Another dying 280?

Status
Not open for further replies.

LstBrunnenG

Supreme [H]ardness
2FA
Joined
Jun 3, 2003
Messages
6,677
My week-old BFG GTX 280 OC is artifacting. I also alt-tabbed out of CS:S (CS! Not even Crysis!) once to notice the temperature had climbed to over 100 C. CS was still running fine, though it was artifacting at random intervals. It was as though it were trying to draw my character model ahead of me, in anticipation of me running forward or something.

Seeing as I already had to wait a week for the replacement PSU apparently necessary for this card to come in the mail, I am *not* happy with the prospect of sending it back. I'm considering using it until the 55nm version comes out, and stepping up.

I will try ghetto-rigging a 120 mm fan to blow in directly its general direction.

On the flip side, I am extremely, extremely happy with this card's performance. Even though the frame counter in Crysis says about the same things as it did when I was using my 4850s in CF, the gameplay experience feels much smoother. The minimum framerate must be much higher. I also see no evidence of the odd artifacting and memory usage issues I encountered under DX10 mode on my ATi cards. I would be entirely happy with my purchase if I didn't start seeing artifacts half an hour into gameplay.
 
The defective GTX 280's seem to heat up to around 100c. Best thing to do is RMA it before the damn thing dies on you. I've experienced that but fortunately I bought mine and was able to return it during the exchange period.

I originally bought two, then I exchanged the one and the two I have now have been totally solid.
 
Hmm.

With the case side off, it seems to hover around 90 C. I also haven't seen an artifact lately.

I should really be using FurMark. With checking for artifacts this fun, I'm going to be doing validation all night :p
 
Well I tried FurMark once and the temperature jumped right to 106 C, even with the case side off. It then BSODed 5 minutes later.

My CPU overclock just passed 22 hours of Prime95, but just in case I decided to back it off. Everything is back at stock now, and I'm running FurMark again. Still going at 106 C after 2000 seconds. It's doing this weird hitching thing, where the doughnut will stop for half a second and then jump ahead, over and over again.

To anyone else with a 280, I ask you to do three things:

1. Open up FurMark. Set it to something slightly less than your monitor's native resolution, set 16X AA, and uncheck the fullscreen box (I'm using 1680x1050 on my 1920x1200 monitor).

2. Open up GPU-Z and set it to the sensors panel.

3. Walk away for ten minutes. When you return, do you notice the hitching? What are your temperatures? Please post your findings.
 
Well I tried FurMark once and the temperature jumped right to 106 C, even with the case side off. It then BSODed 5 minutes later.

My CPU overclock just passed 22 hours of Prime95, but just in case I decided to back it off. Everything is back at stock now, and I'm running FurMark again. Still going at 106 C after 2000 seconds. It's doing this weird hitching thing, where the doughnut will stop for half a second and then jump ahead, over and over again.

To anyone else with a 280, I ask you to do three things:

1. Open up FurMark. Set it to something slightly less than your monitor's native resolution, set 16X AA, and uncheck the fullscreen box (I'm using 1680x1050 on my 1920x1200 monitor).

2. Open up GPU-Z and set it to the sensors panel.

3. Walk away for ten minutes. When you return, do you notice the hitching? What are your temperatures? Please post your findings.

I'd do this for you if I had my GTX 280's installed. Right now I've got a pair of 4870 X2's in my box.
 
My first 280 got up to 106 in furmark. I called up BFG and told them I wasn't happy with it and they told me that I could send it in for a new one, or I could wait until it crapped out (if it did). I went with the former and now have a card that gets up to mid 80's in furmark. I suggest you send your card in and get one that isn't so hot :D
 
You should RMA it. IIRC you cannot step up with a defective card; they will reject your step up.
 
