Who's panicked by all of the 7900 series problems lately? Is there a DESIGN FLAW?

i dont think its a heat pronlem. at least not "only" heat. my cards idle at around 52c and regulary hit 70 any time i game, without issues. this sound like a problem that the 9800 series of cards had, where a pipe or 2 would break, causing artifacts and checkerboarding effect.
 
Couple days ago, my BFG 7900GT started actcing up on me after a month of some kickass gaming on it. I notice when I game, they temps were really high, so I remount the heatsink with some fresh grease. Everything seems ok now.

But the thing that gets me is, everyone is pushing the 7900GT's to some high overclocks. My BFG cant even go 5mhz without locking up or artifacting on me. Its not a heat issue as I checked that, it just wont overclock that good. I going to put a Zalman on it though and try to push it again, but I doubt it will work.
 
My 7900 GTX was bad out of the box, unless you call playing FEAR or COD2 or SS2 "stress testing". Artifacts and wackiness right from the start. No OC, no custom anything....
 
My first EVGA 7900 GTX was bad outta the box, then this morning one of my XFX 7900 GTX XXX cards took a dump, Funny thing is i have never overclocked any onf these cards, yes they come factory overclocked but i expect more then 2 months use out of casual gameplay. It really pisses me off, Im 31 years old and this is the first video card problem i have had and its been in the 7 Series cards. I sold 2 7800 GTX 512 meg cards and i wish i didnt, those were flawless.
 
Just to follow up my first post in this thread.

My third card (replacement from BFG) is now bad. Artifacting like crazy in BF2 and crashing my os after 3dmark 05 & 06.

Initially BFG Tech was emailing me back twice a day and now it's been 2 days since I've heard from them on this latest card. Since they've been so responsive in the past I can only believe that they are formulating what their response is to this issue. I sent them many links to various hardware sites about this issue and I let them know that EVGA is even tinkering with the bios to solve the issue.
 
tim-x said:
even if 20 people post that they got broken 7900GTs here on this forum thread. For every 1 person that got a broken one there are probably 100 people that got one that's fine or more. It's just the same as the people who bought ipod nanos and cried when they scratched them. My ipod nano is fine and my 7900GT is fine as well.

edit: however i would go with the KO version if you can. The fan covers the memory at least.
Click to expand...

agreed, your never going to see threads saying "my 7900 works great!" I didn't know there were problems with the 7900GT, my friend and I bought ours in march and haven't had one problem, the only way I knew there were problems is by checking this section of the forum today lol
 
KompressorV12 said:
agreed, your never going to see threads saying "my 7900 works great!" I didn't know there were problems with the 7900GT, my friend and I bought ours in march and haven't had one problem, the only way I knew there were problems is by checking this section of the forum today lol
Click to expand...


agreed. got my 7900gt on the 10th of march now 2 months and 10 days later the card is working absolutely fine.
 
revenant said:
Have you guys with bad or marginal cards tried the 87.25 drivers yet? I heard of a couple peeps using them and they helped.. they seemed to fix issues for me which I could have sworn were artifacting.. so for what it's worth:

XP 32bit version: http://www.3dchipset.com/drivers/nvidia/beta/nt5/index.php

hope they help!
Click to expand...

Those are the drivers I'm using right now, and I've tried all different sets of drivers. :( No dice.
 
like the "bad capicitors" fiasco a few years back with caps bulging/blowing up on mobo's , maybe we have a similar issue here with a bad batch of voltage regulator or something of the sort that are getting put on these new cards


[F]old|[H]ard
 
blegh

Just got mine last thursday from eVGA's step-up program (I used to have a 7800gt) I did some overclocking (550/799), but with a Zalman VF700LED and RAMsinks. No voltmods, although I did check the volts with a multimeter today and they seem right where they should be.
Now it seems to be dying. Even if I have it at stock speeds (450/660), FEAR and 3dMark06 will crap out, getting a very low score on the benchmark, and, more often than not, will crash. I haven't tried any other games, but that's what I'm gonna do right now... :(
 
I'm now convinced that the reason mine is dying is because I'm being paid back by the hardware gods for starting this thread. :p

I am in the process of RMAing with eVGA (just need to send in a CC authorization form) but in the meantime I've been experimenting a little. I've found out that if I underclock the card to stock nVidia specified GT speeds (450/1320) it runs pretty stable. I can run the memory at up to 1400mhz (DDR) but any higher an artifacts and wild polys like crazy..


