Have you swapped your MSI 7950?

[IC3]

That's definitely too hot, especially at stock. RMA that bad boy.

It's a shame, because people that have 100% good versions of this card are having great temps etc, and it's a sexy looking card too.

Mort, my card runs at 78C at about 67% fans with 1100/1350 @ 1.112V. That's after about 2 loops of Heaven 3.0, although it usually gets there playing other games too. Furmark etc push it a lot higher, so I don't use that to test.

Devrij, is right I had similiar temps to him, but got my cash back under 14 day gurrante and thats what. I would do on your place.
Nice to know that my thread is helping people

 

[rolypolyman]

Because they don't have proper VRM cooling. For example, the MSIs and HIS cards have a heatsink over the VRMs, the Gigabytes (and Asus Direct CU I think) don't have them, so the VRMs run hotter, and in theory could inhibit over-volting related over-clocking.

So, my card running at 1250 core OC returning vrm temps of 65\66c, is that too hot because I've read of vrms getting to 125c, now that is what I would call hot. I would like to know what the cards you mentioned, was it Msi and His, vrm temps are when running OC at 1250 core.

 

[spoffle]

Just because your case in contrary to that, it doesn't mean it can't be the case. As I said "in theory", a card with VRM cooling is more desirable, not that Gigabyte Windforce cards can't and don't overclock.

66c on the VRMs sounds cool, is that definitely the VRMs though? That sounds very very cool for a card with no VRM heatsink.

 

[Geo-Gamer]

So, my card running at 1250 core OC returning vrm temps of 65\66c, is that too hot because I've read of vrms getting to 125c, now that is what I would call hot. I would like to know what the cards you mentioned, was it Msi and His, vrm temps are when running OC at 1250 core.

what have you overclocked on the memory?

 

[IC3]

So, my card running at 1250 core OC returning vrm temps of 65\66c, is that too hot because I've read of vrms getting to 125c, now that is what I would call hot. I would like to know what the cards you mentioned, was it Msi and His, vrm temps are when running OC at 1250 core.

 

[rolypolyman]

what have you overclocked on the memory?

Nothing, dont touch the memory, little point.

 

[rolypolyman]

Just because your case in contrary to that, it doesn't mean it can't be the case. As I said "in theory", a card with VRM cooling is more desirable, not that Gigabyte Windforce cards can't and don't overclock.

66c on the VRMs sounds cool, is that definitely the VRMs though? That sounds very very cool for a card with no VRM heatsink.

Yeh, definately vrms, logged from the Gpu-z log file, usefull that if a little untidy.

Spoffle, I owe you a bit of an apology, I went over those gpu-z logs again and found the vrm temps actually peaked at 73c, still not too bad at 1250 core.

 

[Gandalf185]

Devrij, is right I had similiar temps to him, but got my cash back under 14 day gurrante and thats what. I would do on your place.
Nice to know that my thread is helping people

This thread is indeed very helpful! I found a lot of people with similar issues on various other forums, and most people advise to check the case airflow. I could have spent days trying to sort this out when it's just a bad card.

This morning I received an answer from MSI techsupport: The temperature of 80°C is normal when the GPU reach to 100% loading, at the same time that the loud fans it's inevitable.

So there you go, MSI consider this to be normal. I'm returning it to the shop and will swap for a different brand, no more MSI products for me ever again!

 

[Devrij]

Yeh, definately vrms, logged from the Gpu-z log file, usefull that if a little untidy.

Spoffle, I owe you a bit of an apology, I went over those gpu-z logs again and found the vrm temps actually peaked at 73c, still not too bad at 1250 core.

Not bad at all. I ran my card at 1200/1250 and my VRMs got to 105C within a few seconds of Heaven 3.0 so I stopped the benchmark before it shut down.

This thread is indeed very helpful! I found a lot of people with similar issues on various other forums, and most people advise to check the case airflow. I could have spent days trying to sort this out when it's just a bad card.

This morning I received an answer from MSI techsupport: The temperature of 80°C is normal when the GPU reach to 100% loading, at the same time that the loud fans it's inevitable.

So there you go, MSI consider this to be normal. I'm returning it to the shop and will swap for a different brand, no more MSI products for me ever again!