Well I tried Crysis last night, and the max it got up to playing Crysis was 90 C. It usually hovered around 80 C. No BSODs even though it was a rather long session.
My first 280 got up to 106 in furmark. I called up BFG and told them I wasn't happy with it and they told me that I could send it in for a new one, or I could wait until it crapped out (if it did). I went with the former and now have a card that gets up to mid 80's in furmark. I suggest you send your card in and get one that isn't so hot :D
Did you change anything about your cooling setup between now and then? My setup is rather spartan - I have a P180 with the top and rear fans, but the lower part is pretty much on its own airflow wise.
You should RMA it. IIRC you cannot step up with a defective card; they will reject your step up.
*sigh*

If I do send it back, I won't do it until after this weekend.

EDIT: Would using one rail from the PSU - a single cable with both a 6-pin and 8-pin PCIe connector on it - cause my card to overheat?
 
I believe that you have one major issue going on. You need to have two separate cables, one with a 6-pin & the other with the 8-pin. As far as I know, every PSU I've used has separate cables for PCIe power connectors.
 
Tried using two seperate rails. No artifacts so far, but the card still does go to 105 in Furmark. Playing CS:S at 85 though.
 
Did you change anything about your cooling setup between now and then? My setup is rather spartan - I have a P180 with the top and rear fans, but the lower part is pretty much on its own airflow wise.*sigh*

No sir. Same fans, same direction they are blowing, same everything.
 
I got my BFG 280 OC today and after installing the drivers the temps started out at 70c... when I opened up IE7 the temps went right up to 85c. I tried to play COD4 and the card shut down before I could start the first level. Been messing with it all night and it will not go below 85c unless it is right after I power up after letting it cool down for a while. I replaced an 8800 GTX (moving this to a second computer) plus I have a thermaltake 1000 watt power supply and plenty of cooling. I guess I will RMA this one and try again. I played with the fan speeds and with the fan set at 100% it would cool down to about 77c but it went right back up after I slowed the fan.

wolfie
 
Just got a BFG 280 OC as well and it seems like this is a pretty common occurrence. :( CoD4 hovered in the 90's, Crysis stuttered at about 110 before it shut it down. RMA time.
 
I got my BFG 280 OC today and after installing the drivers the temps started out at 70c... when I opened up IE7 the temps went right up to 85c. I tried to play COD4 and the card shut down before I could start the first level. Been messing with it all night and it will not go below 85c unless it is right after I power up after letting it cool down for a while. I replaced an 8800 GTX (moving this to a second computer) plus I have a thermaltake 1000 watt power supply and plenty of cooling. I guess I will RMA this one and try again. I played with the fan speeds and with the fan set at 100% it would cool down to about 77c but it went right back up after I slowed the fan.
Wow, even mine idled at ~50.
 
Well after trying it again with no luck I just decided to return it. I thought I might have problems since I bought it from Tiger Direct but it everything went great with TD paying for shipping too. I guess I will wait a while to see if BFG can fix the problems, I am not a fan of ATI. Maybe read up on the EVGA and XFX version.
 
Well after trying it again with no luck I just decided to return it. I thought I might have problems since I bought it from Tiger Direct but it everything went great with TD paying for shipping too. I guess I will wait a while to see if BFG can fix the problems, I am not a fan of ATI. Maybe read up on the EVGA and XFX version.

TD site lists the card as non-returnable, and that everything else needed to go through the manufacturer, did you e-mail/call them to sort it out?

How did you return yours? :confused:

This is what they list:
"TigerDirect.com cannot accept returns of BFG GeForce GTX 280 Video Card - OC Edition, 1GB GDDR3, PCI Express 2.0 x16, SLI Ready, (Dual Link) Dual DVI, HDTV, HDMI Support, VGA Support after delivery has been made.

* All BFG GeForce GTX 280 Video Card - OC Edition, 1GB GDDR3, PCI Express 2.0 x16, SLI Ready, (Dual Link) Dual DVI, HDTV, HDMI Support, VGA Support are covered by the manufacturer warranty which may include repair, or replacement service."
 
Wow Metrocall I did not read that. I went to my order history, clicked on the order number and there was a selection to return item. I could exchange or return for a full refund. After I filled out a few question they gave me a number to call for the RMA#. When I talked to the guy he just asked what was wrong and gave me the RMA#. Then he sent me a UPS delivery label so I got free shipping for the return. I was really surprised and thought I might have a hard time returning it. Tiger Direct seems to have improved in the customer service department. Right now the status reads "BFG GeForce GTX OC 280 1GB PCIe w/Dual Link DVI Status: Authorized for Return" Now I just have to wait on my credit card to be credited.
 