Is anyone out there familliar with eVGA's cross shipping policy? I have to send in a CC authorization form, but I'm not clear on how they handle this. Are they going to immediately put a hold on my card for the price of the card during cross shipping or what?
 
Blue Falcon said:
Is anyone out there familliar with eVGA's cross shipping policy? I have to send in a CC authorization form, but I'm not clear on how they handle this. Are they going to immediately put a hold on my card for the price of the card during cross shipping or what?
Click to expand...

Haven't tried one before, but I heard that EVGA just hold your CC info and only after a certain period of time if you don't send in your dead card, then they will charge your CC. Otherwise, no hold amount of any sort.
 
I got my Step Up 7900GTX EGS back in mid April and it's still running great. I did put a ramsink on the memory regulator voltage IC as when I tried to feel it, it was scorching. I did this in the very early stage of using my card. So, that may be why I never have any problem.
 
i certainly think it's a good idea to put a small heatsink or two on those voltage regulators in the bottom corner. I bought some coolermaster copper ramsinks off a FS/FT thread over at Anandtech and one of them fits all three pretty well. Can't hurt anyway :)
 
revenant said:
LOL, I just vmodded mine also, I put the ink on tape with "easy to break away" tabs for quick removal if I wanted to.. anyways, I got a 9494 also. pretty funny! running at 635/1700 now

http://service.futuremark.com/compare?3dm06=307132
Click to expand...

our machines are very similar, except you have more l2 cache than me. i did the same with my vmod, i masked off 2 straight lines with celo tape, and i have a 1/4" wide piece of tape with the defroster paint on it joining the 2 lines so i could just pull it off to go back to stock.
 
Well, I have tried the new bios sent to me from Mario and it did not correct the problem at all. I then called eVGA tech support and a rep Jacob sent me another bios file. Tried that one and still did not work. I still have the crazy artifcating happening. First thing I did after the bios flash was to run 3dmark06. On the scene where the blimp passes over the huge dragonfish...the screen starts to shine in a very bad green color. Then when it gets to the winter scenes, everything just ******s out and you can barely see what is even happening. When I load of world of warcraft and fear, I still get super crazy spike / cone like artifacts shooting out of everywhere.

For you guys that have RMA'd already...have the new cards worked without a hitch?? I want to know what I am in for here. I did not spend 600 bucks on a top end video card to have to keep replacing it!

I called tech support again today and was told they were busy and to give a call back number.

Drep
 
Running two XFX 7900 GTX XXX cards in SLI since the day the 7900's came out, no problems, simply pure joy.
 
For you guys that have RMA'd already...have the new cards worked without a hitch??
Click to expand...

Nope. I'm on my third card right now and am getting ready to RMA it. I'm going to sell the brand new unopened card on eBay to recover my losses (since I don't believe I can get a refund any more after doing my RMA directly with BFG Tech) and buy an x1900xtx.

I'll be in the hole about a hundred bucks (costs of shipping back and the loss of sales tax I paid, etc.) and that's about as far as I want to go.

I don't want to worry about my card crapping out any more. I just want it to work.
 
Sovereign said:
Ya I hope it doesn't hit me, I'm getting dual 7900GTXs STOCK CLOCKED and if those crap out on me......*shakes fist*
Click to expand...

I sold 2 evga 7800gtx ko to upgrade to the 7900gtx for the faster speeds and the extra 256mb of memory, got them last monday from monarch and they both are crapping out with the color dots,tearing,major artifacting, and the blinking screen when exiting d3d games requiring a hard reboot :rolleyes: I requested an RMA yesterday but at this moment I'm not real optimistic because i have read about others being on their 3rd and 4th RMAs.. how can they send out defective RMAs?

