GTX 550 Ti fails at 52°C

Heku

Heku

Member
Hey Guys!
so lately my Club 3D GTX 550 Ti has started failing while Gaming and DogeCoin mining (doing for fun).
The gaming problems happened early on (PlanetSide 2, since march if i remember right), though it was because of drivers or something, and started playing other games (didn't crash).
wasn't really following the temps back then, but since i started mining, in the beginning everything went fine, but since last week my gpu has started failing several times, in both gaming and mining..
So last days i've been following my gpu temps and started wondering why it keeps crashing, and i've come to a conclusion that everytime it's about to hit 52°C it has failed and freezed my computer (had to press the reset button to get the pc back up and going).
or well, today my computer suddenly crashed when i was playing GRID 2, and though about following the temps and play again, after a while of being at 51°C it crashed (i guess just when it was about to hit 52°C it failed).
then i tried mining, and again after a while being at 51°C the computer failed (-||-).
Warranty time is passed, and i've been thinking about getting a new graphics card anyway.. but i'm guessing i need to get it soon.

TL;DR: my computer freezes and i need to reset it when gpu temp gets to 52°C.
Has this happened to anyone else before?
 
That's odd, what is your idle temp? What drivers are you using? Have you checked how much dust is trapped in the heat sink? Have you checked your voltages for stability?
 
Can you tell what your power supply is putting out? If not install something like HWiNFO64 and check your sensors and see what your 3.3/5.0/ and 12V rails are running at.
 
Yeah it sounds like a power supply issue, I had a PSU that was on the way out doing the exact same thing as yours, I was pulling my hair out trying to figure it out.
 
JLuchinski said:
That's odd, what is your idle temp? What drivers are you using? Have you checked how much dust is trapped in the heat sink? Have you checked your voltages for stability?
Click to expand...

idle temp is 36C (at fan 100%) or ~39C (at fan AUTO)
i'm not sure what you mean by drivers... but gpu has the 337.88 [GeForce Experience] (HWiNFO64 says Driver Version: 9.18.13.3788)
i've cleaned out the heatsink (gpu) a couple weeks ago (for like the 3rd time since i got the card)
haven't touched the voltages ever so they should be ok i guess...


StrangleHold, downloaded and installed HWiNFO64, but i'm not totally sure how to read the numbers.....
Nuvoton NCT6771F:
+3.3V | Cur: 3.264 V | Min: 3.248 V | Max: 3.264
+5V | Cur: 5.040 V (Min/Max same)
+12V | Cur: 11.880 V | Min: 11.880 V | Max 11.933 V
[didn't find in the Sensor Status window any else with voltages..]
uh oh, and GPU Core Voltage: 1.075 V (Cur/Min/Max)

I hope i atleast answered some of your questions right..!
 
How do you know its the GPU? Im thinking its the CPU, either overclock or heat.

REMOVE ALL OVERCLOCKS - all of them, CPU, GPU, RAM and so on.!
What have you used to measure temps?
WHat is your ambient temperature?
12V rail is within spec, however I would use a multimeter to check under load.
Clean install your display driver using Add and REmove Programs.
Restart
Download and run Display Driver Uninstaller for nvidia
Restart
Resinstall latest nvidia driver.
 
Do you get green dashes across the screen when it freezes. I have an old GTX 580 (pre DX11) I MUST run with the voltage OC'd on later games or it will crash. I use Afterburner to over-clock to the first increment - a tiny amount but enough, never crashes.

Athough I didn't have to reboot the computer just the game.
 
Okedokey said:
How do you know its the GPU? Im thinking its the CPU, either overclock or heat.

REMOVE ALL OVERCLOCKS - all of them, CPU, GPU, RAM and so on.!
What have you used to measure temps?
WHat is your ambient temperature?
12V rail is within spec, however I would use a multimeter to check under load.
Clean install your display driver using Add and REmove Programs.
Restart
Download and run Display Driver Uninstaller for nvidia
Restart
Resinstall latest nvidia driver.
Click to expand...

