Re: Repeated TDR crashes due to unknown fault

[ogkspaz]

Hi there.

I have had a new PC since September 2016 and somewhere during the last year it has started giving me issues when playing games or sometimes even watching a Youtube video. I have taken it in for RMA as my friend and I was suspecting the PSU to finally be the problem. I received a mail back from the supplier saying the PC is ready to collect with: "No fault"

The original mail that I sent them with all the steps I have tried below.

One of the PC's components seems faulty or does not like something because I am having repeated TDR crashes. I now suspect it might be a faulty PSU but I cannot test it myself.

I have tried all kinds of solutions to exhaustion. Sorry for the long list to follow, but I need to let the technician know what I have tried thus far and the results I have gotten so that you have an idea what we're dealing with.

PC Specs:
Asus Maximus 8 Hero motherboard
Intel Skylake i5-6600 quad 3.3
Corsair 16gb RAM
WD 3Tb HDD
Corsair 9020103 650w CX modular PSU
EVGA Geforce 760GTX SC 2gb
Samsung SSD 850 250gb

The crashes occur mostly when I am playing games or watching videos on chrome, YouTube mostly. It sometimes happens within 15 minutes of starting up, other times it doesn't crash for hours on end. It is not limited to specific games and so far I have encountered it in all of the following:

Diablo II Lord of Destruction
Diablo III
Grand Theft Auto 5
Nier Automata
Tom Clancy's the Division
Tomb Raider 2013
Trackmania Untited Forever
Trackmania2 Stadium
Warframe

I have tried:

-o- New drivers, obvious one this. Uninstalled the old ones with display driver uninstaller, restarted, installed, restarted, played game, CRASH.

-o- Tried old GPU, Geforce 250GTS, NO CRASH, maybe doesn't draw enough power?

-o- Tried Geforce 760GTX in old PC (Phenom XII 550 BE, Crosshair III formula, 4gb RAM, 500 watt PSU, NO CRASH (Before I upgraded to new PC, GF760 was in this PC for maybe a year and I never had a crash)

-o- Maybe GPU does not like one screen on Analog and the other on digital, Connected only 1 screen, CRASH.

-o- Took old PC's Hard drive, moved it to new PC, formatted, did a clean install of ONLY Windows Newest GFX Drivers and Diablo II / Warframe and ran it on one screen only, CRASH.

So from what I can figure out:
-Not the GPU (works fine on old PC with 500watt PSUm 150 less than current PSU)
-Not software (Clean windows and driver install also crashes)
-Not the SSD if that counts (Clean install test used a different HDD and SSD was not connected)
-Motherboard, RAM and CPU checked out fine during November RMA.

There are also a ton of tips online which I have tried. TDRdelay etc registry keys, still crashes.
Switching off advanced 3D rendering in chrome, still crashes.
Changing "Power Management Mode" in Nvidia's control panel to "Prefer Maximum Performance", still crashes.

The crashes themselves come in two different forms.
1. Both screens go black then a popup appears saying: "Display Driver Stopped Responding and Has Successfully Recovered"or

2:The game freezes for 5-10 seconds then moves for maybe half a second then freezes again, rinse and repeat till either both screens go black and the above popup comes up or I give up and hit reset instead

When these crashes happen I can hear the GPU fan stop spinning and then resume when the screens go back on.

Event viewer has a TON of logs of the crashes mentioning nvlddmkm.

They have not tested it with my OS drive yet. I'm running Windows 7 Ultimate so now wondering if that version does not like the hardware. I've also wondered if it could be the wall socket I am running the PC from as our building has had a few tenants power trip every now and then. I ran the old 500PSU PC off the same socket though with no problems.

So TLDR: I have taken a PC in to a supplier with a definite fault which they are unable to replicate on their side (PC is currently still there).

What can I do from here to try and fix the problem or who can I approach in the Cape Town area that I can rely on to actually find the cause of the crashes rather than just "no fault"

Thanks

 

[envo]

"I have had a new PC since September 2016"

It's 2018, it's not new anymore.

0) Why the hell would the supplier not test with the OS?
1) Reinstall the OS if you feel it's the OS
2) If that doesn't work, try getting a UPS. Brown-outs often cause issues with PC's and the old PSU may have been a little more less sensitive to fluctuations than your "new" PSU would be. Brown-outs is even worse especially if your PSU is under load, like playing games.
3) Is it in a new case as well? Did you check the thermals of your GPU? Did you even try to maybe test with a friend's GPU to see if it's not a GPU problem? GPU's tend to overheat and cause issues like that. I had a new case and my GPU temp at load was 90 degrees (fans were spinning etc.) Turns out the airflow in my new case didn't gel so well with how the GPU was expelling the hot air.

