Drivers Cause GTX 590 To Burn (On Video)?

Not to nitpick but amperage would likely actually drop when voltage rises. The voltage goes up but the overall load will stay the same for the most part.

You don't push watts, your load pulls them.

Not to nitpick but your just completely wrong in your assumptions on how this works. As you increase voltage the amperage will increase at a very high rate.
 
Take a look at the LED's on the mobo .. don't those indicate how much power is being drawn? At the same time the spark and smoke show up the mobo appears to indicate a huge power draw.
 
Also why does the blue led on the CPU fan appear to flicker off and back on around that same time?
 
Proof AMD/ATI cards are far more efficient.

But really, for real, I think it's pretty obvious NVIDIA cards aren't as efficient as AMD/ATI. Not being a fanboy here, because all my GPUs are NVIDIA.

Efficiency has absolutely nothing to do with overvoltage/overcurrent protection failing and cascading damge.
 
Not to nitpick but your just completely wrong in your assumptions on how this works. As you increase voltage the amperage will increase at a very high rate.

false and true at the same time. if you increase your inport V the amperage would go down. So lets say that your 12 rail was 13v your card would draw fewer amps form your 12v line but the current would be the same. Now if you increase the v to the gpu core the power circuits have to work harder to transform 12v in to 1.2v so your pulling more form the 12v side then before
 
You Sir, are NOT posting in the "Doily knitting granmas who push prams and eat porridge" Forum.

Here, we revel in stupid :p

I'd prefer Ma'am, but point taken :)

Also, I said that it seems like the system is only booting up as you can see the fans spinning down and stopping, then spin up again a second before the smoke is released. Definitely seems like they were restarting the system.
 
Watch the fans closely the whole system had a drop in power right before it popped.
 
It looks like the power connector de-soldered its self. You can see the solder drip off the back of the card.
 
I'd prefer Ma'am, but point taken :)

Also, I said that it seems like the system is only booting up as you can see the fans spinning down and stopping, then spin up again a second before the smoke is released. Definitely seems like they were restarting the system.

Or a sudden spike in power draw over the motherboard. I've seen this happen on my Tbird set up back in the day when I fried one of my CPU's trying to take it past 1400mhz. Right before it fried everything that was hooked up to the motherboard spiked, LED's quickly flashed once and the fan on my GTS2 stopped for a split second before going back on. Then dead CPU.

If this is during boot up and its at IDLE power using BIOS vCore settings and not software settings then thats even worse!!!!

These things were supposed to be OC'ing monsters, guess that has to be taken with a huge grain of salt now.
 
Also, I said that it seems like the system is only booting up as you can see the fans spinning down and stopping, then spin up again a second before the smoke is released. Definitely seems like they were restarting the system.
That is not consistent with how volt modding is usually done (a bios modification might cause a failure like what you describe but there is no point in taking that approach when there are simple software solutions available).
 
From now on, this should be referred to as "sweding" a videocard.
I second this motion!

It should also be noted (because a lot of folks posting seem to think this is just a bad sample or the PSU failed) that this is occurring across multiple 590 samples (sweclockers popped 2 590's and TPU popped 1... more to come?).
 
false and true at the same time. if you increase your inport V the amperage would go down. So lets say that your 12 rail was 13v your card would draw fewer amps form your 12v line but the current would be the same. Now if you increase the v to the gpu core the power circuits have to work harder to transform 12v in to 1.2v so your pulling more form the 12v side then before

Current is Amps, and it depends on if it was in parallel or series on what amps do, but typically, the only thing that makes amps go down is the loads in the circuit. Increasing voltage does not increase/decrease amps.

if you look at the video, the short went from the connector on the card and jumped through the air to the ground screw on the motherboard, that is a massive jump, wonder what the voltage had to spike to to make it jump that far.
 
As for the 267.52 driver being the "wrong" driver, it is the driver in the box! Do not use driver that comes with the card. (They should have guessed that?)

