and should let the RTX3070 operate with less errors.

 

[InvalidError]

Blown VRM may just have been an unfortunate material or manufacturing defect that took a while to happen through no fault of anything else in the system. If it is a 12V phase near the PCIe AUX connectors, then it is extremely unlikely the PCIe riser had anything to do with it since the blown phase likely gets its power through the AUX cables.

 

[prateek320i]

Blown VRM may just have been an unfortunate material or manufacturing defect that took a while to happen through no fault of anything else in the system. If it is a 12V phase near the PCIe AUX connectors, then it is extremely unlikely the PCIe riser had anything to do with it since the blown phase likely gets its power through the AUX cables.

That’s what I was thinking since the death of card.
VRM gets voltages from PSU. And my HX750 has been running flawlessly for 3 months, and voltage levels are perfect (checked regularly), no oddity seen. My heart can not even try to blame HX750 for this, then what on earth could possibly fry the GPU? This question is mystery for me.

Edit- show me some light

 

[InvalidError]

what on earth could possibly fry the GPU? This question is mystery for me.

Edit- show me some light

One of my friends had a $2000 digital piano with one partially dead key (no sustain) and when I repaired it, the issue was simply a 1N914 diode in the key sensing array that had gone open-circuit for whatever reason. The key sensing operates on mA-scale current, so overload definitely wasn't the cause. Must have been a manufacturing defect such as a micro-crack in the glass diode body cracking the junction over time with vibrations from use.

Very small defects can kill stuff over time. That's why devices that pass QC fail at various points throughout their useful life with the peak failure rate often happening in the few years after the warranty has expired.

 

[prateek320i]

One of my friends had a $2000 digital piano with one partially dead key (no sustain) and when I repaired it, the issue was simply a 1N914 diode in the key sensing array that had gone open-circuit for whatever reason. The key sensing operates on mA-scale current, so overload definitely wasn't the cause. Must have been a manufacturing defect such as a micro-crack in the glass diode body cracking the junction over time with vibrations from use.

Very small defects can kill stuff over time. That's why devices that pass QC fail at various points throughout their useful life with the peak failure rate often happening in the few years after the warranty has expired.

This means I can trust my Corsair HX750 with my new 3070.
In addition to this, I’ll try to visit Thermaltake regarding PCIe Riser cable, and will ask them to verify the same.
For now, I’m planning to buy a closed case (Corsair 4000D Airflow) and try to sell my idea of mounting my GPU vertically 🙁