To RMA or not to RMA: eVGA 6800GT

P

particle9

Weaksauce
Joined
Jun 30, 2004
Messages
103
Okay so here is the scoop.
I have tried four different drivers, finally sticking to 65.73 as that gives me the fewest crashes. I have tried all kinds of tweaks to my BIOS (motherboard, not graphics card). I have had my apeture setting at every different level. I have enabled/disabled blitting, I have turned on/off v-sync, I have enabled fastwrites and disabled them. I have checked voltages, I have done everything I can possibly think to do to get this card to stop crashing to desktop or rebooting my system (blue screen of death) for three weeks. Nothing will stop this thing! It crashes in Rome:Total War, Unreal2004 demo, Call of Duty, Farcry, Postal2, and Wolfenstein Enemy Territory. Even counterstrike sometimes! For some reason Doom3 doesn't crash hardly ever though. As far as I can tell the card is working fine! Fan spins, temps are normal, running stress tests produce no artifacts, work for long periods of time. What could be doing this?
I have been getting all kinds of errors, pfn table corrupt once, irq not less or equal, NV4_mini.sys crash, nv4_disp.dll crash! Aggh! I can't take it anymore lol. Someone give me their opinion. RMA this seemingly great card that quits every game I throw at it for apparently no reason and at totally random times, or keep trying to get it working through drivers?
 
have you tried not overclocking? You know, not every card CAN overclock as high, and yours is like 350mhz stock, and you're pushing it at 400mhz. Try lowering it back and leaving fast-writes off. Then test and tell us if anything gets better
 
If you have the same problems with the stock clock speeds, then I would RMA that thing for sure. Also make sure that you haven't done something to void the chance of it being returned.
 
Yeah, sorry about that, I don't have it overclocked at all. That is just what I had tested it at before I got games and realized none of them would run even at stock for more than a few minutes without locking up. I also forgot to mention the sims2 doesn't work without crashing to the desktop either.

Yeah I'm about at my wits end with this thing. I seriously don't know what else I can do. It runs rthdibl fine... it isn't too hot, I've tried everything! Why will it run 3dmark and rthdibl without incident and get good scores and then totally screw up in games? I could have sworn it was drivers but no, because one driver has to work on something right? I was able to make CoD work for longer stretches of time (an hour at most) without crashing to the desktop with 65.73 drivers, and those also let me play CS for an hour before crashing. But that is hardly acceptable. Again, I don't understand why Doom3 would work on the thing for 3 hours before crashing either. All that led me to believe it had to be a driver/software issue. But four clean installs later, five different drivers, five different games, and more settings than you can imagine (I've even tried playing with 4x agp) the thing is not working!

The only thing that I did that I think might be frowned upon was I took the serial sticker off the card and saved it. Don't ask me why, I just didn't like looking at it lol. But I didn't see anywhere where it said that would void a warranty.

So yeah I think I have to RMA this thing, which blows because I don't have another graphic card while it is out. So a couple questions:
Do you think I should RMA this thing?
If I should, since it has been more than a month since I got it from newegg do I do the RMA through newegg or eVGA?
How long does this process usually take?

Yes I know I should figure out the second two by myself but if it isn't too much trouble to chime in feel free.

Thanks for the replies so far.
particle9
 
Out of curiosity, did you upgrade from another nvidia card or an ati card?

If ATI, did you run driver cleaner or fresh format?

Could be the problem.
 
Nope, fresh build. All components are new, and like I said I did a fresh format install of XP about four times now.
 
If it's a new build, then it could be anything, not just the videocard. Video card problems usually manifest themselves as artifacting or something similar. In my experience, crashing to the desktop has quite often been caused by bad RAM.

If you haven't already, try running memtest.
http://www.memtest.org/#downiso

Just make a floppy, and boot it. Memtest will start running automatically, and will probably take an hour or so to finish.
 
Yes I have run memtest and memtest + and both passed with no errors at all. So not memory related.

I would try to use the card in another computer but I don't have one, sorry.
 
Sounds like what started happening to my card a few weeks after I got it. I sent it back for a replacement yesterday. Test the system with another vid card if you have one lying around, but if everything else is ok then chances are it's a bad card.

If you have a spare hard drive or partition handy, you could try a fresh windows installation, but that's pretty time consuming. If you have your last system's hard drive lying around (I have a stack of drives starting at 400 meg going all the way to 20 gig on my shelf that I keep around after upgrading just in case...) then it probably wouldn't hurt to give it a shot.
 
flenser said:
Sounds like what started happening to my card a few weeks after I got it. I sent it back for a replacement yesterday. Test the system with another vid card if you have one lying around, but if everything else is ok then chances are it's a bad card.