At least 4 more cards frying posted in the thread over at Anands.

NVidia driver fail.
 
nVIDIA thought it had a hot card on it's hands, but this launch just went up in smoke. :)
 
'Nvidia Fanboy Mode++': This goes to show you that there is no way that a single pcb can withstand the COMPLETE and TOTAL AWESOMENESS that is TWO GF110s!!!!1One!!!!Two

(Back to normal) That's crazy! Apparently, the PSU's or motherboard's OCP/OVP kicked in and shutdown the system. Seems that when the system shutdown, one of the 8-pin connector points must have slightly melted.

'ATI/AMD Fanboy Mode': This goes to show how AUSUM the 6990 is!!!!Nvidia Sucks!!!

:D
 
Nvidia has a response for the card burning out with the early beta driver.

http://www.++++++++++++++++++++/news/video/nvidia-responds-complaints-newly-released-gtx-590-frying/

Apparently in the reviewers guide they told the reviewers to use the new driver specifically and that the old one was a beta driver for testing the engineering samples. They also list recommended ranges of voltages to try. It took reviewers specifically ignoring both bullet points for the cards to fry.

I suggest you go over to TPU and tell that to Wizz, as he never was told and used the driver shipped in the retail box. Same at sweclockers.

Same with alienbabeltech.(Chech the Anand thread if you don't believe it.)

This is nVidia trying to spin that they made a mess and sent out wrong drivers. We can only hope everybody that buys a 590 hears about it before they install the driver in the box.
 
Nvidia has a response for the card burning out with the early beta driver.

http://www.++++++++++++++++++++/news/video/nvidia-responds-complaints-newly-released-gtx-590-frying/

Apparently in the reviewers guide they told the reviewers to use the new driver specifically and that the old one was a beta driver for testing the engineering samples. They also list recommended ranges of voltages to try. It took reviewers specifically ignoring both bullet points for the cards to fry.

Heh, well that explains a lot :D Url has been "sanitized" but did they post something that shows that Nvidia did actually supply this to them?
 
Yes sorry about that as the link was sanitized here is what it says (below)

NVIDIA Responds to Complaints of Newly Released GTX 590 Frying
Blog > News > Video Card News > NVIDIA Responds to Complaints of Newly Released GTX 590 Frying
Share |
Posted by FiXT— March 24th, 2011, 5:31 PM

smoking gtx 590 300x162 NVIDIA Responds to Complaints of Newly Released GTX 590 FryingWith a successful launch of their flagship card today, March 24th, 2011, the NVIDIA GeForce GTX 590 was met with positive feedback…for the most part.

It appears that some of those who were out to test the limits of the behemoth, managed to capture the $700 card going up in smoke, literally. Sweclockers, among others, reported and video taped that the GTX 590 was in fact so powerful that the card killed itself during overclocking. While using the beta drivers which were not intended for release, some reviewers willingly took the voltages of their GTX 590s far past NVIDIA’s limits or recommendations. It should be noted, that a quick sampling of many of the web’s premier review websites were using driver version 267.71 in their reviews while others were using slightly older AiB-supplied drivers which did not implement the over current protection properly.

“The few press reports on GTX 590 boards dying were caused by overvoltaging to unsafe levels (as high as 1.2V vs. default voltage of 0.91 to 0.96V), and using older drivers that have lower levels of overcurrent protection. Rest assured that GTX 590 operates reliably at default voltages, and our 267.84 launch drivers provide even more additional levels of protection for overclockers. For more information on overclocking and overcurrent protection on GTX 590 please see our knowledge base article here: http://nvidia.custhelp.com/cgi-bin/nvidia.cfg/php/enduser/std_adp.php?p_faqid=2947.”

In addition, here is an extract from ASUS’ GTX 590 reviewer’s guide:

It is not advised to exceed the 1.050 to 1.065 vcore range as this begins to meet the limits for the OCP/OVP mechanism on the card. Exceeding these values without disabling OCP/OVP or having superior cooling could affect the lifespan and functionality of the card/gpu.

