m17 R3 BSOD - Display Device Reset?

Hi All,

I have the GTX 580m GPU in my m17 R3 that has been throtteling some since I got it (very subtle) but now I have a ticket into Dell to send out a tech to replace the fan/heat sinks within the next 4-5 days because things are getting worse.

Its getting so bad that my machine will BSOD with the message "display device failed to reset" or somthing like that. I know, sorry for the lack of detail on that but at work now and will edit/follow up with better info when I get home.

The machine resets when that happens, which is around every 45min of gameplay (so frustrating). My teamspeak after the BSOD reset sounds like crap too so its getting unplayable at this point.

All drivers are updated per the dell driver update monitor. So I wanted to see if the BSOD symptoms will be cleared up with the new fan/heat sinks or is it a software issue?

Unfortunatley I am out of warrenty as I bought the machine Dec 2011, but Alienware is being nice enough to send out a tech for free anyways. Having already read alot of threads here about the 580 issues, I am nervous haveing let it go for so long I have messed up more then just the GPU, or its just a driver issue and I need to do a serious restore/install.

Thoughts? Thanks to all you awesome peeps in advance!

 

Generally that BSOD message means the driver crashed and failed to restart. If you're using the most current nvidia driver that's probaly your problem. It's unstable.

 

I see. I have rolled back the driver with still no improvement. How far back do I need to go, and/ or does someone have the best driver to use for the card to be most stable?

 

On Fri 7/26/2013 1:15:17 AM GMT your computer crashed
crash dump file: C:\Windows\Minidump\072513-23400-01.dmp
This was probably caused by the following module: nvkflt.sys (nvkflt+0x31704)
Bugcheck code: 0x116 (0xFFFFFA8007C114E0, 0xFFFFF88004031704, 0xFFFFFFFFC000009A, 0x4)
Error: VIDEO_TDR_ERROR
file path: C:\Windows\system32\drivers\nvkflt.sys
product: NVIDIA Windows Kernel Mode Driver, Version 310.90
company: NVIDIA Corporation
description: NVIDIA Windows Kernel Mode Driver, Version 310.90
Bug check description: This indicates that an attempt to reset the display driver and recover from a timeout failed.
A third party driver was identified as the probable root cause of this system error. It is suggested you look for an update for the following driver: nvkflt.sys (NVIDIA Windows Kernel Mode Driver, Version 310.90 , NVIDIA Corporation).
Google query: NVIDIA Corporation VIDEO_TDR_ERROR

 

tech came out and replaced the heat sink and re pasted the card. He cleaned out alot of dust and its really clean....still getting crashes.

It has to be the driver at this point. This is insane that a driver from the manuf site can do this. What a mess?

So I need to know what driver to use at this point?

 

Called Alienware, they are sending a tech down to install a new 580m in the machine. He said this new card has an updated bios and does not have the sme issues of th old 580m.

I told him I was scared of the same card and asked if I could get another. He offered an ati card of similar performance but no 3d. I wanted to get that one but strongly recommended the 580. I agreed against my will. We will see if this works. But it's really cool to se them doing all this for a machine out of warranty.

 

I'm actually have the exact same problem as you. Same BSOD with the same video card. Was the 580m back ordered? I have to wait for mine, they're going to replace it for me. I'm worried as well, I've seen others with the same issue and it turned out to be the motherboard.

 

Unfortunately many people that have tried the recent NV drivers have been having a LOT of these issues with older cards. Even the latest ones are having it to some extent.

The real shame is that many are finding that rolling back is not helping as the cards are permanently damaged. It is overvolting older cards and all sorts of issues are arising. NV refuse to even acknowledge the issue exist - as they would since they are liable for all the replacements that are happening (many!)

See my thread: http://forum.notebookreview.com/alienware/721878-nvidia-320-18-driver-warning.html

Damn you nvidia

 

These video cards use lead-free solder on the processor, and the resetting problem is a classic symptom of cracks developing in that solder due to the expansion/contraction of heating then cooling. The older leaded solder would flex during expansion/contraction, but the lead-free solder instead accumulates stress and eventually develops cracks.

When that happens, the video card begins failing to respond to commands, the commands time out, then the system tries to reset the card.

Like MickyD1234 says, make sure you don't use a damaging video driver in the first place. Also, run on a cooler. All my 17" laptops are on coolers, the Cooler Master NotePal U3 Laptop Cooling Pad with Three Configurable 80mm Fans (R9-NBC-8PCK-GP). Center a cooler fan beneath each of the two laptop fans, then have the blades of the third fan blow into the vent beneath the video RAM.

If you're having the resets and you're under warranty, call Dell to have the video card replaced. The problem will only continue to get worse over time, until the video card finally dies. The GPU requires much more power, and therefore generates much more heat, than the CPU. So it's much more prone to this problem.

Some people have brought dead video cards back to life by doing a reflow on the card. But that's only a temporary fix, and often damages or shortens the life of components on the card.

Kudos to MickyD1234 for bringing the driver problem to everyone's attention in the first place.

 

So when the tech comes and puts in a new (refurb?), card in there won't I have the same problem? What the hell can I do at this point sine it's just going to keep happening?

 

In general, I advise users to keep a machine under warranty as long as they plan to rely on it. And run laptops with higher-end GPUs on a cooler. And avoid the faulty video drivers that MickeyD1234 describes.

Other than that, it's just a fact that we have to live with, given the requirement for lead-free solder to meet ROHS compliance. ROHS was adopted for environmental reasons.

Lead-free solder is also a headache for spacecraft designers because, over time, it causes crystals of tin to form (called "tin whiskers") which expand from the solder joints, causing short-circuits and thus circuit failure.

Fortunately, with the M17x, at least you can drop in a variety of new video cards if you need to so, on the bright side, it's as an opportunity to upgrade. Lots of laptops from other manufacturers don't allow that option, with the GPU either soldered directly onto the motherboard, or requiring specific video cards that are almost impossible to find.

BTW, as soon as the technician completes the repair, ask him/her to stick around long enough for you to fire up a game for a minute or two, to make sure that nothing overheats. That helps ensure that the heatsinks are properly pasted and mounted. Also make sure you don't see any tearing or color cast in the display (and tilt the display forward toward you and backward away from you while looking for these symptoms), which can indicate damage was done to the display cable during the repair. And make sure that the keyboard is working properly before the tech leaves, because technicians can sometimes damage a keyboard cable or connector.

Good luck!

 

Thx for the tips and best wishes! What a great community we have here.

 

Thank you for the advice, you guys are awesome! Love this community!