I'm done with my GPU, Wondering if I can get RMA?

[Akhilcool]

Hey guys so I am quite pissed off at this point. My Sapphire 6870 w/ Dual fans (DiRT 3 edition), might I stress again that it is has dual fans, is overheating. Now I've already posted many threads on this forum asking of how to troubleshoot this problem, yet I had no luck with any of the suggestions I was given. I pushed my OC back to stock (I was barely able to get any higher of a stable OC anyways), I even replaced the thermal paste with Arctic Silver 5, two times because I thought I might have added too much the first time. My temps when running Furmark or OCCT are easily 82*C + and this is at a stock overclock of 900mhz core and 1050mhz memory clock. When I play Just Cause 2, my temps are around 65*C - 70*C which are generally low compared to my stability testing temps yet they are still pretty high considering that the GPU is dual fan and the game isn't even that graphic intensive as opposed to BF3 or any of the latest games. When I play Witcher 2 however, my temps are between 70*C and 75*C. Now obviously this is not even close to normal as other people are getting between 50*C -70*C max even when stress testing. When my card idles it is at 33*C as soon as the system boots up. After about 10-20 min of idling it is at 37*C-45*C. Now I pretty much gave up on the card and I'd like to RMA it A.S.A.P.

My system is well cooled I'd say. I have a HAF 912 case that has two CM sickleflow 120mm fans as front intake. One stock 120mm side case fan that blows directly at my GPU, a 200mm megaflow as exhaust on the top and a stock 120mm rear case fan as exhaust. I even thought that maybe my airflow pressures were too neutral, or maybe there might have been a vortex that was disturbing the airflow so I turned off all my case fans except for my rear 120mm fan. This didn't change anything much so I turned on my side case fan along with my rear fan which also didn't help. After playing around with which fans were on or off I realized that the fans had nothing to do with the problem and that it was only the GPU that was full of crap. I'm getting sufficient power since I have an iCAN 600W PSU that should be more than enough even for two GPUs. So in conclusion, I am wondering if I can still get a RMA done even though I changed the thermal paste. The thing is that the way it was applied by Sapphire was using a perfectly rectangular applicator that had a shape a lot bigger than the actual TIM. So when I applied it I only obviously put the thermal paste on the TIM itself. If they open the GPU it's obviously going to be noticeable. I had previously already filled out the RMA form regarding a different problem. What I thought was that my GPU was rattling when at a fan speed of around 65%-80% but I think the problem was actually that one of those PCI-e metal slot cap things that block empty PCI-e slots from being open from the back of the PC case ( lol sorry I forgot the actual term,) was actually touching the metal around the cap that was on the input interface metal cap thingy of the GPU itself. This caused some sort of resonance that created the rattling noise at certain fan speeds. I think I fixed this problem myself but the temperatures are still killing me.

Sorry it is a long post but I gave up and really need to know if I can replace this faulty card. Thank you!

 

[alvine]

mmmm well you opened the card up so that has voided your warranty...

 

[Akhilcool]

mmmm well you opened the card up so that has voided your warranty...

But some other people have said they've been able to get away with it regardless.

 

[Akhilcool]

I also made sure there wasn't any stickers hidden in the card.

 

[Kari]

well you could buy aftermarket cooler for it...

 

[Akhilcool]

well you could buy aftermarket cooler for it...

But if the GPU gets to those temps even with a highly rated dual fan cooling system...doesn't it mean that there is a problem with the GPU itself and that I should RMA it?

 

[Kari]

is this your card?
http://www.sapphiretech.com/presentation/product/?cid=1&gid=3&sgid=1037&pid=1312&psn=&lid=1&leg=0

well it looks nice... Only thing I could think of is that the heatpipes on the cooler have released their contents and are now dry, and not conducting heat properly...

anyhoo i think you could just rma the card, if they give you trouble over the fact that you've reapplied the Thermal Interface Material, telll them you knew what you were doing and wore a antistatic wrist straps and stuff...

 

[RetiredChief]

Myself, I'd try the RMA route, with the statement that you were tring a realitively simple fix. And yes you could indicate that you wore an ESD stap, which you should have and being a good tech did.