Yeah, I had a similar response on sending in a webnote to OcUK, saying that up to 90C was fine. It most certainly is not! Maybe if I'm overclocking the nuts off it, but not with a small bump to clock speeds. Was playing FC3 last night and was running 79C at 75% fans. That's pretty toasty. Just waiting for my RMA email so I can ship it off and get an IceQ to replace it. Hoping I get it soon because I'm giving my 5850 to my bro for xmas (and other stuff! I'm not that cheap ) and it's my only backup card.

 

[Rusty0611]

Give OcUK credit here as well - they're replacing them no questions asked.

Good service.

I think you can thank 5UB for that - I messaged him asking him for clarification and he said they were discussing it. He didn't get back to me on what the outcome was but we can safely assume this is it.

 

[Devrij]

Give OcUK credit here as well - they're replacing them no questions asked.

Good service.

I think you can thank 5UB for that - I messaged him asking him for clarification and he said they were discussing it. He didn't get back to me on what the outcome was but we can safely assume this is it.

Yeah they've been great. Once you speak to someone who's aware of the issue it's plain sailing.

 

[Gandalf185]

I have the right to return any product within 8 days after purchase anyway, no questions asked. It's the law in my country
Will try to have it replaced with a Gigabyte card. I'd rather have a proper working MSI TF3, but after such a BS response I just can't be bothered anymore.

 

[ScottiB]

Hi guys,

So sorry for the lack of clear response. We've had a hell of a time trying to narrow down the issues involved and identify the cause.

firstly, let me just clarify something. The V2 models are based on 7970 PCB/GPU, hence they will run hoter than the V1 models that have been in the channel and in the reviews. That said, the differences reported in this thread are unacceptable, our testing suggests 5C and 3% fan increase in Furmark when V1 & V2 are tested side by side. I've included our data below.

We are 99% sure that the thermal grease application is to blame and will be happy to provide a aid and warranty cover to anyone who wishes to reapply the grease themselves.

The reason for the delayed response is that found it impossible to accurately identify the affected cards. Many hundreds of cards went to OcUK who booked them in using the same stock code as the V1. We were left with a needle and haystack situation.

We will offer OcUK every possible support with this issue. Please feel free to contact me via trust if you have any questions.

Tahiti Pro Tahiti Pro H
Idle 31℃ 35℃
fan duty 30% 30%
FurMark 65℃ 70℃
fan duty 49% 52%
Total board Voltage 225W 260W
Fan Duty: 40℃--30% ; 70℃--50% ; 90℃--100%

 

[Devrij]

Hi guys,

So sorry for the lack of clear response. We've had a hell of a time trying to narrow down the issues involved and identify the cause.

firstly, let me just clarify something. The V2 models are based on 7970 PCB/GPU, hence they will run hoter than the V1 models that have been in the channel and in the reviews. That said, the differences reported in this thread are unacceptable, our testing suggests 5C and 3% fan increase in Furmark when V1 & V2 are tested side by side. I've included our data below.

We are 99% sure that the thermal grease application is to blame and will be happy to provide a aid and warranty cover to anyone who wishes to reapply the grease themselves.

The reason for the delayed response is that found it impossible to accurately identify the affected cards. Many hundreds of cards went to OcUK who booked them in using the same stock code as the V1. We were left with a needle and haystack situation.

We will offer OcUK every possible support with this issue. Please feel free to contact me via trust if you have any questions.

Tahiti Pro Tahiti Pro H
Idle 31℃ 35℃
fan duty 30% 30%
FurMark 65℃ 70℃
fan duty 49% 52%
Total board Voltage 225W 260W
Fan Duty: 40℃--30% ; 70℃--50% ; 90℃--100%

Thanks Scotti, it's great to get an official response on what is normal and what isn't. I've always been happy with MSI products, so I'm sure this is just a one-off issue with these cards. All the same, I'm glad I'm getting a different brand as a replacement if they haven't been able to identify which cards are affected.

 

[IC3]

Hi guys,

So sorry for the lack of clear response. We've had a hell of a time trying to narrow down the issues involved and identify the cause.

firstly, let me just clarify something. The V2 models are based on 7970 PCB/GPU, hence they will run hoter than the V1 models that have been in the channel and in the reviews. That said, the differences reported in this thread are unacceptable, our testing suggests 5C and 3% fan increase in Furmark when V1 & V2 are tested side by side. I've included our data below.