This has to be killing their bottom line with all the bad cards we keep seeing.
 
For those that haven't seen any sort of official reply, here's what an evga representative put up on their forums yesterday:

[font=Verdana, Arial, Helvetica]
EVGA would like to thank its forums members for posting system specs and the details about the issues they are having. EVGA takes these matters very seriously and is looking into all post to extract the data from them.

Below are some key facts regarding this issue that we want to share with everyone:
Click to expand...
[/font]

  • [font=Verdana, Arial, Helvetica][/font][font=Verdana, Arial, Helvetica]
    [*] EVGA’s 7900GT Series total RMA rate is from 0.04% to 1.9% per different SKU’s (which includes cards that had no problems found)
    [*]19% of the users who requested a 2nd time RMA out of the all RMA still experienced either the same issue or new issues unrelated to the VGA card itself.
    [*]EVGA will continue to perform better product quality control ensuring a reliable and stable product. At the same time we are also improving our technical support skills marking sure to provide a better customer service experience with EVGA’s product.
    [*]EVGA has observed some users pushing the card to an unreachable amazing clock speeds and causing the GPU and/or memory to fail permanently and this has gotten our attention.
    [*]All over-clocking exceeding our product preset clock/memory speeds out of the box is NOT recommended. Some users may be able to over-clock with higher speeds, but this may lead to hidden issues like artifacting or random lock ups. That is why we have created several different clocked products for 7900 series.
    [*]We encourage EVGA customers who have problems to contact our technical support staff first to get the problem fixed ASAP. Our support number is 888-880-3842 24 hours a day 7 days a week.    [/font]
[font=Verdana, Arial, Helvetica]
Click to expand...
[/font]
 [font=Verdana, Arial, Helvetica]
[/font]
 [font=Verdana, Arial, Helvetica]At least one company out there acknowledges there's some sort of problem. The post is here: http://www.evga.com/community/messageboard/topic.asp?TOPIC_ID=15412 [/font]
 [font=Verdana, Arial, Helvetica]
[/font]
 
It doesn't really look like they acknowledged the problem to me, more like "Nothing to see here, these aren't the droids you're looking for... move along".

Still at least we know they've taken notice of the issue... I am certainly hoping for the best. I've got an RMA in progress for my 7900GT and all I can hope is that eventually I'll get a good one.

I still tend to believe that there is a serious design flaw especially in light of the recent discoveries with the voltage regulation IC's that are heating up more than any other component on the cards (save the GPU itself). Those photos taken with the FLIR thermal camera are quite impressive (and worrysome).

[EDIT: Added FLIR photos]

Here are photos taken of someones XFX 7900GT card using a FLIR heat-sensitive camera. The red and white hot area in the images is focused in the area of the ICs said by many to control memory voltage. These are the ones everyone has been putting aftermarket ramsinks on to try and cool off.

videogroot.JPG


xfx%20probleem1.JPG


xfx%20probleem.JPG


NOTE: These are not my photos, I simply uploaded them to an alternate source so as not to steal bandwidth from someone else.
 
Blue Falcon said:
It doesn't really look like they acknowledged the problem to me, more like "Nothing to see here, these aren't the droids you're looking for... move along".
Click to expand...

best quote ever
 
legrand said:
Yes, I've seen those pictures as well, but I'm not convinced that's really the issue. Those chips are rated to 155 degrees celsius, well past the 80 shown in the picture: http://www.datasheetcatalog.com/datasheets_pdf/I/R/F/7/IRF7821.shtml
Click to expand...

Where do you see this chip on the area of the board in question? :confused:

Anyway I highly doubt that ANY chip on there is rated to operate at up to 150C, that's insane. That might be a max storage/inert temperature or something...





Oh and I found the funny on eVGA's forums...