Some additional comments.
I'm a little confused. You said "So when I applied it I only obviously put the thermal paste on the TIM itself"
Tim is the thermal paste, Hopefully you did not apply new theram paste OVER the top of the Old thermal paste (TIM) and that was just averbage error.

Overheating of an IC is caused by one or more of the following factors.
1) Incorrect application of the TIM causing poor thermal conductivity between the IC and the Cooler base.
2) Ineffecient HS and fan
3) Increased current thru the IC. Increased current is caused by (A) Incorrect voltage or (B) the Impedeadance (Resistance) has decreased.
.. (A) could be caused by PSU, or voltage regulator on the GPU (ie onboard regulator that converts the +12V to a lower value (simular to using the +12 V and regulating it down to the 1 Volt used by the CPU, or the 1.5 V used for your Ram.
.... (B) is the chip itself going Bad.

While I dougt it is your PSU as I think you would have other indication Your statement " iCAN 600W PSU that should be more than enough even for two GPUs" is not really accurate.
Yes 600 Watt is more than enough for your system as you are probably maxing out @ 350 Watts while running furmark. I have simular system - I5-2500k, w/16 gigs ram, 2 SSDs, 6870 GPU (and may be a Sapphire model, at work so can not verify), 1 HDD and one Blu-ray writer. That does NOT mean ALL 600 Watt PSUs are created the same, Some are high quality with great regulation while at the other end _______, well you can fill in the Blank
For most, the only way to verify the output of a PSU is to swap with a known good unit.
Or you caoud invest in a digit O'scope (Digital Votl meter reall not good enough) at about a grand. PS I do have both.
The bottom line is just a question - How do you KNOW your PSU is good.

 

[vrumor]

My card idles at 42c and when im under load it routinely gets above 65c and 70c-80c during benchmarking. Not sure why you are complaining about those temps, they seem pretty much inline for me. RMAing wont change a thing with your temps.

 

[Akhilcool]

Myself, I'd try the RMA route, with the statement that you were tring a realitively simple fix. And yes you could indicate that you wore an ESD stap, which you should have and being a good tech did.

Some additional comments.
I'm a little confused. You said "So when I applied it I only obviously put the thermal paste on the TIM itself"
Tim is the thermal paste, Hopefully you did not apply new theram paste OVER the top of the Old thermal paste (TIM) and that was just averbage error.

Overheating of an IC is caused by one or more of the following factors.
1) Incorrect application of the TIM causing poor thermal conductivity between the IC and the Cooler base.
2) Ineffecient HS and fan
3) Increased current thru the IC. Increased current is caused by (A) Incorrect voltage or (B) the Impedeadance (Resistance) has decreased.
.. (A) could be caused by PSU, or voltage regulator on the GPU (ie onboard regulator that converts the +12V to a lower value (simular to using the +12 V and regulating it down to the 1 Volt used by the CPU, or the 1.5 V used for your Ram.
.... (B) is the chip itself going Bad.

While I dougt it is your PSU as I think you would have other indication Your statement " iCAN 600W PSU that should be more than enough even for two GPUs" is not really accurate.
Yes 600 Watt is more than enough for your system as you are probably maxing out @ 350 Watts while running furmark. I have simular system - I5-2500k, w/16 gigs ram, 2 SSDs, 6870 GPU (and may be a Sapphire model, at work so can not verify), 1 HDD and one Blu-ray writer. That does NOT mean ALL 600 Watt PSUs are created the same, Some are high quality with great regulation while at the other end _______, well you can fill in the Blank
For most, the only way to verify the output of a PSU is to swap with a known good unit.
Or you caoud invest in a digit O'scope (Digital Votl meter reall not good enough) at about a grand. PS I do have both.
The bottom line is just a question - How do you KNOW your PSU is good.

