Cooked video card?

Usually after ive been playing a game for a few minutes, my screen gets covered with red squares, or red lines. The computer locks up, and then my monitor loses its input signal and the video card fan kicks up to 100% I havent had a problem with any program outside of games that i play.

is it deep fried?

 

Sounds like it.

Install GPUZ (available here) and check your temps after a couple of minutes of game play.
By the sounds of it you'll see high temps.

 

While im gaming my temperature never goes above 87 C according to Catalyst Control Center. The card is less than a year old... i didnt think temperatures like that could cook a card this fast.

 

Yeh, 87C is at the limit of most gpu's, also gpu temp sensors are sometimes inaccurate on the cheaper cards, so it could be anything up to 100c.

Sounds like it is dead! Is it still under warranty?

Out of interest, what gpu do you have and how old is it?

 

Acording to his sig, he has an HD4870 x2, shouldn't be getting that hot though.
Clean it up and get some good airflow in your case and you may extend the life of it, if you continue on the road you're on it will die 100%

 

87 is NOT hot enough to causes full on crashs the cards are rated for 115c check and make sure you have a good psu and the latest drivers

 

115c? Where are you getting that info from?
My HD4890 runs Crysis maxed at around 60C
87C is way too high and if it's been running at that for a while then it's possibly damaged resulting in a shutdown at lower temps.
Also once damaged the temp goes up and down much faster, I've been through this exact senario in the past 12 months and I've seen a card idle at 60C and jump to 94C in seconds with a windowed game, simply because it's been burnt out.
I've also never seen the results he's having because of PSU failure, very low FPS yes, graphical glitching, no. This is my experience anyway, but 115C is way too high for ny card on the market and I've never seen anyone say above 85C is ok.... until now.

@Deamon... Have you tried several different games, is it game specific?
Have you tried running the same game at lowest settings?
Do you have decent airflow in your case?

I know some 4870 x2's run hot and the hotter they run the faster they die, you can probably still save it with good cooling and by cleaning the GPU heat sink.

 

"rated" temps are "death" temps. It is the max temp the GPU can run with out suffering damage, not the max temp they will run and still be stable.

Odds are this is a heat issue. Try getting your temps below 60c. If this helps but doesn't solve the problem entirely the card might have been weakened due to running so hot. If that is the case try clocking it down some.

 

87c is WITHINSPEC all gpus are rated to 100c Before they will start crashing
I have had 3 video cards that ran in the 90's 8800GT A 4870 x2 and a 3850 that ran all the way up to 120c before it powers it self down those temps are alarming but they are WITHIN SPEC ati and NV both tests there cards to at least 100c that is WARM and you should only get those temps when getting close to 100% load
and a 4870 x2 will get upwards of 90
http://www.digital-daily.com/video/sapphire_radeon_hd_4870_x2/index2.htm
I suspect there is something else going on ( lack of power / bad drivers faulty card) its NOT temp related that I can assure you

 

Ok, you could be correct in it not being temp rated but your findings are not absolute, my X1950 pro would crash out at 88C because it was damaged, 87C is not acceptable.

Find me an ATI or Sapphire info thats says 90C is normal and I will start agreeing with you.

Also attempting to drop the temp is an easyish move, replacing the PSU is an expensive move, go for the easy one first

 

Posting links to reviews and using all caps does not help Deamonstrom. If you want to argue about who is right and who is wrong take it some where else. The bottom line is that temps that high are very unhealthy and could very easily be the case of the problem.

 

What PSU are you using in you computer Deamon? I would check it and make sure it has enough power for your computer. Also, remove the video card and make sure there is no dust build up in the fins of the heatsink.

It seems that for a while, GPU manufacturers are really getting cheap with the stock cooling, with quality and shoddy installation. You might consider removing the stock cooler (may void your warranty) and replace it with a aftermarket cooler, or maybe just clean up the card and stock heatsink, apply some TIM, and replace the old on. I had a few cards in the past, where the stock heatsink was not even contacting the GPU.

Also, as suggested already, try cleaning up the wiring in your case, and make sure you are getting good airflow.

Good luck! :major

 

I have a Thermaltake Toughpower 850W PSU. I looked over the heatsink and the fins are pretty clean... the temperature has run around 86 C from the beginning. the fan speed on the card doesnt even start to increase until the temperature gets up around 88 C.

Earlier today i set the fan speed manually to 60% and the computer crashed after about 20 minutes of game play.

Im glad i got the extended warrenty

 

Ok well I stand by a heat issue probably caused by overheating from the start has weakend the card, my opinion anyway.
And it's quite possible that your 86-87C reading is much higher durin game play, probably up in the low 90's and possibly higher in certain scenes.

I would contact Sapphire, their online support ticket are not bad, fairly quick response. Just explain to them whats happening and what they think may be causing it. They may suggest you return the card.

As a matter of interest, what kind of cooling do you have, fans , case etc?

 

My case is a Lian Li 60BPLUSII
i have one 120 mm intake fan on the front, one 129 mm exaust on the side, an 80 mm on the top and on the back.

Motherboard temperature is usually around 30 C

 

Just to let you know, my card is "rated" at 110c, but i have never seen it go above 65c. EVER.

 

I used to have this problem when my PSU was not powerful enough for the upgrades I put in my machine. That GPU is pretty power hungry however a 850w PSU should cope. What else have you got running in your machine?

Just to let you know I changed my PSU for a better new one and have had no problem's since.

