Solved KA-Blam goes my graphics card. Bye bye PSU.

[TanyaC]

I thought I'd share an experience I had with my sons PC last night.

He was playing DOTA 2 and just as he quit there was a puff of smoke, a loud "bang" and a horrible stench permeated the air.

His EVGA GTX 570 SC graphics card over heated and blew up the power supply.

The thermal monitor in the PC didn't go off (it was set to 80C).

We had to wait several minutes for the card to cool down so we could remove it from the case.

This is the second of two identical cards that have failed in less than 6 months.

The PSU was a coolermaster RS-650-PCAR-E3 650w.

I can get the graphics card replaced under warranty (I hope), but the PSU was out of warranty. According to the ACCC (the consumer regulatory body here in AU), I should be able to get compensation for the PSU, but that could take months.

I believe the card failed first, and then took the PSU with it. I don't believe the PSU could have caused the problem, because in that case, the card theoretically would not have died because the power would have dropped.

Can anyone think of an argument EVGA might make to blame the PSU and attempt to refuse a replacement?

thanks
tanya

 

[Zepher]

Well, it's possible that the PSU failed and caused a huge voltage spike to the PCI-E Power port and fry the card.

It's also possible that the PSU was failing and the heat from the PSU rose up to the GPU causing that to overheat.

 

[Lebon14]

Wow, that blows dude. I personnaly believe what Zepher said : the psu went boom and brought the gpu with it... You are lucky that the rest survived...

 

[Maguscreed]

For future reference if you ever purchase evga again use their precision tool from their website. It will monitor the card and alter the fan in real time to avoid heat.
The built in monitor nvidia provides is by and large useless.
I do have to say I'm shocked to hear this as EVGA has general praise from myself and the internet at large even.
Here's a screen from it... as you can see it even clocks the core down to next to nothing if you aren't using it. Gives it about half juice for hd video and only cranks it up when I actually launch a game, at which point it rarely lets the temp over 60 , it's warm right now because I just popped out of tera to post this.

 

[TanyaC]

Wow, that blows dude. I personnaly believe what Zepher said : the psu went boom and brought the gpu with it... You are lucky that the rest survived...

Don't know if the rest survived yet. Have to get a new PSU to install before I can test. Foruntately I have spare parts - identical to the original PC just lying around waiting for such an event (except the PSU of course)

I spoke with the techs at the store and they are in agreeance with comments here - it is more likely that the PSU took out the GPU and not the other way around. I hope EVGA replace it under warranty otherwise I'm up for a new card as well as a PSU...

I stand corrected.

I'm going to get the Antec TruePower 650 this time around. It has 3 12V rails instead of one. I might buy two I think, so I'll have a spare.

The techs also commented it is not unusual for a GPU to be so hot it can't be touched.

Interestingly, the PSU warranty expired 4 weeks ago...

 

[carwiz]

For one thing, that PS doesn't put out enough current. You have to look beyond the 650W rating. Sure the video card says a "Minimum of 550W" supply is needed but the requirement also states a Minimum of 38 Amps on the 12v line. Your PS was rated at 36A MAXIMUM on the 12v line and that's probably it's solder-melting short-term limit. A more realistic maximum would be 70-75% of that or 25-27A continuous. The PS probably smoked the 12v circuit and left the 5v line running which in turn smoked the video card. I would certainly be testing the CPU after getting everything up and running too.

Minimum of a 550 power supply.
(Minimum recommended power supply with +12 Volt current rating of 38 Amps.)

 

[Trucidation]

Yeah, when picking a PSU I tend to pad the requirements really generously. This site helps you calculate your usage, it's useful as a rough guide:
eXtreme Power Supply Calculator

Then I throw in ~50% more, and start looking at PSUs in that range.

 

[Wishmaster]

Wow, that blows dude. I personnaly believe what Zepher said : the psu went boom and brought the gpu with it... You are lucky that the rest survived...

Don't know if the rest survived yet. Have to get a new PSU to install before I can test. Foruntately I have spare parts - identical to the original PC just lying around waiting for such an event (except the PSU of course)

I spoke with the techs at the store and they are in agreeance with comments here - it is more likely that the PSU took out the GPU and not the other way around. I hope EVGA replace it under warranty otherwise I'm up for a new card as well as a PSU...