Oh lol I'm sorry yeah I got my vocabulary mixed. What I mean is that I only applied thermal compound over the heat spreader on top of the core chip. Not anymore or less. The stock thermal paste that was previously used was applied to the copper of the heatsink instead and then put together. This meant that the heatsink had thermal paste on it with a surface area greater than that of the area of the heat spreader so in order for me to replicate that look would mean I would have to waste some thermal paste and use a rectangular tracer in order to make a perfect enlarged rectangle shape application of thermal paste. They'd be able to tell by the better quality of the thermal paste that it isn't their own anyhow. Oh and you are right that my PSU can't be good enough just based on the wattage. Unfortunately I don't have another PSU that can be used as my other ones are less than 500W

It runs fine in games I guess although I'm having a bit of a micro-stuttering issue while playing The Witcher 2 and Assassin's Creed Brotherhood. There is absolutely no issue in Just Cause 2 however. I would have liked to overclock the card a bit more but I guess it's fine as it is.

The last comment by 'vrumor' may be right that RMAing the card won't be able to fix the heating issue but I have realized that the buzzing noise I mentioned earlier is still ongoing. I realized that it isn't my case but the card itself. When I boost the fan from 30% straight to 100% in one step, for a brief 2 seconds my GPU makes a rattling noise near the fan area close to the input interface. The reason why I thought the sound went away from what I did to the case before was because when the metal from the input interface of the GPU was touching any of the other metal caps on top or below it, it would increase the vibrations and make more of a noise. Good thing I still caught this sound. What should I do about it? Should I try fixing this somehow or RMA the card and hope that the problem along with the many other ones may go away?

 

[Akhilcool]

lol and I was just playing around with my Sapphire 5670 with Arctic Cooling. Man that card was a bad boy even though it didn't have as high of a performance as my 6870. It ran at 28*C on idle and 58*C when stress testing on full load. Not only that but it had a great overclocking ability. I was able to max out the cards clock speeds beyond its maximum without even touching the core voltage ( which I couldn't change unless I flashed the BIOS.) I went from 775mhz core clock to 900mhz. That is 125mhz increase on stock voltage. T_T I know that not all cards are the same and I'm comparing two completely different cards but I was just wondering why I could only go up 20mhz of core clock on stock voltage on my 6870.

 

[cmcghee358]

My card idles at 42c and when im under load it routinely gets above 65c and 70c-80c during benchmarking. Not sure why you are complaining about those temps, they seem pretty much inline for me. RMAing wont change a thing with your temps.

^ This

GPUs run hotter than CPUs. Your temps are actually quite normal.

 

[Kari]

5670 is a low end card, so the chip is small and it doesn't put out much heat.... 6870 is two 'steps' higher

 

[Akhilcool]

is this your card?
http://www.sapphiretech.com/presentation/product/?cid=1&gid=3&sgid=1037&pid=1312&psn=&lid=1&leg=0

well it looks nice... Only thing I could think of is that the heatpipes on the cooler have released their contents and are now dry, and not conducting heat properly...

anyhoo i think you could just rma the card, if they give you trouble over the fact that you've reapplied the Thermal Interface Material, telll them you knew what you were doing and wore a antistatic wrist straps and stuff...

Yeap that's the one. It is supposed to have really good cooling. It's not the case with my specific card.

 

[Kari]

^ This

GPUs run hotter than CPUs. Your temps are actually quite normal.

normal for reference desings, that card has custom cooler with 2 fans on it so the temps seem a bit high... or the cooler isn't as effective as it seems...

 

[Akhilcool]

^ This

GPUs run hotter than CPUs. Your temps are actually quite normal.

Well than I only have the problem with the buzzing then "The last comment by 'vrumor' may be right that RMAing the card won't be able to fix the heating issue but I have realized that the buzzing noise I mentioned earlier is still ongoing. I realized that it isn't my case but the card itself. When I boost the fan from 30% straight to 100% in one step, for a brief 2 seconds my GPU makes a rattling noise near the fan area close to the input interface. The reason why I thought the sound went away from what I did to the case before was because when the metal from the input interface of the GPU was touching any of the other metal caps on top or below it, it would increase the vibrations and make more of a noise. Good thing I still caught this sound. What should I do about it? Should I try fixing this somehow or RMA the card and hope that the problem along with the many other ones may go away?" - quoted from me. I guess the heat is normal. But it does limit me almost completely from overclocking as if I go any higher my temps will shoot above 80*C while gaming which is terrible.

 

[Akhilcool]

normal for reference desings, that card has custom cooler with 2 fans on it so the temps seem a bit high... or the cooler isn't as effective as it seems...

