EVGA 670 red and blue squares, crashing after about 20 minutes.

#1 Posted by Doctorchimp (4078 posts) -

So at first I thought it was just CS GO being a buggy beta, but after playing Diablo 3 for 20 minutes I am starting to see red squares and crashes too. Is anybody else with this card experiencing anything like this? The card is like a month and a half old in a brand new PC...

Or does it sound like I have to do an RMA?

#2 Posted by Robinson (174 posts) -

RMA that bad boy, sounds like a failure of the cooling system

#3 Posted by Doctorchimp (4078 posts) -

@Robinson said:

RMA that bad boy, sounds like a failure of the cooling system

That's the thing...when I come out of a game and I check temps it's never above 55 degrees

#4 Posted by SlasherMan (1725 posts) -

@Doctorchimp said:

@Robinson said:

RMA that bad boy, sounds like a failure of the cooling system

That's the thing...when I come out of a game and I check temps it's never above 55 degrees

Temperatures drop too quickly when you alt-tab or exit a game, so it's not a good way to gauge it. Try running something with an in-game overlay or a log (MSI Afterburner or Rivatuner are good options).

#5 Posted by Doctorchimp (4078 posts) -

@SlasherMan: I had Witcher 2 running on my HDTV and I had the EVGA precision on a different monitor and the temps were like around 62 a couple days ago.

#6 Edited by SlasherMan (1725 posts) -

@Doctorchimp: Is your card by any chance overclocked?

#7 Posted by Doctorchimp (4078 posts) -

@SlasherMan said:

@Doctorchimp: Is you card by any chance overclocked?

No, I just have it set on that option where it overclocks itself and when it gets too hot or it doesn't need the power it'll clock itself...this sort of situation is exactly the sort of reason I didn't want to touch the overclocking\

Also my card didn't come with a backplate and there were stickers on it, I didn't put on a backplate and I just slotted it in. Did I do something stupid or am I good?

#8 Posted by thedj93 (1237 posts) -

i used to get those squares when i had an overheating card back in 2004. but if your temperatures are cool then i dunno

#9 Posted by SlasherMan (1725 posts) -

@Doctorchimp: I'm not sure I understand what you mean by "backplate". Could you elaborate?

Anyway, this sounds to me like either the card is clocked beyond its capacity (but if you're not overclocking then that's probably not it), or the VRAM somehow got corrupted. My first guess would have been temps as well, but they seem to be fine from what you're reporting.

#10 Posted by Doctorchimp (4078 posts) -

@thedj93 said:

i used to get those squares when i had an overheating card back in 2004. but if your temperatures are cool then i dunno

Sounds like I have to lug back my computer to my HDTV room, actually I think I'll just RMA it. Is it painless? How do I ship it?

Also I love those cookies, I don't know where they're from. Probably Bangladesh or India right?

#11 Posted by Robinson (174 posts) -

If it is a bad paste job, half the GPU could be getting cooled while the other half just heats up more and more, eventually reaching a point where the other side can't keep it cool enough while under load, when you alt-tab out it stops having to work so hard and cools off fast. That 670 pumps out some real heat under load, I have two 570s (which do run hotter than a 670, but not insanely) in mine and during the winter I heat a floor of my house with it.

The temp sensors are also just a particular point on the die, not the whole picture. So while it may see the GPU die as being 62 degrees, the other side could be near 90.

And after years of messing with cards and adding aftermarket parts to make them run correctly, ive found it is almost always easier to just RMA and get a whole new card.

try using hwinfo64, it is a very detailed sensor report of your pc, lets you see everything, stuff that EVGA precision doesn't even have access to. it also will keep a saved record of all the sensor reports while you play, and once the computer hits issues you can go back and see what exactly occured.

Heck, your card might even just have a bad temp sensor and it is actually running significantly hotter than it realizes.

Have you tried manually engaging the GPU fans to run at 80-90% all the time while you play?

If it runs like that for more than half an hour, id say its def. a bad sensor, and it just didnt know to up the fan speeds.

#12 Edited by SlasherMan (1725 posts) -

@Doctorchimp: You don't need two monitors to check out your temps while in-game. I don't know about EVGA Precision, but there are plenty of other tools that will show this information in-game on an overlay or log it to a file. Others, such as Speedfan, will keep track of the highest and lowest temperatures the card gets to during their runtime.

#13 Posted by Doctorchimp (4078 posts) -

@SlasherMan said:

@Doctorchimp: I'm not sure I understand what you mean by "backplate". Could you elaborate?

Anyway, this sounds to me like either the card is clocked beyond its capacity (but if you're not overclocking then that's probably not it), or the VRAM somehow got corrupted. My first guess would have been temps as well, but they seem to be fine from what you're reporting.