We are 99% sure that the thermal grease application is to blame and will be happy to provide a aid and warranty cover to anyone who wishes to reapply the grease themselves.


The reason for the delayed response is that found it impossible to accurately identify the affected cards. Many hundreds of cards went to OcUK who booked them in using the same stock code as the V1. We were left with a needle and haystack situation.

We will offer OcUK every possible support with this issue. Please feel free to contact me via trust if you have any questions.

Tahiti Pro Tahiti Pro H
Idle 31℃ 35℃
fan duty 30% 30%
FurMark 65℃ 70℃
fan duty 49% 52%
Total board Voltage 225W 260W
Fan Duty: 40℃--30% ; 70℃--50% ; 90℃--100%

What are the temps with re-applied thermal compound?

 

[Hybridtheory]

Hi guys,

So sorry for the lack of clear response. We've had a hell of a time trying to narrow down the issues involved and identify the cause.

firstly, let me just clarify something. The V2 models are based on 7970 PCB/GPU, hence they will run hoter than the V1 models that have been in the channel and in the reviews. That said, the differences reported in this thread are unacceptable, our testing suggests 5C and 3% fan increase in Furmark when V1 & V2 are tested side by side. I've included our data below.

We are 99% sure that the thermal grease application is to blame and will be happy to provide a aid and warranty cover to anyone who wishes to reapply the grease themselves.

What Thermal Grease do you recommend for reapplying ?

The reason for the delayed response is that found it impossible to accurately identify the affected cards. Many hundreds of cards went to OcUK who booked them in using the same stock code as the V1. We were left with a needle and haystack situation.

We will offer OcUK every possible support with this issue. Please feel free to contact me via trust if you have any questions.

Tahiti Pro Tahiti Pro H
Idle 31℃ 35℃
fan duty 30% 30%
FurMark 65℃ 70℃
fan duty 49% 52%
Total board Voltage 225W 260W
Fan Duty: 40℃--30% ; 70℃--50% ; 90℃--100%

What Thermal Grease do you recommend ?

 

[spoffle]

Hi guys,

So sorry for the lack of clear response. We've had a hell of a time trying to narrow down the issues involved and identify the cause.

firstly, let me just clarify something. The V2 models are based on 7970 PCB/GPU, hence they will run hoter than the V1 models that have been in the channel and in the reviews. That said, the differences reported in this thread are unacceptable, our testing suggests 5C and 3% fan increase in Furmark when V1 & V2 are tested side by side. I've included our data below.

We are 99% sure that the thermal grease application is to blame and will be happy to provide a aid and warranty cover to anyone who wishes to reapply the grease themselves.

The reason for the delayed response is that found it impossible to accurately identify the affected cards. Many hundreds of cards went to OcUK who booked them in using the same stock code as the V1. We were left with a needle and haystack situation.

We will offer OcUK every possible support with this issue. Please feel free to contact me via trust if you have any questions.

Tahiti Pro Tahiti Pro H
Idle 31℃ 35℃
fan duty 30% 30%
FurMark 65℃ 70℃
fan duty 49% 52%
Total board Voltage 225W 260W
Fan Duty: 40℃--30% ; 70℃--50% ; 90℃--100%

I replaced the thermal paste on both of mine, it "helped" to an extent, but they were both still pretty bad.

With a new application of thermal paste the fans still hit 100% with temps over 80C.

Something to note though, one of my cards, the thermal paste application was really bad. There were even grey finger prints around the card that were clearly from the person who assembled the card, in thermal paste used.

 

[IC3]

Did you have this problem only in uk?

 

[ScottiB]

to answer the questions above:

The temps shown in the data are from non-affected cards

Any non-metallic thermal grease - we can provide some if you wish

to my knowledge only one batch, all of which were shipped to OcUK

 

[IC3]

to answer the questions above:

The temps shown in the data are from non-affected cards

Any non-metallic thermal grease - we can provide some if you wish

to my knowledge only one batch, all of which were shipped to OcUK

So good working MSI should have temps of around 28-32 idle and 59-61c load?