Your old GPU perhaps expels air differently to the current GPU (there's a difference sometimes, even on the same type/model) or didn't get as hot under load, never crashing.

Doubt it's your PSU or other components. Hopefully some of these tips help

 

[ogkspaz]

0) This is what I was thinking. I have requested yesterday that they test it with my drive and have supplied them with my password but have yet to hear something back
1) I have tried a fresh OS reinstall on a different drive with only a game and the GPU drivers and it still crashed.
2) I may have to go the UPS route so still need to find out pricing and which to go for.
3) It is in a new case (Stryker Stormtrooperwith a ton of fans and outlets) as the old one was way too constricted and in need of replacing anyway. GPU temperature was fine but I can't give you numbers from memory. New PC runs a whole lot cooler than the old one I can tell you. I have tested with the only other GPU I have available which was the 250GTS, no crash, possibly due to it not pulling enough power.

Thanks though.

 

[WAslayer]

Hi there.

I have had a new PC since September 2016 and somewhere during the last year it has started giving me issues when playing games or sometimes even watching a Youtube video. I have taken it in for RMA as my friend and I was suspecting the PSU to finally be the problem. I received a mail back from the supplier saying the PC is ready to collect with: "No fault"

The original mail that I sent them with all the steps I have tried below.

They have not tested it with my OS drive yet. I'm running Windows 7 Ultimate so now wondering if that version does not like the hardware. I've also wondered if it could be the wall socket I am running the PC from as our building has had a few tenants power trip every now and then. I ran the old 500PSU PC off the same socket though with no problems.

So TLDR: I have taken a PC in to a supplier with a definite fault which they are unable to replicate on their side (PC is currently still there).

What can I do from here to try and fix the problem or who can I approach in the Cape Town area that I can rely on to actually find the cause of the crashes rather than just "no fault"

Thanks

Question regarding this: -o- Tried Geforce 760GTX in old PC (Phenom XII 550 BE, Crosshair III formula, 4gb RAM, 500 watt PSU, NO CRASH (Before I upgraded to new PC, GF760 was in this PC for maybe a year and I never had a crash)

Is that testing with a different OS drive to the one you are having issues with..? If so, what driver version does this OS use and have you tried that version on the problem OS..?

 

[ogkspaz]

Question regarding this: (SNIP)

Is that testing with a different OS drive to the one you are having issues with..? If so, what driver version does this OS use and have you tried that version on the problem OS..?

Same OS: WIndows 7 Ultimate
Different Drive: Samsung 250 Gb SSD vs 320 Gb SATA either Seagate Or Western Digital.

Before testing I uninstalled the current driver with Display Driver Uninstaller and installed the newest (same) driver for 760 on 64bit Win7 .

EDIT: Just called them and asked that they retest with my drives and also check the event viewer's log, so let's see what happens.

 

[ogkspaz]

Okay so an update.
My wife fetched the PC yesterday while I was at work. They did not charge me a workshop fee as they could not find the fault. It behaved itself last night in Division and Warframe but had its first crash again in Borderlands this morning (the long freeze type). I just checked and there Eventviewer has no log between 17 Jan @ 15h39 when they probably switched it on (I forgot to give them my password initially, derp) and 26 Jan @ 19h01 when I switched it on again which shows me that they did not test it with my drives attached as I asked on the 24th.

This still leaves me with the original question:
Does anyone know who I can approach in the Cape Town area that I can rely on to actually find the cause of the crashes?

 

[Jladan]

So I’ve also had a few odd TDR issues over my time with various boxes. Some things that have worked for me beyond the usual suggestions.

1. Make 100% sure GPU is seared properly, especially with bigger and heavier cards they sometimes move ever so slightly and cause crashes

2. Turn nvidia audio components that come with the nvidia drivers off. U don’t need them usually and this has helped in a few cases

3. Lastly if you have any RBG or light software remove it and see if that helps. I’ve had Corsair and Aorus RBG apps causing TDR’s

 

[ogkspaz]

I've reseated the GPU so many times for testing, it is definitely firmly in.
I disabled all the nvidia audio bits and deleted AURA for the LEDs.

Just crashed again in World of Warships.
Just about ready to throw the pc off the roof and pick up a different hobby. Still got a few model kits that need building from years ago.