Vortex II, Windows 10 Killed Graphics Card

[Bluesteel]

Having already searched the Internet and it has been all doom and gloom.

I decided to update my 2011 Vortex II (GTX 580M) to windows 10 yesterday, all was fine until after it had finished auto updates. I shut it down to do some other work, after a couple of hours I switched it back on and discovered that the display had all sorts of weird lines across it. Opening control panel the following status is displayed 'Windows has stopped this device because it has reported problems. (Code 43)'.

The card is still visible in BIOS, Speccy sees it but has stopped reporting temps.

I have tried all sorts, Re-installing windows 7 doesn't allow me to install the drivers, Linux mint didn't even display anything.

Is there anyone still on here using one of these laptops, and has anyone else come across this problem.

If it is goosed, does any one have any experience with getting what appears to be compatible cards off ebay. Or is there a secret stash of unused new ones hidden away somewhere.

I know it is an old laptop but it has served me well and still did what I mostly needed it to, and eventually I was hoping to hand it down to the Kids who pretty much only play Minecraft or Roblox.

Thank you in advance

Blue

 

[ubuysa]

I'd be amazed if Windows 10 alone really has damaged your graphics card, perhaps it was on its last legs anyway and Windows 10 has just pushed it over the edge?

One thing to try would be to remove the CMOS battery for a minute or two (and TBH given the age of the laptop it would be wise to replace it with a new one). You may need to enter the BIOS setup on next boot to select the default configuration. Then try clean installing Windows 10 from installation media, delete all partitions on the system drive (or whatever part of the drive you use for the system), select the unallocated space and click Next. The installer will create the appropriate partitions and install Windows into the right one. Once Windows is installed run Windows Update repeatedly until no more updates are found, if you're told to reboot then do so. That will (should) install all necessary drivers, Windows 10 makes a good job of finding the right drivers.

If the driver installs fail there or if it's not working properly afterwards then it most likely is a hardware failure. Sadly graphics cards in laptops are almost always soldered in and are not replaceable.

 

[Bluesteel]

Thankyou for the tip, I did try it but to it didn't work, all the research I did seems to point to a very specific failure to do with the combination of either GTX480M or GTX 580M, the 1709 build of windows and a specific NVIDIA driver which automatically gets downloaded and installed which can cause problems with those cards.

Still researching it on the internet, I was just wondering if anyone else on here had a similar experience.

 

[ubuysa]

Thankyou for the tip, I did try it but to it didn't work, all the research I did seems to point to a very specific failure to do with the combination of either GTX480M or GTX 580M, the 1709 build of windows and a specific NVIDIA driver which automatically gets downloaded and installed which can cause problems with those cards.

Still researching it on the internet, I was just wondering if anyone else on here had a similar experience.

If that's the case I would phone PCS and ask their advice. They are gamers too and they should have a handle on issues like this.

 

[Scott]

Manually download the latest drivers from NVidia. Download DDU and do a clean uninstall of all the drivers for the GPU. Disconnect your internet so that there is no auto-update interference. Install the drivers manually downloaded.

If it's the latest driver set that is the problem simply download a previous driver set from a known good configuration.

 

[Bluesteel]

Latest update: At least temporarily fixed, definitely a hardware issue, put graphics card in the oven today at 140C Fan for a couple of hrs, and let it cool down with the oven. Re-tim and install and all is well for now, I'm quite sure this isn't a permanent fix and I will have a similar issue later.

Thanks for all the suggestions Blue

 

[Scott]

What a coincidence that it presented after a windows 10 upgrade! Really unlucky with that.

 

[Tony1044]

Latest update: At least temporarily fixed, definitely a hardware issue, put graphics card in the oven today at 140C Fan for a couple of hrs, and let it cool down with the oven. Re-tim and install and all is well for now, I'm quite sure this isn't a permanent fix and I will have a similar issue later.

Thanks for all the suggestions Blue

I'm curious on this - 140C is beyond the degradation point for silicon, but well below the melting point of solder, so I'd love to know what this might have achieved as on the face of it, it shouldn't resolve dry joints but should have damaged the chips.

Also, forgive my ignorance, but what is a re-tim?

 

[Scott]

I'm curious on this - 140C is beyond the degradation point for silicon, but well below the melting point of solder, so I'd love to know what this might have achieved as on the face of it, it shouldn't resolve dry joints but should have damaged the chips.

Also, forgive my ignorance, but what is a re-tim?

140C for hours should definitely have killed the card, it's normally recommended around 15-20 minutes..... so that the surrounding temp is 140 but the components don't soak to 140 themselves.

The chips are usually soldered using low temperature solder, the melting point is much lower than conventional solder and oven tricks are often used to re-flow the chips and make the contacts. If you search "solder balling" you will find out a bit more about it.

Re-tim is the Thermal Interface Material. He basically re-pasted the cooler onto the chip

 

[Tony1044]

I am aware of low-melting point solder but I thought it needed around 180C to melt properly. Edit: a quick Google suggests this may be the case, though with many stating they have a melting point of around 138C and a peak flow of around 178-180C

Hmm I'd be very dubious of the success of it with temperatures so close to the actual melting point. And as above 140C is 35C above the degradation point of silicon so I'm amazed it didn't kill it outright.

Thanks for the info on the re-TIM. Just a fancy way of saying repasted.

 

[Scott]

I am aware of low-melting point solder but I thought it needed around 180C to melt properly. Edit: a quick Google suggests this may be the case, though with many stating they have a melting point of around 138C and a peak flow of around 178-180C

Hmm I'd be very dubious of the success of it with temperatures so close to the actual melting point. And as above 140C is 35C above the degradation point of silicon so I'm amazed it didn't kill it outright.

Thanks for the info on the re-TIM. Just a fancy way of saying repasted.

It's a well known method for recovering GPUs. The Xbox360 was famous for it too. The red light of death was often cured by turning the system on while wrapped in a blanket for an hour. The heat generated inside the console allowed the solder to melt and the deflection (due to excess heat to begin with) would sort itself and the console would work again..... for a time at least.

The number of hours in the oven is definitely throwing me as I'm very surprised that the silicone didn't take a pasting (geddit). Normally, thanks to thermal dynamics, the solder melts before the silicone gets close to it's degradation temp but I think hrs in the oven is pushing things a bit

And yes, re-tim = re-paste