If you have a spare hard drive or partition handy, you could try a fresh windows installation, but that's pretty time consuming. If you have your last system's hard drive lying around (I have a stack of drives starting at 400 meg going all the way to 20 gig on my shelf that I keep around after upgrading just in case...) then it probably wouldn't hurt to give it a shot.
Click to expand...

Yeah, I've done a clean install, so a new harddrive isn't going to fix anything and this drive is running perfectly anyway so it isn't a problem with the SATA drive. Like I said above, I don't have another card which is why it sucks so bad to RMA this one, but at this point I feel I have tried everything and tested every single component of the system and it has to be the card. The thing I don't get is what is wrong with the card that it would work fine in 2d great and work in 3d great with no artifacting and then all of a sudden crash to desktop all the time. That sooo seems like a driver/software problem and yet it can't be because other people aren't having this problem and I've done a million clean installs of everything from the os up. IT BOOGLES THE MIND!
 
My only other suggestion, then, is to send it back.

eVGA doesn't have a very good track record with me; I had an eVGA GeForce 2 MX400 32MB and it was nothing but trouble.
 
I concur. I got my RMA approved from newegg. I'll be sending that back to them fedex standard overnight tommorow and hopefully I'll get a new one by next friday or the following monday. I really can't fault eVGA as they have been great on the support side and have a very flexible upgrade program so as long as I get a card that actually WORKS I'll be very happy with them. I'll keep everyone posted on my RMA experience. (Like you care.... :p )
 
RMA after you have OC'd for ANY reason = bad in my book. Even if it was for just ONE second, you ran the product out of spec.

Had you not overclocked at all - I would endorse your RMA as the product sounds faulty. :)
 
it could be your cpu i had the same probelms i think its your cpu and memory i had the exact same problems with 3500
 
I appreciate the sentiment about the oc, however I will take it as a lesson learned if anything, not a financial hit on my bank statement. These cards cost at least three times currently what they cost to make so I don't feel like I'm ripping off anybody. Sorry, just the way I feel about it. I also ran it out of spec for about 10 minutes as I auto-detected oc at 420/1120 and then put it on 400/1100 and decided why am I doing this when Doom3 is running fine at stock. If I ran it oc'd for more than 20 minutes I would maybe begin to feel worse about it, but maybe I'm just a douche.

I thought it could be the cpu too, but the thing is that it runs without error on cpuburn. The memory also runs flawlessly. I think if anything it is an error with the memory on the graphic card, but then again I don't really know. The other thing I was wondering is if it could be the motherboard. But seriously, how do I test these components any more than I am? The problem only occurs in 3d games, only after a few minutes, not related to any cpu load (as I checked task manager when it crashed one time and load was not even 50%). I tested every component to the best of my ability and spent weeks, hours every night trying to cover all my bases. I believe I have. If the new card does the same thing then I'm really going to go ape, but seeing as how it only seems to occur during high 3d use, and not high anything else use, and the mem tests fine, I don't think it could be anything else.
 
it ran the cpu stress test and memory test all postive so it think its still your cpu when i switched cpu and memory worked right what it say when it crashes i am 90% postive its your cpu


because when i had that problem it was getting postive marks on memory and cpu

but when i switched the meomory and cpu everything worked great rma your cpu and memory if you can :D
 
How about your PSU? Same thig happened to me, I killed 3 psus Antec 550,430 and Thermaltake480 w/ MSI NX6800U. Finally settled w/ ENERMAX 660W and no more crashes when loading 3d game.
 
I'm sorry, I'm really glad you are trying to help me, but I am not going to RMA my cpu and memory when everything about them seems fine unless you can explain to me why I should in a way that is more than, "something similar happened to me, I tested everything and it was fine, and I got them replaced and it worked great". Sorry, not going to happen. I don't know what about a 3d game would cause the cpu and memory to exhibit problems that would not be evident in other uses. If it were the cpu and memory testing fine and being the problem than I would consider that a total freaking weird scenerio that I would persue AFTER I RMA'd the video card. I mean, I think that is only logical. Again I don't mean to sound patronizing or condecending, but do you really expect me to RMA my cpu and memory based on your one line saying that is the problem with no explaination of how that could be? What makes you think it is the cpu and ram over the videocard? I need more convincing. Thanks for the replies though. :)
 
At least you are taking into consideration the fact you overclocked. Some just say F it, and RMA even though they were clearly at fault.