I stand corrected.

I'm going to get the Antec TruePower 650 this time around. It has 3 12V rails instead of one. I might buy two I think, so I'll have a spare.

The techs also commented it is not unusual for a GPU to be so hot it can't be touched.

Interestingly, the PSU warranty expired 4 weeks ago...

If its an eVGA card with a SKu that ends with -AR ...AND .. you registered the card within 30days, they will replace that card. Its a Lifetiome warranty.
Otherwise, probably only 3 years.

I agree. use Precision and make a custome fan profile first thing. I have mine set to ramp up slowly and will be at 60% when it hits 70C. But, for most games it stay around 60-65C.

Secondly, the use of Vsync will help alot for some games. No point running a game at 300FPS, making the GPU run at full throttle, when it can easily run the game capped at the refresh rate of your monitor.
I/E if your monitor has a 60 refresh it physically can not display more than 60FPS anyway. No point working the card so hard. It will also run much cooler.


A good quality 650W PSu should be more than adequate for a 570 though.

 

[jeepmann4x4]

I have two EVGA GTX 570sc'ed cards and yes getting hot is the normal but in most games mine run in the mid 60's, if I play GTA IV for a long peroid of time then I hit the mid 70's.

 

[TanyaC]

Ok, here's the low down;

The problem was caused by the power supply blowing. According to technicians it "blew a cap". The safety features in the power supply were sufficient to protect the other components. They said it appears to have dropped power instantly when the PSU blew, and did not drop any voltage or current spike on the 12V or 5v rails.

Good news.

All other components survived the "explosion", including the graphics card, which ScorpTec tested for 20 minutes. We also found that the card was running in the high 80s when running DOTA, hence why it was too hot to touch.

This despite the fact that the case has good airflow and 7 fans. But we added a spot cooler near the graphics card anyway.

Interestingly, the other son who has an identical system, his card runs in the high 60s to low 70s when running the same game.

I bought a corsair TX-750M modular power supply. This one has all the same safety features, and a 7 year warranty.

In all my years of building/repairing systems I've never actually seen a PSU explode. What fun

As far as the graphics card goes; it is only a 3 year warranty. They were pretty good about replacing the other one which was definately faulty. We will install precision and keep our eye on things.

thanks for the tips and suggestions.

I didn't know that the FPS was tied to the monitor's refresh rate. Will check that out.

 

[MrNeeds]

Ok, here's the low down;

The problem was caused by the power supply blowing. According to technicians it "blew a cap". The safety features in the power supply were sufficient to protect the other components. They said it appears to have dropped power instantly when the PSU blew, and did not drop any voltage or current spike on the 12V or 5v rails.

Good news.

All other components survived the "explosion", including the graphics card, which ScorpTec tested for 20 minutes. We also found that the card was running in the high 80s when running DOTA, hence why it was too hot to touch.

This despite the fact that the case has good airflow and 7 fans. But we added a spot cooler near the graphics card anyway.

Interestingly, the other son who has an identical system, his card runs in the high 60s to low 70s when running the same game.

I bought a corsair TX-750M modular power supply. This one has all the same safety features, and a 7 year warranty.

In all my years of building/repairing systems I've never actually seen a PSU explode. What fun

As far as the graphics card goes; it is only a 3 year warranty. They were pretty good about replacing the other one which was definately faulty. We will install precision and keep our eye on things.

thanks for the tips and suggestions.

I didn't know that the FPS was tied to the monitor's refresh rate. Will check that out.

glad it's sorted Tanya, I had a similar incident with my motherboard, the caps popped because the southbridge was at about 110c, scary moment when you hear it go like that, but it's all good now.

 

[Lebon14]

*snip*

That's real good news! Happy ending.

 

[essenbe]

I'm glad it worked out well for you and your choice of the PSU is an excellent choice. I also echo the others who advocate EVGA's Precision X. It will also let you customize a fan profile that will keep the card cool. Most monitors have a refresh rate of 60HZ, meaning it is only capable of actually showing 60 FPS. Anything the card puts out over that is wasted because the monitor can't handle it. I believe that is the short version as I understand it.