PDA

View Full Version : EVGA says to return my 7900GT KO for RMA. Do you agree?


John Adams
October 14th 07, 10:03 AM
I have the infamous corrupted screen issue on my 7900GT KO and Vista.
Have also noticed this issue in some games on XP too. But I can alt+tab
in XP games and the screen comes up clean again. I contacted EVGA about
it and they said to do an RMA for a replacement. Thing is that I have
seen people do RMA's over this issue and the replacement 7900GT does it
too so I'm not so sure this is a hardware issue and is really a Nvidia
driver issue with those cards. For one thing, the drivers that come with
Vista don't cause it so that is why I think it is a driver issue. I
could be wrong though. I've already had to do one RMA for the same card
and don't want to go through the hassle of doing another if it doesn't
really fix it. EVGA charges for shipping both ways too. Anyone know for
certain if this is a driver issue or a hardware issue?


Here is a video on youtube about the issue I am talking about.

http://www.youtube.com/watch?v=DZVPZse53Tg

Augustus
October 14th 07, 04:08 PM
"John Adams" > wrote in message
m...
>I have the infamous corrupted screen issue on my 7900GT KO and Vista.
Anyone know for
> certain if this is a driver issue or a hardware issue?

From what I've read, it's a driver and Vista AA issue not a hardware issue.
I suspect you'd be out of a videocard for a long time, then get the exact
same thing when you get it back. There's a discussion about it here, and
what did and didn't work for some...
http://forums.nvidia.com/lofiversion/index.php?t26830.html

deimos[_2_]
October 14th 07, 06:40 PM
John Adams wrote:
> I have the infamous corrupted screen issue on my 7900GT KO and Vista.
> Have also noticed this issue in some games on XP too. But I can alt+tab
> in XP games and the screen comes up clean again. I contacted EVGA about
> it and they said to do an RMA for a replacement. Thing is that I have
> seen people do RMA's over this issue and the replacement 7900GT does it
> too so I'm not so sure this is a hardware issue and is really a Nvidia
> driver issue with those cards. For one thing, the drivers that come with
> Vista don't cause it so that is why I think it is a driver issue. I
> could be wrong though. I've already had to do one RMA for the same card
> and don't want to go through the hassle of doing another if it doesn't
> really fix it. EVGA charges for shipping both ways too. Anyone know for
> certain if this is a driver issue or a hardware issue?
>
>
> Here is a video on youtube about the issue I am talking about.
>
> http://www.youtube.com/watch?v=DZVPZse53Tg

Things to try:

* Disable power management.
* Check voltages and amp ratings on your PSU.
* Use a CRT (bypass EDID mode information sent to the DVI port).
* Disable AA/AF.
* Borrow another card, preferably a different brand and series.
* Update your chipset drivers, make sure no incompatibilities exist.

That's a weird issue you got there, I've never experienced it on any of
my EVGA cards or any of them for that matter. At least the forums are
pretty well documented on it.

John Adams
October 14th 07, 07:57 PM
Augustus wrote:

> From what I've read, it's a driver and Vista AA issue not a hardware issue.
> I suspect you'd be out of a videocard for a long time, then get the exact
> same thing when you get it back. There's a discussion about it here, and
> what did and didn't work for some...
> http://forums.nvidia.com/lofiversion/index.php?t26830.html
>
>
But it's not just a Vista issue. As I said i have seen it happen in a
few games under XP too. Quake4 and CM:Shock Force for example. If it is
purely a driver issue then why doesn't it happen on 8800 cards too?

John Adams
October 14th 07, 08:00 PM
deimos wrote:

> * Use a CRT (bypass EDID mode information sent to the DVI port).

I am using a CRT. How do I bypass EDID mode though?

Mr.E Solved!
October 14th 07, 08:50 PM
John Adams wrote:
> I have the infamous corrupted screen issue on my 7900GT KO and Vista.
> Have also noticed this issue in some games on XP too. But I can alt+tab
> in XP games and the screen comes up clean again. I contacted EVGA about
> it and they said to do an RMA for a replacement. Thing is that I have
> seen people do RMA's over this issue and the replacement 7900GT does it
> too so I'm not so sure this is a hardware issue and is really a Nvidia
> driver issue with those cards. For one thing, the drivers that come with
> Vista don't cause it so that is why I think it is a driver issue. I
> could be wrong though. I've already had to do one RMA for the same card
> and don't want to go through the hassle of doing another if it doesn't
> really fix it. EVGA charges for shipping both ways too. Anyone know for
> certain if this is a driver issue or a hardware issue?
>
>
> Here is a video on youtube about the issue I am talking about.
>
> http://www.youtube.com/watch?v=DZVPZse53Tg

I've never seen anything quite like that and I've seen it all, not that
it is dramatic, just atypical. At first blush it doesn't seem like
hardware failure, since there is motion and you can recover from
it...once a ram chip goes it is gone.
I'd pin the blame on the Forceware drivers first and probably be right.