Think about this. Your car has a red line. If you exceed it just once it can have detrimental effects. Some of these effects can be seen immediately, others take time. Then again you can exceed red line and it never ever hurt it. It is honestly just a YMMV thing. :eek:

To add to this I will say that the first 6800 GT I owned was an eVGA. The memory went bad at stock speeds and I RMA'd it. I have been happy with my BFG ever since. :)
 
I'm all for legitimate RMA.... but if what he says is true you guys are barking up the wrong tree. I wasn't by his side when he did this stuff, but the the chances he borked his card doing that are like .0000001% In all my volt modding and freezing experience, I've yet to break anything. You guys act like these things are so ridiculously fragile. Imagine throwing in a new cpu, how long do you go before you try overclocking it? Imagine it up and dying without even a change in volts? Don't have to worry because it doesnt happen, unless the chips was probably heading for death anyway. It crashes... that's it. You guys think he really caused some kind of physical problem with the transistors in his chip doing this?

I've told members who's cards failed after switching sinks and stuff that they should do the right thing.... this is crazy though. Anyone who thinks the guy borked his card either doesnt believe the poster's story or doesnt know fiddle sticks about this stuff. Technicalities.... Guess you guys wouldnt RMA your DOA AMD cpu because you didnt use the included thermal pad right? BS. I think the man should look at himself, and ask if he is doing the right thing.
 
because i had the exact same problems with my cpu and ram and i have a BFG 6800gt i thought it was the card turns out it was the cpu and ram belive me i had a 3500 and ocz pc 3200 1gig plantuim the good stuff
 
texuspete00 said:
I'm all for legitimate RMA.... but if what he says is true you guys are barking up the wrong tree. I wasn't by his side when he did this stuff, but the the chances he borked his card doing that are like .0000001% In all my volt modding and freezing experience, I've yet to break anything. You guys act like these things are so ridiculously fragile. Imagine throwing in a new cpu, how long do you go before you try overclocking it? Imagine it up and dying without even a change in volts? Don't have to worry because it doesnt happen, unless the chips was probably heading for death anyway. It crashes... that's it. You guys think he really caused some kind of physical problem with the transistors in his chip doing this?

I've told members who's cards failed after switching sinks and stuff that they should do the right thing.... this is crazy though. Anyone who thinks the guy borked his card either doesnt believe the poster's story or doesnt know fiddle sticks about this stuff. Technicalities.... Guess you guys wouldnt RMA your DOA AMD cpu because you didnt use the included thermal pad right? BS. I think the man should look at himself, and ask if he is doing the right thing.
Click to expand...

Looks like someone has returned a few things they've broken in their time. ;)

If the manufacturer had a way to tell that he overclocked his card at ANY point for ANY amount of time...guess what? They would cancel his RMA or at least have the grounds to do so.
 
texuspete00 said:
I'm all for legitimate RMA.... but if what he says is true you guys are barking up the wrong tree. I wasn't by his side when he did this stuff, but the the chances he borked his card doing that are like .0000001% In all my volt modding and freezing experience, I've yet to break anything. You guys act like these things are so ridiculously fragile. Imagine throwing in a new cpu, how long do you go before you try overclocking it? Imagine it up and dying without even a change in volts? Don't have to worry because it doesnt happen, unless the chips was probably heading for death anyway. It crashes... that's it. You guys think he really caused some kind of physical problem with the transistors in his chip doing this?

I've told members who's cards failed after switching sinks and stuff that they should do the right thing.... this is crazy though. Anyone who thinks the guy borked his card either doesnt believe the poster's story or doesnt know fiddle sticks about this stuff. Technicalities.... Guess you guys wouldnt RMA your DOA AMD cpu because you didnt use the included thermal pad right? BS. I think the man should look at himself, and ask if he is doing the right thing.
Click to expand...


Thanks for the vote of confidence. This is the first system I have ever built and believe me, I have not done anything insane to damage the thing with overclocking. I RMA'd this morning. I would also like to say that on a personal note I see what you are saying with the overclocking breaking things and people sending them back and that not being honest. I also have to say that I don't think that RMA'ing a computer product will result in a loss for these companies. Let me explain why I say that, and you can prove me wrong if I am so. First, the amount of hardware that is RMA'd is very very small. This in and of itself makes me believe that they can not offset profit margins by any significant percentage. Secondly, with these enthusiast graphic cards the price to manufacture is magnitudes smaller than what they sell them for. Let me explain further. The cards themselves cost 1/4 - 1/3 of what they sell them for. The high price is justified by the expense of "inventing" and fabricating one to make a model of. Once the card hits the production line it itself is worth very little, the hard work and intellegence that went into designing and producing the first one was magnitudes of expense larger than any of the cards after it. If I pay $400 for a card that cost $140 to make, and maybe hundreds of millions to invent, I have done right by that company. If I get a defective one, or even if someone broke theirs willfully and sent it back, and get it replaced that is costing the company that made it in my mind very little. Think about it, nvidia sells the card marked up to resellers who then mark those markups up three or four times that. One could probably get two or three replacement cards at such a price before even bringing the profit margin on a single sale to zero. Now if everyone was sending them back, and sending them back multiple times, this would be a huge problem. However because A) Few people RMA since most units work and even more people don't screw them up willfully and B) the cost of manufacture is low I don't believe that RMA'ing anything can be construed as a significant "theft" from a company. It is like buying a game disc and then willfully scratching it so it no longer works after paying $50 dollars for it. Is it really going to kill that developer or publisher or game merchant to replace a $.02cent CD-ROM? No. That is just my opinion.
 