The backplate didn't come with the graphics card I bought. I just put it in bare ass naked. Was that dumb? Or was the backplate necessary? Also there are stickers along the back underneath the card. The glue from those stickers didn't melt down and mess the card up did it?

I'm assuming these fears are irrational. But my back really hurts from some injury I don't know how I got and my HDTV is getting a purple band down the screen everytime there's a bright white screen. It has not been a good week for me..

#14 Posted by Robinson (174 posts) -

@Doctorchimp: RMAing isn't exactly painless, but it shouldnt be an ordeal. Register the Card with the company. Contact the company and say you need to RMA an overheating card and describe the problem as best you can. I believe they email you a mailing address, you put it back in the box you bought it in, then put that in a shipping box + packaging.

I've had companies send me a box with the shipping box and label inside, I've had others insist that I pay the shipping and materials cost and print my own mailing label.

Never had to RMA with eVGA tho

#15 Posted by SlasherMan (1725 posts) -

@Doctorchimp: No, the backplate is certainly unnecessary. First time I've even ever seen these sold as accessories to a card. I'd have assumed each card had its PCB reinforced enough on its own without the need to go and spend more money on stuff like this. I could understand the use of that if you're buying a particularly heavy aftermarket cooler (and even then, it's a bit of a stretch to actually need something of the sort), but otherwise no. And no, the stickers didn't melt down and mess the card up. They're paper stickers, aren't they? And even if they weren't, the back of the card isn't where the heat is emitted, so it should be fine either way.

#16 Posted by Doctorchimp (4078 posts) -

@Robinson: Yeah the card's registered already. From the screen it sounded like they wanted to see the graphics card first and then they'll ship me a card. I think I'm gonna call them on Monday and see what a guy can help me out with. Because I'd like it if they cross shipped me a card...but they want collateral.

#17 Posted by Doctorchimp (4078 posts) -

@SlasherMan said:

@Doctorchimp: No, the backplate is certainly unnecessary. First time I've even ever seen these sold as accessories to a card. I'd have assumed each card had its PCB reinforced enough on its own without the need to go and spend more money on stuff like this. I could understand the use of that if you're buying a particularly heavy aftermarket cooler (and even then, it's a bit of a stretch to actually need something of the sort), but otherwise no. And no, the stickers didn't melt down and mess the card up. They're paper stickers, aren't they? And even if they weren't, the back of the card isn't where the heat is emitted, so it should be fine either way.

Alright I'm just paranoid EVGA will laugh at my incompetence when they see my card.

Now I'll laugh at theirs goddamnit.

#18 Posted by Robinson (174 posts) -

@Doctorchimp: that backplate is totally not necessary, just an extra thing to sell you.

Stickers? Did you put them there? If not, probably not the culprit

have you updated the firmware lately? like tried to flash the bios?

#19 Posted by Robinson (174 posts) -

@Doctorchimp: they might be willing to cross-mail them if you put a credit card on file with them....but id rather not have them have mine, if i didn't have to.

This isnt the nicest thing to do to a B&M store, but.....

I believe Best Buy has a return policy of 30 days open box for full price, you could buy a 560 to hold you over, then return it when you get the replacement.

I've had to do this a few times when i had my PSU go out or having to RMA a card (been years). But i would pay in cash, then receive the exact same amount back. Just had to tell them that the part wasn't what you needed/didn't fit and they take it back.

#20 Posted by Doctorchimp (4078 posts) -

So I found out that there was a bios update for my card, I downloaded it and ran it. So I just started playing Diablo 3 to see it fixed my problem, and it crashed really hard this time. Red screen and my PC had trouble starting windows...

Yuck, this is is leaving the worst taste in my mouth for EVGA. I probably should have waited on the Asus one and I wouldn't have gotten a bad card.

RMA it is

#21 Posted by RIDEBIRD (1233 posts) -

@Doctorchimp: Red screen is not entirely uncommon with the 600-series. Have you updated drivers and the like? I would guess you have bad VRAM, and the only cure for that is a new card. EVGA are awesome except for their pretty shit always reference coolers, but what makes them so great is their return policy and warranty.

You will get a new card, hassle free, pretty much instantly.

#22 Edited by VACkillers (1075 posts) -

Try changing the drivers first, try a couple different ones, if you are still getting the same crap....RMA it, the GPU might just be defected, it doesn't need to be overheated to cause a defect, the GPU just might be fucked....

#23 Edited by Doctorchimp (4078 posts) -

@RIDEBIRD said:

@Doctorchimp: Red screen is not entirely uncommon with the 600-series. Have you updated drivers and the like? I would guess you have bad VRAM, and the only cure for that is a new card. EVGA are awesome except for their pretty shit always reference coolers, but what makes them so great is their return policy and warranty.

You will get a new card, hassle free, pretty much instantly.