Mr.E Solved!
October 14th 07, 08:54 PM
John Adams wrote:

> few games under XP too. Quake4 and CM:Shock Force for example. If it is
> purely a driver issue then why doesn't it happen on 8800 cards too?

The architectures of the 7-series cards and 8-series cards are
completely different so that bugs can and do appear distinctly between
them. Even cards of very similar design manifest gpu-specific bugs, no
surprise.

Mr.E Solved!
October 14th 07, 08:57 PM
John Adams wrote:
> deimos wrote:
>
>> * Use a CRT (bypass EDID mode information sent to the DVI port).
>
> I am using a CRT. How do I bypass EDID mode though?

By using a CRT, substituting VESA for EDID. Glad you asked?

First of One[_2_]
October 14th 07, 10:53 PM
Bring the core and memory clocks down to nVidia reference speeds (450/1320)
and see if the corruptions persist. Even though the higher clocks are
"guaranteed" by eVGA, their testing and binning process isn't bulletproof.

--
"War is the continuation of politics by other means.
It can therefore be said that politics is war without
bloodshed while war is politics with bloodshed."

"John Adams" > wrote in message
m...
>I have the infamous corrupted screen issue on my 7900GT KO and Vista. Have
>also noticed this issue in some games on XP too. But I can alt+tab in XP
>games and the screen comes up clean again. I contacted EVGA about it and
>they said to do an RMA for a replacement. Thing is that I have seen people
>do RMA's over this issue and the replacement 7900GT does it too so I'm not
>so sure this is a hardware issue and is really a Nvidia driver issue with
>those cards. For one thing, the drivers that come with Vista don't cause it
>so that is why I think it is a driver issue. I could be wrong though. I've
>already had to do one RMA for the same card and don't want to go through
>the hassle of doing another if it doesn't really fix it. EVGA charges for
>shipping both ways too. Anyone know for certain if this is a driver issue
>or a hardware issue?
>
>
> Here is a video on youtube about the issue I am talking about.
>
> http://www.youtube.com/watch?v=DZVPZse53Tg

John Adams
October 15th 07, 08:03 PM
First of One wrote:
> Bring the core and memory clocks down to nVidia reference speeds (450/1320)
> and see if the corruptions persist. Even though the higher clocks are
> "guaranteed" by eVGA, their testing and binning process isn't bulletproof.
>
Yea, tell me about it. I had to RMA my first 7900GT KO because the ram
couldn't do the factory overclock in deep freeze bench of 3DMark. By
downclocking the ram by 50mhz it was fine though. Last night I
downclocked the ram by 10mhz and haven't seen the issue in the game I
tested it in but am not running Vista on that PC any longer so don't
know for sure if it has fixed the issue. I've read of other's fixing
this issue that way too, some have overclocked by 5-10mhz instead of
downclocking and it still fixed it. That hasn't worked for everyone
though. If it works for me then great and I will modify the bios by
10mhz on the ram if it does but need to do further testing first to make
sure it fixes it. I really don't want to do another RMA with EVGA as
that will put the PC out of commission for 2-3 weeks. No thanks.

John Adams
October 15th 07, 08:10 PM
Mr.E Solved! wrote:

> By using a CRT, substituting VESA for EDID. Glad you asked?

Think I'll not bother as I have seen no one else mention that as a fix.
Also I highly doubt it is a PSU issue as others no longer had the issue
when they put in an 8800 instead. 8800 draws more power than 7900,
right? I have a 600w Seasonic S12 so doubt that is the issue, although I
will not rule it out. I don't have a better PSU to test with though. I'm
testing the ram frequency fix I read about right now and will see how
that goes. Had no issues when I tested it last night but need to test
more to be certain. All I did was downclock the ram on the vid card by
10mhz.

Hawk[_2_]
October 22nd 07, 08:41 PM
John Adams wrote:
> First of One wrote:
>> Bring the core and memory clocks down to nVidia reference speeds
>> (450/1320) and see if the corruptions persist. Even though the higher
>> clocks are "guaranteed" by eVGA, their testing and binning process
>> isn't bulletproof.
>>
> Yea, tell me about it. I had to RMA my first 7900GT KO because the ram
> couldn't do the factory overclock in deep freeze bench of 3DMark. By
> downclocking the ram by 50mhz it was fine though. Last night I
> downclocked the ram by 10mhz and haven't seen the issue in the game I
> tested it in but am not running Vista on that PC any longer so don't
> know for sure if it has fixed the issue. I've read of other's fixing
> this issue that way too, some have overclocked by 5-10mhz instead of
> downclocking and it still fixed it. That hasn't worked for everyone
> though. If it works for me then great and I will modify the bios by
> 10mhz on the ram if it does but need to do further testing first to make
> sure it fixes it. I really don't want to do another RMA with EVGA as
> that will put the PC out of commission for 2-3 weeks. No thanks.

I've done the RMA shuffle with eVGA on 7900 series cards and my
impression is that their general testing process is "weak" to put it kindly.