That's what I'm thinking but from other reviews of this card people are getting much better temps. So it might just be my card.

 

[cmcghee358]

Why do you assume the inability to overclock is somehow the manufacturer's problem?

Does your card function at the rated speed?

Does your card stay below the maximum recommend temperature at the rated speed?

If you answered yes to both, I see no RMA based on temperature.

As for the buzzing sound, that could possibly be an RMA. But if I was the manufacturer, and I found aftermarket TIM on the GPU, I'd assume the customer was at fault for the buzzing/clicking/vibe. Sorry to be such a downer

 

[Dogsnake]

Looking this entire post over, the temps you report may not be off at all. You can not say if they are high. What is the room ambient temperature? This makes a big difference. Your temps seem 5-10 degree C on the high side maybe. They seem inside the cards operating envelope. Your case and air flow are good. Do you get artifacts or crashes during game play? Did it crash during stress testing? From what you have said I am guessing the answer to the crashes is no. The stuttering is often due to the game calling new data from the HD to the memory Cache and is very common in games when moving in and out of areas. If the thing is making noise send it back. Most likely they will send a replacement with no problem. Otherwise keep it and enjoy your games. Please correct me but the problem besides the noise, is that you think the temps. should be lower. Not that there is an on going operational issue. Right now my 6950 is idling at 45C and that is normal to me.

 

[Akhilcool]

Why do you assume the inability to overclock is somehow the manufacturer's problem?

Does your card function at the rated speed?

Does your card stay below the maximum recommend temperature at the rated speed?

If you answered yes to both, I see no RMA based on temperature.

As for the buzzing sound, that could possibly be an RMA. But if I was the manufacturer, and I found aftermarket TIM on the GPU, I'd assume the customer was at fault for the buzzing/clicking/vibe. Sorry to be such a downer

lol the buzzing noise has been going on even before I changed the TIM so it can't be my fault.

 

[cmcghee358]

normal for reference desings, that card has custom cooler with 2 fans on it so the temps seem a bit high... or the cooler isn't as effective as it seems...

The card is advertised as having a more advanced cooling system, but it never states the anticipated temp, or any guarantee as to a maximum temperature.

You can't submit an RMA because something doesn't perform well in your opinion. It has to fall short of specifications.

 

[cmcghee358]

lol the buzzing noise has been going on even before I changed the TIM so it can't be my fault.

Re-read what I posted please.

I didn't say you made the buzzing sound.

 

[Akhilcool]

Looking this entire post over, the temps you report may not be off at all. You can not say if they are high. What is the room ambient temperature? This makes a big difference. Your temps seem 5-10 degree C on the high side maybe. They seem inside the cards operating envelope. Your case and air flow are good. Do you get artifacts or crashes during game play? Did it crash during stress testing? From what you have said I am guessing the answer to the crashes is no. The stuttering is often due to the game calling new data from the HD to the memory Cache and is very common in games when moving in and out of areas. If the thing is making noise send it back. Most likely they will send a replacement with no problem. Otherwise keep it and enjoy your games. Please correct me but the problem besides the noise, is that you think the temps. should be lower. Not that there is an on going operational issue. Right now my 6950 is idling at 45C and that is normal to me.

Very true, my ambient air temp however is cool because my PC is in my basement and the temp down here is currently 20*C (I live in Canada). And yes the temps could be lower but from what I've just read on this thread, I think it should be fine. My only real issue is the noise which definitely needs to be looked at by Sapphire.

 

[Akhilcool]

Re-read what I posted please.

I didn't say you made the buzzing sound.

Yes I know what you said, that regardless of the fact that if the TIM was the reason for the noise or not, even if I still put it on, Sapphire wouldn't want to lose money by replacing my card when they could just blame me for the noise because I changed the TIM. Yet I don't think they would do that as they may be more than happy to just replace the card if I explain to them that this problem has been ongoing since I bought the card.

 

[cmcghee358]

No you still missed it.

No one would think TIM causes a buzz/click/sound, that's obsurd.

But to apply the TIM requires some amount of disassmbly of the cooling mechanism, which is making an unusual sound.

If I was Sapphire, I'd assume you damaged the product when you voided the warranty doing maintenance to the card.