Yeah, last time I called they told me to update the bios I did that. The PC crashed super hard to where it almost didn't boot up windows.

Now they want me to change the PCI slot....and then put it under load again. Even though I'm using the card right now in that slot and it's fine, only when it goes under load and the temp rises does it break.

#24 Posted by FritzDude (2279 posts) -

Doesn't the 600-series have something called GPU boost technology that automatically clock the core speeds or similar? Maybe there's a collision with the drivers there? I haven't done research on this, but maybe it's related to this and maybe it's possible to turn this off. Unless it's already turned off by default.

#25 Edited by Franstone (1149 posts) -

Sounds like a faulty card to me, I'd stop wasting time troubleshooting and just RMA that thing as soon as ya can.

Had a problem like that a long time ago with an old Radeon card creating purple boxes all over my screen.

Swapped out the card and was good to go.

#26 Posted by Doctorchimp (4078 posts) -

@RIDEBIRD: @Robinson: @SlasherMan:

Crisis solved!

For some reason my fan software controller was turned off, so even though I had the fan set to AUTO, it wouldn't change the speed for the temperature...

I should be fine right? I played Counter Strike GO for about an hour, no damage was done to my card right?

#27 Posted by Deusx (1910 posts) -

I'm pretty sure it's bad cooling. Either get more fans, a better case, or change the heatsink.

#28 Posted by Stonyman65 (2826 posts) -

@Doctorchimp said:

@RIDEBIRD: @Robinson: @SlasherMan:

Crisis solved!

For some reason my fan software controller was turned off, so even though I had the fan set to AUTO, it wouldn't change the speed for the temperature...

I should be fine right? I played Counter Strike GO for about an hour, no damage was done to my card right?

Maybe. Download unigine heaven or 3dmark and run it through a stress test - if the card is gonna fail, that will do it. Also, run GPUz while running the benchmark to check temps.

http://www.3dmark.com/3dmark11/download/

http://unigine.com/products/heaven/download/

http://www.techpowerup.com/gpuz/

If it survives all that with temperatures under or around 80C, then you should be okay. If it doesn't, RMA that card and get a new one and then do the same thing and make sure it is stable.

#29 Posted by VACkillers (1075 posts) -

Its too hard to say if there was any serious damage done, how long have you had it, and had it running like that?

#30 Posted by RIDEBIRD (1233 posts) -

If all you did was play CSGO for one hour when it started to buck out it's totally fine even if it was at a 100c or so

#31 Posted by Doctorchimp (4078 posts) -

@RIDEBIRD: @VACkillers: Not very long, and the PC would buck out after about 20 minutes of playing a game. I maybe did that 3 times when it started to die on me. I'm pretty sure I had the fans controller software turned on before. I was playing Diablo 3 and Skyrim for about a month with no problems.

@Stonyman65: I ran the 3D mark 11 and I got a Nvidia Kernel stopped working error on the High Temple...

#32 Posted by Stonyman65 (2826 posts) -

@Doctorchimp said:

@RIDEBIRD: @VACkillers: Not very long, and the PC would buck out after about 20 minutes of playing a game. I maybe did that 3 times when it started to die on me. I'm pretty sure I had the fans controller software turned on before. I was playing Diablo 3 and Skyrim for about a month with no problems.

@Stonyman65: I ran the 3D mark 11 and I got a Nvidia Kernel stopped working error on the High Temple...

RMA that card dude.

What case are you using? Just wondering.

#33 Posted by Doctorchimp (4078 posts) -

@Stonyman65: Cooler Master HAF 932

#34 Posted by RIDEBIRD (1233 posts) -

@Doctorchimp: If it failed 3Dmark it's probably busted, so yes, RMA it.

#35 Posted by MordeaniisChaos (5730 posts) -

@Doctorchimp said:

@Robinson said:

RMA that bad boy, sounds like a failure of the cooling system

That's the thing...when I come out of a game and I check temps it's never above 55 degrees

Doesn't matter, you are almost certainly experiencing hardware failure on a pretty damn new GPU. RMA. Any time you see artifacts like that, update drivers immediately, and then RMA that bitch if that doesn't do the trick.

#36 Posted by VACkillers (1075 posts) -

As people have already stated, RMA it back, I dont think its doing very well, it definitely sounds like a faulty card something not quite right with it, and you'll either get a new one back, or the card will get returned if there is nothing wrong with it, the only thing you loose is just time having the card... but best you sort that out now, then it completely fail in a few months time when you think its all good.

This edit will also create new pages on Giant Bomb for:

Beware, you are proposing to add brand new pages to the wiki along with your edits. Make sure this is what you intended. This will likely increase the time it takes for your changes to go live.

Comment and Save

Until you earn 1000 points all your submissions need to be vetted by other Giant Bomb users. This process takes no more than a few hours and we'll send you an email once approved.