I now have three BFG Geforce GTX 280 OC cards. I've owned a total of four, had to return one thus far. The three I've got seem to be working very well.
 
My last 3 or 4 vidieo cards have been BFG, this is the first one I had to return. I feel like I am cheating by looking at the XFX and EVGA cards.
 
Wow Metrocall I did not read that. I went to my order history, clicked on the order number and there was a selection to return item. I could exchange or return for a full refund. After I filled out a few question they gave me a number to call for the RMA#. When I talked to the guy he just asked what was wrong and gave me the RMA#. Then he sent me a UPS delivery label so I got free shipping for the return. I was really surprised and thought I might have a hard time returning it. Tiger Direct seems to have improved in the customer service department. Right now the status reads "BFG GeForce GTX OC 280 1GB PCIe w/Dual Link DVI Status: Authorized for Return" Now I just have to wait on my credit card to be credited.

That's my luck alright. :eek: "BFG GeForce GTX OC 280 1GB PCIe w/Dual || Returns NOT Accepted"
 
I have a standard eVGA GTX 280. When I first got it going, it idled at ~55C and loaded at ~82C in Crysis. It's stock-clocked, so I can definitely see how an OC variant would hit 100+C in Crysis and etc (TrackMania: Unlimited Forever can also work it a ton at 1920x1200, 8^2 lightmaps, max, etc... kinda wondering why [H] is using Grid instead of TMUF in benchmarks atm 'cause with the right settings TMUF can really push a gpu).

However, my stock fan was set only to 40%. I amped it up to 80% and it idles at 45C and loads only into the high 60's now (generally no higher than 65 or 66 but Crysis and TMUF can occasionally shove it up to 67-68C).
 
However, my stock fan was set only to 40%. I amped it up to 80% and it idles at 45C and loads only into the high 60's now (generally no higher than 65 or 66 but Crysis and TMUF can occasionally shove it up to 67-68C).
But can you be in the same room without going deaf?
 
But can you be in the same room without going deaf?

It's quieter than one of my roommate's 360 despite that my case is an Antec 900 with all the fans set to max ;)

Btw, I've been using the cable from eVGA that converts two 6-pins to a single 8-pin, but Corsair sent me what amounts to a Y cable with two 6-pins and a 6-pin plus 2. The interesting part though is that the 6-pin plus 2 is on one of the arms of the Y as opposed to the base and that seems to indicate it wants me to hook one 6-pin into the PSU and then hook the 6-pin plus 2 to the GTX 280's 8-pin and the other 6-pin to the GTX 280's 6-pin... that doesn't seem right though. But the eVGA cable is "only for temporary usage" and I haven't done much OCing w/it b/c of that. Not quite sure what to do.
 
To be honest it's not the temperature that bothers me(directly) it's the fact that the card starts to stutter and jerk at about 105C even though you may be getting 60+ fps. I had a faulty 8800gt that did that same thing.

I can possibly try to take a video of what I mean if anyone cares to see...
 
Update:

I called BFG last Friday about it. The guy was very helpful and understanding, but seemed tired or something. When I he asked for my zipcode, I said "Uhhhh - 12345 ," and he said, "Oh, you're in Canada?" Apparently he thought that "uh" was an "R" and I had to go back and correct this.

He said I would get an e-mail immediately, and here I am with no e-mail a week later. Poor guy probably misheard my address too.

In the meantime, my card has taken to freezing the video on screen with some crazy fractal color pattern. Then the video cuts out until I power cycle the machine. Sigh.

Looks like the predictions were right on the money. If you have a 280 that breaks 100 C in FurMark, RMA it ASAP!
 
I called them last night and the automated answering sounded like the guy just got outta bed to record it. I found the person that I talked to helpful as well but told me 24-48 hours for the email.
 