There is a thread there advising people not to run 3DMark 06 and advising that it could be actually damaging these cards. :D (NOTE: This thread wasn't started by official eVGA personel or representitives)

I needed a good laugh right about now. :)
 
OK, wrong data sheet, here's the one for the 6549 series: http://www.datasheetcatalog.com/datasheets_pdf/M/C/P/6/MCP6549.shtml
Where you can see on page 4 that the operating temperature range goes to 125 deg. C

And it's not a thread about not running 3DM 06, it's a single post in a thread that could have been made by a 10 year old for all we know. There are also replies stating how silly it is to think that 3DM 06 could physically damage the card. I guess if you just thought it's funny, that's fine, but it seems like you're saying that people at evga are saying not to run 3DM 06.

Also, I don't know how you can see that they're trying to do some sort of Jedi mind trick, they freakin' say there's up to a 2% RMA rate for those cards. That's huge. Are you people reading something I'm not?

The rep. from evga also goes on to say:
[font=Verdana, Arial, Helvetica]
I just thought about it after I made the post and some members may think that we are trying to point the finger back at them for over-clocking, this is not the case.

At this moment I just wanted to share some RMA statistics back to the community. We are working on narrowing down the issues that customers are experiencing. Like I stated, we definitely need to improve our quality control and offer users reliability along with stability to prevent even the first RMA.
Click to expand...
He's saying there's a quality control problem (i.e. bad chips or something, for those of you that can't seem to read between the lines) as they need to improve their quality control.

What more do you want?!?
[/font]
 
Just to chime in...

Ive been posting and following the posts on the eVGA forums. Ive had 4 bad eVGA 7900GTX EGS cards (SLI setups). First set worked for about a day. Second set worked for 4 days. No overclocking, in a brand new system.

Same problems as everyone else, artifacting, stuttering, lockups, crashes in 3dmark05, 06, Oblivion, Farcry, BF2, etc.

For me to get 4 bad cards in a row, their return rate is a hella lot higher than 2% or the bad cards were nicely grouped in a batch Monarch computer received...lol
 
knox1711 said:
Just to chime in...

Ive been posting and following the posts on the eVGA forums. Ive had 4 bad eVGA 7900GTX EGS cards (SLI setups). First set worked for about a day. Second set worked for 4 days. No overclocking, in a brand new system.

Same problems as everyone else, artifacting, stuttering, lockups, crashes in 3dmark05, 06, Oblivion, Farcry, BF2, etc.

For me to get 4 bad cards in a row, their return rate is a hella lot higher than 2% or the bad cards were nicely grouped in a batch Monarch computer received...lol
Click to expand...


I bought my defective pair of 7900's from Monarch on May 12th so it's not out of the question ;)
 
legrand said:
OK, wrong data sheet, here's the one for the 6549 series: http://www.datasheetcatalog.com/datasheets_pdf/M/C/P/6/MCP6549.shtml
Click to expand...

The operation temperature range (datasheet shows Ta) is the ambient temperature. This doesn't tell you anything. The junction temperature (die temperature) of the device is 150C (Tj).

However, the thermal photos in the previous post show only the case temperature (Tc).

There is not enough information (in this thread) to determine whether or not those chips are functioning within or outside of design limit.

To find out the actual junction temperature:

Tj = Thetaja*Power + Ta

or

Tj = Thetajc*Power + Tc

The datasheet only gave Thetaja (per package config) but we don't know the power or the ambient temperature.

The thermal photos only gave Tc but we don't know the Thetajc and power....
 
legrand said:
And it's not a thread about not running 3DM 06, it's a single post in a thread that could have been made by a 10 year old for all we know. There are also replies stating how silly it is to think that 3DM 06 could physically damage the card. I guess if you just thought it's funny, that's fine, but it seems like you're saying that people at evga are saying not to run 3DM 06.
Click to expand...

http://www.evga.com/community/messageboard/topic.asp?TOPIC_ID=15435

Let's see, the first post in the thread advises people that the Deep Freeze test in 3DMark 06 could be damaging the cards, followed by a few people agreeing with him, and finally the last couple of posts some people with half a brain explain how rediculous the idea is.

