Question Intermittent operation of a used GPU ?

[El_Caminero]

Hi all

First post, hope to be able to repay the favour.

So, I bought a used/faulty Zotac GTX 1050 Ti OC where the owner stated it was working intermittently.

I thought I'd have a go at trying to fix it (and in turn upgrade my standard 1050 ti), but more than a month later I'm still scratching my head.

The card will run fine one day, will run for hours with zero issues, even allowing me to restart the PC it's attached to, but the moment I SHUTDOWN the PC, it will not boot up/POST/fan spin/generate any heat or signs of life whatsover....until I try starting the PC up a few days later (as long as it was disconnected from the MOBO).

What I've tried so far:

• Visually inspected for issues (nothing obvious, missing, broken, burnt)
• Pulled card apart completely, deep cleaned and reapplied thermal paste
• Tried on different PCs (my current gaming PC and a test PC), no difference (both PC PSUs with plenty of wattage)
• Tried changing monitor cables (uses HDMI), no luck
• Tried different monitors (1080s), no luck
• Updated to latest drivers on both PCs (both running latest MOBO BIOS too), no luck

I'm waiting a day or so to see if it'll start, so I can update the GPU BIOS?

I suspect it could be a power issue though because up until the PC shuts down, it runs perfectly (PC detects no issues with it using DxDiag and temps, voltages, etc all seem fine).

Any ideas would be appreciated.

(I've refrained from posting system specs as I've tried it on two different set ups, and I think I can safely assume the issue is the card itself, but I will post specs of both set ups if you really think it could help eliminate everything - let me know).

 

[Lutfij]

Welcome to the forums, newcomer!

What does GPU-Z show you? Do you see an error code 43 on Device Manager?

Please list the specs to your build like so:
CPU:
CPU cooler:
Motherboard:
Ram:
SSD/HDD:
GPU:
PSU:
Chassis:
OS:
Monitor:
include the age of the PSU apart from it's make and model. BIOS version for your motherboard at this moment of time.

Since you used another PC as well, we will need the specs as well, as requested above to make sure if anything might be out of place or if there's a common denominator, outside of the obvious which was not to buy an used GPU that was faulty to begin with.

 

[El_Caminero]

Welcome to the forums, newcomer!

What does GPU-Z show you? Do you see an error code 43 on Device Manager?

Please list the specs to your build like so:
CPU:
CPU cooler:
Motherboard:
Ram:
SSD/HDD:
GPU:
PSU:
Chassis:
OS:
Monitor:
include the age of the PSU apart from it's make and model. BIOS version for your motherboard at this moment of time.

Since you used another PC as well, we will need the specs as well, as requested above to make sure if anything might be out of place or if there's a common denominator, outside of the obvious which was not to buy an used GPU that was faulty to begin with.

Thanks for the reply. Purchasing faulty gaming PC components is a hobby of mine, and only spent AUS$30 so if the gamble pays off I get a cheap upgrade; if not I may learn a thing or two from you good folk

Okay, here goes:

Faulty GPU: Zotac GTX 1050 ti OC edition

Test PC 1:
CPU: Intel Core i7 4770, 3.4Ghz
Cooler: Antec Kuhler H2O 620r AIO
MOBO: Asus Gryphon Z87, BIOS ver. 2103
RAM: Corsair Vengeance Pro Series 16GB DDR3 1600MHz
STORAGE: Corsair Force LS 120GB SATA 3 SSD and Seagate Barracuda 7200.12 1TB HDD
GPU: Asus GeForce GTX 1050 ti
PSU: Be Quiet! Pure Power BQ L8-600W (not sure of age)
CASE: NZXT CA-H442W-W1 Case
OS: Windows 10 Home 22H2
Monitor: LG IPS FullHD 27" monitor

Test PC 2 (don't laugh):
CPU: Intel Core i5 750
Cooler: Coolermaster CPU heat sink/fan combo
MOBO: Gigabyte P55A-UD3, BIOS ver. F10
RAM: Corsair CMY16GX 16GB DDR3 666MHz
STORAGE: WD 2TB SATA HDD
GPU: AMD Radeon HD 5450
PSU: Some unknown brand PSU, 475W (2010 model but does the job)
CASE: No case (until tomorrow, picking one up for $25 lol)
OS: Windows 10 Home 22H2
Monitor: same monitor as above

 

[El_Caminero]

Welcome to the forums, newcomer!

What does GPU-Z show you? Do you see an error code 43 on Device Manager?

Please list the specs to your build like so:
CPU:
CPU cooler:
Motherboard:
Ram:
SSD/HDD:
GPU:
PSU:
Chassis:
OS:
Monitor:
include the age of the PSU apart from it's make and model. BIOS version for your motherboard at this moment of time.

Since you used another PC as well, we will need the specs as well, as requested above to make sure if anything might be out of place or if there's a common denominator, outside of the obvious which was not to buy an used GPU that was faulty to begin with.

Sorry, just realised you'd asked what GPU-Z reported - nothing stands out at all, temps low and constant after running for about an hour, board power draw: 35.5W, PCIe slot power: 31.7W, PCIe slot Volts: 9.5v.

Also just updated the GPU BIOS to 86.07.42.00.72, driver is ver. 31.0.15.3623 (Jun 08 2023, believe it's the latest).

Device Manager same, no errors or anything out of the ordinary.

Have this GPU installed in a new case, where my testing components are finally housed - will post this update, then restart and see if any changes. Then will shutdown and see if the BIOS update helped or not, and report back.

 

[geofelt]

It could be a thermal issue.
When running, components heat up and expand.
That may make or break s small connection.
Probably a very difficult problem to identify.

 

[El_Caminero]

Okay, the GPU BIOS update made no difference, restarted = fine, shutdown = fails to post. Any ideas?

 

[El_Caminero]

It could be a thermal issue.
When running, components heat up and expand.
That may make or break s small connection.
Probably a very difficult problem to identify.

True but when this card boots, it runs for over an hour with no dramas. It's the moment power is cut from it that it doesn't want to start up again (until MUCH later).

I might stress test it and see if it fails too, just to rule this out - thanks for the suggestion.

 

[El_Caminero]

Sorry for the bump guys, but just a quick update - I booted it up last week and it posted fine, ran GPU-Z and all seems normal so I stress test it for about an hour using Furmark, and no probs, temps peaked at about 71c. Restarted and all good too. Shutdown and rebooted, no good.

If it is an issue with the card requiring a VERY long cooldown, is this something fixable?