Removed CPU OC (was at 3.4GHz)
i use CoreTemp on CPU and GPU Meter (Desktop Gadget) / MSI Afterburner for GPU.
Ambient temp is 25-27C..
i don't have a multimeter.. but i'll see what i can do about it..
EDIT: did the driver installation as you said, but still crashes at 52C.. could the problem then be in the PSU?

melledog, nope, screens just go black and there is a weird sound for a sec, and then No Signal..
 
Last edited:
mellodge said:
Do you get green dashes across the screen when it freezes. I have an old GTX 580 (pre DX11) I MUST run with the voltage OC'd on later games or it will crash. I use Afterburner to over-clock to the first increment - a tiny amount but enough, never crashes.

Athough I didn't have to reboot the computer just the game.
Click to expand...


GXT580 IS Dx11, and IS a very different card to the 550ti. Most vendors reduced the voltage on the core to the absolute minimum to make it work, and some games required a slight bump say 100mV. This 550ti, however is not that card...
 
My CPU is around 50C in idle, i don't think it should be that high.. running stock cooler and haven't changed thermal paste ever.. maybe i should change thermal paste (or buy a better cooler, so i can OC) and clean the heatsink from dust that i can't get off when cooler on place..?
 
Lets cover all bases. GPU and CPU.

This is your card right?

Club-3D-GeForce-GTX-550TI-1536MB.jpg


Ambient temperature of 27oC seems high too given its currently 16oC there, you run heating at 27oC wholly molly!!!

Lets start from the beginning if you wish?

Do this (in order - no short cuts):

  1. Download PC Wizard (from my sig), run it and go to FILE, SAVE AS and click OK. Copy the text out of that file into this thread.
  2. Update Direct X
  3. Click on Windows button + R
  4. Type cmd, right click on cmd and run as Administrator
  5. In the command prompt, type (with spaces) sfc /scannow
  6. Let it complete, then
  7. Update Windows (do it) - let it finish. Completely update windows including any restarts before progressing.
  8. Shutdown.
  9. Remove CPU cooler
  10. Remove CPU.
  11. Clean both with isopropyl alcohol
  12. Clean heatsink from dust.
  13. Replace CPU and replace thermal paste with Arctic Cooler 5 using this methodology .
  14. Replace heatsink ensuring you plug fan into motherboard.
  15. Restart computer.
  16. Download this and install it
  17. Restart.
  18. Download nVidia driver to your desktop, do not install it yet.
  19. Download Display Driver Uninstaller - do not run it yet.
  20. Uninstall all nvidia software in your Add and Remove Programs (within Windows 7).
  21. Restart.
  22. Run Display Driver Uninstaller for nvidia (as an admin) and check the boxes shown in this image.
  23. It should automatically restart. You now have a completely clean display driver stack.
  24. Restart if it doesn't do it automatically.
  25. Install nvidia driver (default install) downloaded at 11.
  26. Restart
  27. Navigate to http://www.evga.com/precision/
  28. Press the Download Button
  29. Create a new registration if you don't already have one, then...
  30. Download EVGA Precision and install EVGA_PrecisionX_Setup_410.exe
  31. Run EVGA_PrecisionX.
  32. Refer to Figure 1 below. Click on the the 'cog' i have put a green square around.
  33. Refer to Figure 2 below. Check the boxes i have highlighted.
  34. Refer to Figure 3 below. Check the 'Enable software automatic fan control' check box that i have highlighted.
  35. You should see a a drop down graph like the one in Figure 4. Notice the squares on the line graph (that i have highlighted yellow). You can use your mouse to modify these).
  36. Refer to Figure 5. Move each square so that at 50oC it hits maximum (100%) fan - as shown in Figure 5.
  37. Minimise (don't close) EVGA Precision.
  38. Try to game and see how you go.