breadman said:
Looks like someone has returned a few things they've broken in their time. ;)

If the manufacturer had a way to tell that he overclocked his card at ANY point for ANY amount of time...guess what? They would cancel his RMA or at least have the grounds to do so.
Click to expand...

Are all you nay-sayers this skeptical of everything? I told you I never broke a damn thing! Why would I have to RMA it. I've been lucky enough to never get anything already borked. Yeah you're right I'm full of shit, lots of reasons for me to lie about it. :rolleyes:

There are morals and then there are technicalities. What if you're asus board really runs 201 fsb? Is it ok to return a borked cpu then. Hate to get this rediculous with an example but you guys can't seem to tell the difference between what is reasonable and what is not. I like the AMD paste example... fits in again... if they could tell like you said, they'd have the right to decline your RMA. Well they can't so whatever. If they wanted to implement an OC detection I'm sure they could too.

Particle9- there may be some holes in your theory, like the cost of GDDR3. They didnt invent that so the markup is not there's to control. I dont know that we can begin to speculate the cost of it. Others will use this to criticize you. They've been lucky enough to not get anything borked and probably haven't put hardware through the wringer enough to know what these things can take. Point is you didnt break it. Sleep well.
 
texuspete00 said:
Particle9- there may be some holes in your theory, like the cost of GDDR3. They didnt invent that so the markup is not there's to control. I dont know that we can begin to speculate the cost of it. Others will use this to criticize you. They've been lucky enough to not get anything borked and probably haven't put hardware through the wringer enough to know what these things can take. Point is you didnt break it. Sleep well.
Click to expand...

Oh I'm sleeping very well, provided I get my new card ASAP. I see your point about the GDDR3 but the fact is I left a pretty wide margin for my cost analysis. If people want to nitpick it then that is their perogative. The fact remains the same, these cards pay for themselves many times over, that is a simple fact of the business model. No matter how you look at it these cards cost at the most a third of what they sell for to manufacture. No computer hardware manufacturer is ever going to sell something for a few dollars more than it costs to make. Not going to happen. I was just trying to explain my rationalization to the demon seekers. I think you are spot on with your point about reasonable RMA's. If I wanted to test how well my card ran without the heatsink attached then I would be a total tard to RMA it and expect to be taken care of even though they probably would. Those kinds of people suck, but even those types do not affect these companies profit margins to a percentage point. If this was really the problem some want to make it out to be then we would see much more stringent restrictions on returns.
 
texuspete00 said:
Are all you nay-sayers this skeptical of everything? I told you I never broke a damn thing! Why would I have to RMA it. I've been lucky enough to never get anything already borked. Yeah you're right I'm full of shit, lots of reasons for me to lie about it. :rolleyes:

There are morals and then there are technicalities. What if you're asus board really runs 201 fsb? Is it ok to return a borked cpu then. Hate to get this rediculous with an example but you guys can't seem to tell the difference between what is reasonable and what is not. I like the AMD paste example... fits in again... if they could tell like you said, they'd have the right to decline your RMA. Well they can't so whatever. If they wanted to implement an OC detection I'm sure they could too.
Click to expand...

Geez, I thought they made the winky face for a reason. :rolleyes:

Now as far as AMD with the paste difference. Let's run with it. You buy a retail CPU with their HSF combo that has their gook on it. You decide to scrape off the gook and use the same CPU with a different paste. The cpu fries for whatever reason. You call in for an RMA and they say okay, send the HSF too. WOOPS, looks like you're caught with your pants down on that one.