Oh yeah, you're right... it's not a thread about running 3DMark 06 at all.... :rolleyes:

And watch your fucking accusations there, kiddo. I even went out of my way to explain that I found it funny and that it was NOT a thread started by eVGA personell or support staff. Jesus.

legrand said:
Also, I don't know how you can see that they're trying to do some sort of Jedi mind trick, they freakin' say there's up to a 2% RMA rate for those cards. That's huge. Are you people reading something I'm not?
Click to expand...

Dude, seriously... you might want to pull that stick outta your ass. It really could do some long term damage if you're not careful.

legrand said:
The rep. from evga also goes on to say:
[font=Verdana, Arial, Helvetica] He's saying there's a quality control problem (i.e. bad chips or something, for those of you that can't seem to read between the lines) as they need to improve their quality control.

What more do you want?!?
[/font]
Click to expand...

ROFL..

Wait wait wait... no one told me we have to read between the lines. Oh my GOD now I can see everything so clearly! :D I better get over there and re-read all the support staff posts to see if I've missed anything 'between the lines'. Jedi mind tricks indeed....

Thanks for the laughs. :)


Anyway I want everything to turn out for the best believe me. The RMA rate of 'a certain cardbase' seems pretty alarming and anyone want to guess which cards those are?

I'm hoping that I'll eventually get a good 7900GT CO SC out of this whole mess.
 
And watch your fucking accusations there, kiddo
Click to expand...
Dude, seriously... you might want to pull that stick outta your ass.
Click to expand...
Woah, look out, we've pulled out the obscenities to get our point across... ooohhh, I'm so impressed.

Grow up already.
 
My 7900GT Superclocked seems to be running fine...even ocing a little bit more to 560 1600.
 
revenant said:
whoa.. check this

http://www.xtremesystems.org/forums/showpost.php?p=1468059&postcount=439

that could account for all of this.. or most of it.
Click to expand...

Wow, that's pretty interesting indeed. Although he does say that they can be damaged at or above 2.5v, Samsung just doesn't guarantee them above 2.0v. Some very interesting stuff popping up the past few days. Granted we should all take it with a good sized grain of salt, but still...

Thanks for that post, Revenant. :)
 
EVGA hears our customer’s concerns loud and clear about the issues on EVGA 7900 series products. We definitely need to improve our quality control and tighten the testing procedure on the EVGA CO, KO and SuperClocked™ product line. There is no excuse for not providing the best quality products to gamers, that’s EVGA’s spirit. We don’t fear the problems or escape from the problems, we are confident the problem will be resolved shortly. Thank you for everyone’s patients and encouragement.



EVGA’s senior tech staff has already started calling individual customers who have stated and left their concerns. Please follow the steps below if you think you are experiencing some issues. Our staff will contact you promptly during business hours. Please remember this is ONLY for EVGA’s customers.



1. If you have an EVGA 7900GT/GTX series with problems (artifacting, freezing, etc.), Please visit the following page and fill out the quick form. http://www.evga.com/7900.asp

2. We will contact you directly and walk thru the issues one by one.

3. We’ll test individual card per customer using games & 3DMark06 tests.

4. We’ll cross-ship you a new card with a prepaid (ARS) shipping label to return the board back to EVGA freight paid.

EVGA will make every effort to get in contact with you in the next few days, if you require a special time or phone number other than the one registered with EVGA product please note that in form you fill out.

Thanks,
Joe Darwin
[email protected]
 
revenant said:
whoa.. check this

http://www.xtremesystems.org/forums/showpost.php?p=1468059&postcount=439

that could account for all of this.. or most of it.
Click to expand...

All -BJ chips are rated for 2.0 V +/- .1 V (1.1 and 1.2 ns)

All -BC chips are rated for 1.8 V +/- .1V. (1.4, 1.6, 2.0 ns)

So, if GT cards are using 2.0 V, then they would be out of spec. The GTX cards do need 2.0V.

And, looks like EVGA has stepped up big time on the 7900 series issues. Have to recognize that effort.
 
You must log in or register to reply here.
Back
Top