    Figure 1: EVGA Precision Initial GUI

    Imate1_new.png~original



    Figure 2: General Tab, Check the Start with Windows and Start Minimised

    Image2_new.png~original



    Figure 3: Enable Software Automatic Fan Control to view/modify fan curve vs temp profile

    Image3_new.png~original



    Figure 4: Shows default fan curve with control points (highlighted yellow)

    image4_new.png~original



    Figure 5: Desired fan curve for removing temperature as an issue (for now).

    image5_new.png~original

 
Last edited:
did everything as explained above (except CPU removal, don't have any thermal paste atm).. but still crashed at 52oC!
i'm starting to lose hope with this problem..
i'm hoping this is a GPU problem and not a PSU problem, because i'm anyways gonna upgrade my GPU soon..
 
Already forgot the PC Wizard report.. anyways:

PC Wizard 2013 Version 2.12
------------------------------------------------------------------------------------------

Owner: Heku
User: Heku
Computer Name: HEKU-PC
Operating System: Windows 7 Ultimate Professional Media Center 6.01.7601 Service Pack 1
Report Date: torstai 05 kesäkuu 2014 at 16:09

------------------------------------------------------------------------------------------


<<< System Summary >>>

> Mainboard : ASRock 880G Pro3

> Chipset : AMD 880G

> Processor : AMD Phenom II X4 955e @ 3200 MHz

> Physical Memory : 8192 MB DDR3-SDRAM

> Video Card : NVIDIA GeForce GTX 550 Ti

> Hard Disk : Western Digital WD5003AZEX-00K1GA0 ATA Device (500GB)

> Hard Disk : Seagate ST320005 42AS (2000GB)

> CD-Rom Drive : DTSOFT Virtual CdRom Device

> DVD-Rom Drive : Toshiba-Samsung DVD-ROM SH-D163C

> Monitor Type : Dell Computer DELL 1907FPV - 19 inches

> Monitor Type : Samsung T24B300 - 23 inches

> Network Card : Attansic (Now owned by Atheros) AR8151 PCIe Gigabit Ethernet Controller

> Operating System : Windows 7 Ultimate Professional Media Center 6.01.7601 Service Pack 1 (64-bit)

> DirectX : Version 11.00

> Windows Performance Index : 5.9 on 7.9
Help : Example of usage


------------------------------------------------------------------------------------------
***** End of report *****

i'll check the BIOS settings on sunday or monday, leaving for the weekend soon.. but haven't touched much in BIOS, except the UCC thingy for overclocking...
 
Im beginning to think your CPU is overheating when gaming.

Complete steps 9 - 14 above exactly.

If that doesnt solve the issue (by reducing CPU temps), then you are left with a few options.

Download Memtest here, burn it to a DVD/CD, boot from it and let it run overnight.

If not issues, then get a multimeter and test your 12V rail on a molex connector while the computer is gaming or under load.

If that doesn't show any issues, then well, time for a new machine imo.
 
alright, thanks for the help.
I'm going to the army in 1 month and i'll be there for a long time (6/12 months) i think i'm gonna upgrade my system after that (when i have enough money). i already have some ideas, but there will probably be better/new components available then, so maybe i shouldn't think about that now..
 
Last edited:
I'm back!
i tried changing PSU (SolyTech 500W) but that didn't solve it, now it's either got to be the graphics card, or the cpu..
i might just as well get a decent graphics card for cheap, and get some new TIM for the cpu (maybe new cooler).. and when i'm back from the army i should update whole rig...
 
No, that PSU is garbage and the COrsair is much better quality. The Solytech 500W has 33A (peak) on the 12V rail, the Corsair has 46A (continuous). So unfortunately mate, you have (if anything) made things worse, unless of course the Corsair is faulty, which, clearly it isn't (becuase it swapping it out, didnt solve anythign).

Download Memtest, burn it to a CD, run it and test your memory.

Download PC Wizard from my sig install it and tell me what your idle GPU and CPU temps are.
 
You must log in or register to reply here.
Back
Top