I would be very surprised if they do any in depth statistical analysis
on their products before deciding on clock speeds for their stuff. From
what I can tell, they don't test large enough samples and generally
don't allow enough margin.

My impression is that if the first few they try work ok....thats how
they ship them. I don't know this for a fact, but that's how it looks
from my perspective. There's no other way to explain the vast number of
cards they've had to replace.

Also, I have seen too many instances of mismatching bios revs (the label
on the card has one rev, the bios reports a different rev, the web site
yet a different rev, etc) to have much confidence that they have their
processes under control. In my view they fall into a chaotic
firefighting mode and essentially lose track of what they are doing.

The clock speeds they choose for some of their products seem too
aggressive compared to the statistical capability of the product itself.



(*>

Mr.E Solved!
October 23rd 07, 01:22 AM
Hawk wrote:

> I would be very surprised if they do any in depth statistical analysis
> on their products before deciding on clock speeds for their stuff. From
> what I can tell, they don't test large enough samples and generally
> don't allow enough margin.
>
> My impression is that if the first few they try work ok....thats how
> they ship them. I don't know this for a fact, but that's how it looks
> from my perspective. There's no other way to explain the vast number of
> cards they've had to replace.
>
> Also, I have seen too many instances of mismatching bios revs (the label
> on the card has one rev, the bios reports a different rev, the web site
> yet a different rev, etc) to have much confidence that they have their
> processes under control. In my view they fall into a chaotic
> firefighting mode and essentially lose track of what they are doing.
>
> The clock speeds they choose for some of their products seem too
> aggressive compared to the statistical capability of the product itself.


It's even worse than that from a technical perspective. eVGA is merely a
VAR, they do none fo those technical things, they don't even do them badly.

The worst fubar by eVGA's overseas BIOS 'technicians' was when they
forgot or didn't care that shader clocks ran faster than spec in certain
sections of the 6800GT chip (nv70 something or other). So as they
cranked up the speed in the BIOS, there was zero chance of the card
actually working once the shader routines were activated.

The good news is the cards are reference so generic model BIOSes will
work and fix a problem like that. You can easily change your PCI ID to
brand it to any major maker so you never have to see the eVGA name again!

gamefixer
October 24th 07, 05:24 AM
On Oct 14, 1:50 pm, "Mr.E Solved!" > wrote:
> John Adams wrote:
> > I have the infamous corrupted screen issue on my 7900GT KO and Vista.
> > Have also noticed this issue in some games on XP too. But I can alt+tab
> > in XP games and the screen comes up clean again. I contacted EVGA about
> > it and they said to do an RMA for a replacement. Thing is that I have
> > seen people do RMA's over this issue and the replacement 7900GT does it
> > too so I'm not so sure this is a hardware issue and is really a Nvidia
> > driver issue with those cards. For one thing, the drivers that come with
> > Vista don't cause it so that is why I think it is a driver issue. I
> > could be wrong though. I've already had to do one RMA for the same card
> > and don't want to go through the hassle of doing another if it doesn't
> > really fix it. EVGA charges for shipping both ways too. Anyone know for
> > certain if this is a driver issue or a hardware issue?
>
> > Here is a video on youtube about the issue I am talking about.
>
> >http://www.youtube.com/watch?v=DZVPZse53Tg
>
> I've never seen anything quite like that and I've seen it all, not that
> it is dramatic, just atypical. At first blush it doesn't seem like
> hardware failure, since there is motion and you can recover from
> it...once a ram chip goes it is gone.
> I'd pin the blame on the Forceware drivers first and probably be right.- Hide quoted text -
>
> - Show quoted text -

Ugh... After all the "riping a new asshole" you did to me in my
original thread about this... Now you agree that its a driver
issue? :-)

Glad to have you on my side of the street! I guess I should have
posted a Youtube video! LOL!

Matt

gamefixer
October 24th 07, 03:50 PM
On Oct 14, 3:03 am, John Adams > wrote:
> I have the infamous corrupted screen issue on my 7900GT KO and Vista.
> Have also noticed this issue in some games on XP too. But I can alt+tab
> in XP games and the screen comes up clean again. I contacted EVGA about
> it and they said to do an RMA for a replacement. Thing is that I have
> seen people do RMA's over this issue and the replacement 7900GT does it
> too so I'm not so sure this is a hardware issue and is really a Nvidia
> driver issue with those cards. For one thing, the drivers that come with
> Vista don't cause it so that is why I think it is a driver issue. I
> could be wrong though. I've already had to do one RMA for the same card
> and don't want to go through the hassle of doing another if it doesn't
> really fix it. EVGA charges for shipping both ways too. Anyone know for
> certain if this is a driver issue or a hardware issue?
>
> Here is a video on youtube about the issue I am talking about.
>
> http://www.youtube.com/watch?v=DZVPZse53Tg

I started a thread a while back that came from this exact problem. I
ended up RMA'ing the 7900GT with eVGA but the replacement still does
this.

I personally think its the drivers too.... If Nvidia would take a
little time and look into this issue I'm sure they would figure it
out.

Matt