Have you added anything to your comp? How old is your PSU?

 

Well looking at that case, it's not exactly a power house when it comes to cooling.
Lets put it this way, my card is an HD4890 Vapor X, it runs around 55-65C
It has the Vapor X cooling, my case has 2 x 120mm intakes on front, 1 x 120mm exhaust on back and 1 x 210 mm outlet on top.

The 4870 x2 does run hot, everyone says so, most well respected info sites will tell you that if you are going to have that card you need good cooling.
Given the card and the crappy stock cooler it comes with, plus the low end cooling system in your case would tell me that it's been running too hot for too long.
I can't say 100% that it is your problem, but like I said in my experience when a GFX card acts the way yours is acting, not so much the PC crash but the red boxing and the fact you said it powers off before your PC shuts down would lead us to believe it's a heat problem.

If I was going to have an HD4870 x2 (which I would love one lol) I would be upping my already ample cooling.

The PSU idea, it could be causing the problem but your PSU is plenty enough to handle your setup unless you have 10 external drives running and the likelyhood that it's crapping out is lower than you GFX card given the evidence.

Like I said, I have seen underpowered PSU's cause heavy lag and extreme low FPS, which has on ocasion lead to PC crash, but I've seen graphical glitches like your red boxes many times and every time it has been caused by a dying GFX card, 90% of the time caused by excessive temperatures.

Now, if I were in your position I would be giving myself a few options.

1: Test games on lowest settings and compare temps.
2: Contact Sapphire and ask opinion.
3: Increase airflow by adding fans, re-routing cables to allow airflow.
4: Try a different PSU as suggested by others, although the kind of power you require this could be an expensive waste since it's unlikely you could borrow one powerful enough.

Bare in mind I am not you, but I have had a lot of experience with GFX cards over the past couple of years and have done a huge amount of research on them.
These are still my opinions and I'm simply trying to find you a solution that doesn't require you to spend much money lol

 

I've been reading up on this card in more detail, many say that it simply does run at those temps due to Dual GPU's and poor cooling/casing.
But what is still bugging me is the above quote.
If it is the PSU or something else other than a heat issue, why would the fan kick upto max?

All the specs say 80-90C on an HD4870 X2 are within range, but many many people are giving advice to owners of these cards to manually set the fan to 50% or higher.

I'm quite puzzled by this since research suggests it is running fine but eveidence in the problems you are experiencing suggest you have heat issues.

Many people also reported idle temps in the normal range of 55-60C, which IMO is fine, but some also said they are getting idle temps of 75-85C which is not fine.

So, with this new information in mind if this were my problem I would now be doing the following:

1: I would check temps at idle.
2: I would run ATI tools little test software and watch the temps.
3: Then I would try the same tests again but first I would set fan speed manually higher, I personally would start at about 50-60% and work with it until stable.

I seriously hope for your sake it IS the PSU as it's much cheaper to replace, but before spending any money I would want to eliminate all other possibilities first.

And since evidence suggests temperature issues, I would start there. If all results come back with normal ranges such as 87C under full load, 60C at idle then I would maybe start looking elsewhere for problems.

I have to admit I feel a little humbled here since research into this card has taught me a few things, but I am just niggled by the Artifacting and Fan behavior you are experiencing.
Your pc is acting in exactly the same way mine did when my 9800 died AND when my X1950 died, actually both still work with intense cooling but operate at much higher temps than they should and they will both crash out in exactl the same way your's is if installed into a case with mediocre cooling.

I'm certainly puzzled, IF you can get your hands on another PSU for testing then it would be nice to eliminate it one way or another, but I know that is a tall task since you need so much power.

 

I have also been looking into this card and as it says above it does run very hot and the temps your getting are kind of normal. If it was my card I would not be happy with that at all.

I also would agree it seems like a heat issue and that is exactly what I thought when it happened with my lower spec card. I tried all sorts of stuff before deciding it was the PSU because they are so expensive to buy. Well a decent one any way. Buy cheap buy twice and all that.

The advice given above is good sound advice and I would try all of these before going down the shop.

Also if you do buy a PSU and that is not the problem then there is no reason you could not take it back saying that your girlfriend is a computer geek and she bought you the right one for your birthday.

I dont think I have seen any post about the set up of your comp anywhere.

 

I also suspect heat is the issue. Why not try manually maxing out the fan on the GPU to 100%, and see if you can game longer. I think the stock cooling is not making proper contact with the card, or is just a poor design. As stated above, contacting Sapphire is not a bad idea.

 

I contacted sapphire and they suspect it is a "hardware issue" they sent me all the RMA forms so hopefully ill be getting a new one.

I also clocked the GPU speed down to 600 MHz. Running the fan manually at 60 % i have been able to run everything that was crashing before.

 

Glad it's getting sorted out, that card is a beast and I'd hate to see it go down.
Figured Sapphire would ask you to send it back, they are pretty good like that, tech help aren't exactly tech gods but they are as helpful as they can be.

I would strongly advise you to up your cooling though to prolong the life of the replacement, if you can afford it I would suggest a new case.
I personally always recommend the Antec 900 because it's a good price and has worked wonders for me, but any case with lots of room and fans is good.

http://www.newegg.com/Product/Produ...129021&cm_re=antec_900-_-11-129-021-_-Product

Be aware that many o the gaming cases have the PSU mounted at the bottom so make sure your PSU has decent length cables.