Of course if you knew ahead of time that they were going to ask for their HSF with the gook showing use, there would be people who would work around that as well.
 
texuspete00 said:
Particle9- there may be some holes in your theory, like the cost of GDDR3. They didnt invent that so the markup is not there's to control. I dont know that we can begin to speculate the cost of it. Others will use this to criticize you. They've been lucky enough to not get anything borked and probably haven't put hardware through the wringer enough to know what these things can take. Point is you didnt break it. Sleep well.
Click to expand...


If he had put his hardware through the wringer enough before overclocking it, he may have realized that he had a bad card. (If that turns out to be the problem.) In the end, it doesn't matter whether or not he damaged the card by overclocking it. The real point is, if he is not willing to deal with the fact that he voided his warranty by overclocking his card, then he should either take more caution with his equipment or be prepared to buy a new one if something goes wrong.
 
OurZar said:
If he had put his hardware through the wringer enough before overclocking it, he may have realized that he had a bad card. (If that turns out to be the problem.) In the end, it doesn't matter whether or not he damaged the card by overclocking it. The real point is, if he is not willing to deal with the fact that he voided his warranty by overclocking his card, then he should either take more caution with his equipment or be prepared to buy a new one if something goes wrong.
Click to expand...

That I can agree with, but you damn [H]'s got me so worked up to oc everything lol. When I get my new card I do not intend to oc and I will definitely make sure it works to spec before I do if I so choose later. If I then actually break the thing oc'ing it I will not complain and I will get a new one.
 
Sounds like what mine does when i OC it. Around the power connector it will get real hot and when it gets to hot it will reboot... When i run my fans at full speed it reduces the amount of time it takes to do it or even sometimes doesn't at all. i guess this is why ultras have 2 power slots rather than one.

I'd imagine if i didn't have a cheap case with some loud ass fans where i'd actually run them it wouldn't be to much of a problem but oh well. :)
 
The best thing AMD, INTEL, NVDA, ATTYT can do to us is LOCKED (to Standard specs) ALL of their products, So, no more OC and in turn no RMA
 
breadman said:
Now as far as AMD with the paste difference. Let's run with it. You buy a retail CPU with their HSF combo that has their goop on it. You decide to scrape off the goop and use the same CPU with a different paste. The cpu fries for whatever reason. You call in for an RMA and they say okay, send the HSF too. WOOPS, looks like you're caught with your pants down on that one.

Of course if you knew ahead of time that they were going to ask for their HSF with the gook showing use, there would be people who would work around that as well.
Click to expand...

AMD has a list of specific thermal pastes that they approve for use other than the stock thermal goop that comes with it. They realize that the chip will probably be taken out of a motherboard more than once in its lifetime.

I RMA'd a chip to AMD with a cleaned off heatsink... they didn't care.
 
I've massively overclocked tons of processors and video cards, pushing them to their crashing limits, and running them just a few mhz before that point for years on end. My CPUs idle at 60C, my videocards idle at 70C, my RAM will burn you if you touch it, and in all my experience, I've never killed a single piece of hardware (I've had a few hard drives die over the years, but those weren't related to overclocking). Maybe I'm just lucky (knock on wood) but in all my experience, damaging hardware by overclocking or increasing the voltage is next to impossible unless you run something without a heatsink and fan on it.
 
Hi again, well here is the update.
The RMA went swimmingly, they turned around my card and had it back to me a week after I sent it back (newegg). The new card works great however I just figured out that my graphic card was not actually at fault. The new one locked up the same way the old one had (although this one is noticably louder and the fan seems to be pushing more air). So I sighed and opened up a few sensor monitors. Now maybe I'm a retard, maybe I'm unlucky, but anyway after running CPU burn for 3 and a half minutes I noticed something awful. My CPU hit 50C and then THROTTLED DOWN! The heat from the cpu dropped in half for about 20 seconds and then shot straight back up to 50C. Well needless to say I took off the heatsink, replaced the AS5 and reseated and now everything works beatifully. No lockups in anything, ever. It's amazing. So I guess it was my CPU after all, but I swear the thing was not getting this hot before and I have no clue what the hell was going on but glad it did freak out while I was paying close attention.
RTWar now works, U2k4 works, everything works great. The new card is a bit cooler (about a degree). My only complaint is that I sent back my FarCry game that came for free and they didn't send me one back. WTF? So I'll call about that on Monday and see if they will send it back or give me another one (most likely I imagine they will not. Anyway, everything works now like I said. Thanks for all the help.
 
that's awesome...glad it worked out

they should send you farcy, i mean, technically its not like you got a new "crippled" card, and you sort of payed for it, so i would assume they'd give it to....well, good luck with ur card!
 
glad to hear it wasn't the card - love my eVGA GT so far. $360 at newegg was a steal. :D

congrats
 
You must log in or register to reply here.
Back
Top