Cant believe this is still going on with the 280s. I feel bad for all those that had to return thier cards, even more so for those that this has happened multipls times!!...Talk about MEGA return ratios...:mad:
 
Cant believe this is still going on with the 280s. I feel bad for all those that had to return thier cards, even more so for those that this has happened multipls times!!...Talk about MEGA return ratios...:mad:

They're all OC variants though? Ofc, once again, I haven't had my GTX 280 conk out on me and ofc there's no way I can confirm that the cards that are having these issues are defective or not and I do not even mean to touch on that issue. I just want to reiterate that a stock clocked GTX 280 at stock speeds can already hit 82C in Crysis. It seems logical that even a moderate OC could push that into the 100's where the card itself is going to be damaged via the heat. But anyway, amping the fan speed from 40% to 80%, my max temp has dropped 15C. Once again, I can't say that the low fan speed is the issue or that amping the fanspeed on a GTX 280 OC will help any or that there is any precedent for it having helped, but when/if you get a GTX 280 OC or any GTX 280 at all, imo bump the fan speed up before the card can melt itself.
 
I forced the card to run at 100% fan and at underclocked it to 260ish speeds and it goes from 40-40 to 105C in about 1 minute... It's not just that they are slightly OC'ed. For me or the few I have personally delt with anyway.
 
I forced the card to run at 100% fan and at underclocked it to 260ish speeds and it goes from 40-40 to 105C in about 1 minute... It's not just that they are slightly OC'ed. For me or the few I have personally delt with anyway.

K- No idea wtf's going on then, but thx for clearing that up.
 
That reminds me of one of the 680i SLI problems. I have seen the temperatures of the north bridge just shoot up randomly one day to temperature ranges of 70c to 80c while idle. Even water cooling couldn't change that. The boards that did this always died shortly after this.

I wonder if this is some kind of quality control issue by the semi-conductor manufacturer or what.
 
These issues are exactly why I decided to skip the 280 and just get a 260 instead. A good savings of money, and performance is still great. Perhaps some 260's have this issue as well, but it isn't nearly as common. I have my 260 clocked higher than FTW speeds and in Furmark running for 10h in a 30c room, it never hit above 86c with the fan at auto.
 
These issues are exactly why I decided to skip the 280 and just get a 260 instead. A good savings of money, and performance is still great. Perhaps some 260's have this issue as well, but it isn't nearly as common. I have my 260 clocked higher than FTW speeds and in Furmark running for 10h in a 30c room, it never hit above 86c with the fan at auto.

The whole thing about Geforce GTX 280 failures is overblown. I've got three of them that work fine. I did have one bad one the first week I they came out but after exchanging it I've never had any problems.
 
The whole thing about Geforce GTX 280 failures is overblown. I've got three of them that work fine. I did have one bad one the first week I they came out but after exchanging it I've never had any problems.

Couldn't agree more. I like you had to exchange one of my 280's when I got them right at release in June but the three I have now have been fine for over two months.
 
The whole thing about Geforce GTX 280 failures is overblown. I've got three of them that work fine. I did have one bad one the first week I they came out but after exchanging it I've never had any problems.

Given that mine works very well I'd be inclined to agree, but my exploded 360 and my time at EA QA over the summer and my contact w/just about everyone that has a 360 inspires a modest degree of caution in terms of offering such agreement. Granted, the failure rates of GTX 280's are definitely nowhere near the failure rates of 360's, but still something to consider that even a 50% failure rate brings it down to relative chance as to whether that failure rate affects you or not.
 
So far all the reports of defective 4870s and GTX280s are just BS.
Every card from every generation has had problems it just part of life.
I say if you want a 280 and can afford one dont let a couple ppl with issues change your mind.
Hell if you get a bad card you can always RMA it.
 
So far all the reports of defective 4870s and GTX280s are just BS.
I guess you are master of the universe to be able to know all and see all. Being acutely aware of failure rates for various products and manufactures must necessitate omnipotent awareness and intellect. Since you yourself have obviously not experienced any problems with your products, those who do are obviously fanboys planted here to threaten your dominion. They must be crushed. You must rule with an iron fist to succeed.
 
Status
Not open for further replies.
Back
Top