The lesson of the day is this folks: ignoring manufacturer’s recommendations and overclocking your card way beyond its limitations (especially when using beta drivers and beta software) can in fact lead to the unfortunate killing of hardware. Fancy that eh?

I highlighted in Green the quote from nvidia, underlined the sentiments of the author of the article
 
I suggest you go over to TPU and tell that to Wizz, as he never was told and used the driver shipped in the retail box. Same at sweclockers.

Same with alienbabeltech.(Chech the Anand thread if you don't believe it.)

This is nVidia trying to spin that they made a mess and sent out wrong drivers. We can only hope everybody that buys a 590 hears about it before they install the driver in the box.

I believe it depends on which card they got. Many peoeple here in the states got the Asus Board or EVGA and on our side both of those companies provided the notes as far as I know. Brent posted the notes in the article regarding asus and their recommended driver and voltages to use.

Heh, well that explains a lot :D Url has been "sanitized" but did they post something that shows that Nvidia did actually supply this to them?

They have quotes around it, I highlighted it in Nvidia green so it's clearly standing out. :D
 
lol, reminds me of a tech I knew back in the day who used to repair power supplies. If one came in for repair with say a blown 5 amp fuse, he'd replace it with a 20 amp and fire it up. Whichever component blew off the pcb is the one he'd replace...

That's actually a pretty good way to force flaky components into failure. The problem is when more than just that one component blows up... or the component decides to take more than just itself off the board. I'd do the same thing to old CRT monitors that I knew were on their way out... switch the thing on, flick the power switch on the switch rapidly back and forth. Sometimes you get that night *POOF!* of ozone and fried components.
 
lol, reminds me of a tech I knew back in the day who used to repair power supplies. If one came in for repair with say a blown 5 amp fuse, he'd replace it with a 20 amp and fire it up. Whichever component blew off the pcb is the one he'd replace...

That is actually not a bad way to find out what is faulty in some cases.
 
I'd be so ticked if any of my hardware ever did that brand new.

That is pretty much what happened to my 285 video card. I was running an md sim in linux when...pop....and out comes the magic smoke. I think it might've been from the beta drivers....
 
Apparently in the reviewers guide they told the reviewers to use the new driver specifically and that the old one was a beta driver for testing the engineering samples. They also list recommended ranges of voltages to try. It took reviewers specifically ignoring both bullet points for the cards to fry.

What sort of fool believes nV damage control spin like this? The blame is too obviously on nV, not the reviewers, many of whom have a past history you can look at and see that they are obviously not incompetent.

The GTX590 may be a fast GPU but something is clearly wrong here and until nV or the board makers fix it you'd be tempting fate to get one.
 
That was just amazing to see the solder shoot off the board. Not good, Wonder what is really going on here.
 
So that's now THREE drivers which have blown up nvidia video cards, within the last year and a half?
 
What sort of fool believes nV damage control spin like this? The blame is too obviously on nV, not the reviewers, many of whom have a past history you can look at and see that they are obviously not incompetent.

The GTX590 may be a fast GPU but something is clearly wrong here and until nV or the board makers fix it you'd be tempting fate to get one.

This is being accepted globally from what I understand. In all instances where the card burnt out, as Its said nvidia communicated with each reviewer and/or etailer the old driver was being used. They specifically found it to be driver related.

At the end of the day I agree with you, as this sounds all nice and sugar coated but you wont catch me buying them until we see that these cards are indeed stable and durable and this is all one big driver fluke.
 
Waiting for Brent's overclocking review...

Brent, be sure to have the camera running.

And a fire extinguisher.
 
What happened here were the 1.7 woodscrews they re-used from the Fermi project... those damned things just went *poof* like matches when hit with that voltage... :D

woodscrews.jpg
 
I had the same thing happen to a stick of GSkill Ripjaws memory once.

I noticed the mobo warning light flashing in the video too. Hope the board was alright.

JW
 
Back
Top