Wanted advice before submitting RMA: HD 7970

T

tonspike

Limp Gawd
Joined
May 24, 2008
Messages
255
I recently got a Diamond HD 7970, and it has been a wonderful card so far. I have not tweaked any voltages, but have come to a happy place between noise, temperature, and OC.

For the past 20 days of ownership, it has idled at a consistent 34-35C. Load temperatures vary, but it never reached 80C.

Today, however, I noticed that the card was idling at 55-60C, and was steadily climbing. As of typing this, it is resting at 61C. I have not changed a single thing, and all settings are untouched. Thinking it was possibly a fan control issue, I reinstalled all drivers to no avail.

Before I send this back to Diamond for an RMA, what would you recommend I do/look at? I've been physically examining the card inside my case and notice a faint scraping noise coming from the fan, which I am unsure whether or not was there before. It is devoid of any dust or debris, I like to keep my case impeccable.

I'm bummed, because I feel like Diamond will take their time to get another card out to me if it comes to it :(
 
Did you buy it within the past 30 days? If so, where?

Usually, computer stores will allow returns/exchanges within 30 days.
 
Tsumi said:
Did you buy it within the past 30 days? If so, where?

Usually, computer stores will allow returns/exchanges within 30 days.
Click to expand...

I forgot to mention this. I purchased it through Amazon, and am well within their return window. Of course all of their 7970s are out of stock, though.

Would you consider an Amazon return the more practical solution? Also, what would happen to the registration--would I simply register the new card under the same name the returned card was under? Would this conflict somehow?
 
Do you have a second monitor connected?

Idle temps of 60C or so are normal on auto fan if you have a second display, as it raises the clock speeds.
 
tonspike said:
I forgot to mention this. I purchased it through Amazon, and am well within their return window. Of course all of their 7970s are out of stock, though.

Would you consider an Amazon return the more practical solution? Also, what would happen to the registration--would I simply register the new card under the same name the returned card was under? Would this conflict somehow?
Click to expand...

Do not under any circumstance RMA it to diamond. Return it to amazon for a refund, the turnaround time is about a week - get the refund instead of an exchange if you have to...amazon takes a week and you can repurchase, while diamond will likely take a lot longer. Not a knock against diamond, but i've never done an RMA to any company that hasn't taken 2-4 weeks. In the meantime, you can order it from another vendor.
 
I'd say pull the card out and make sure theres nothing touching the fan blades like the fan cable.
 
tonspike said:
I recently got a Diamond HD 7970, and it has been a wonderful card so far. I have not tweaked any voltages, but have come to a happy place between noise, temperature, and OC.

For the past 20 days of ownership, it has idled at a consistent 34-35C. Load temperatures vary, but it never reached 80C.

Today, however, I noticed that the card was idling at 55-60C, and was steadily climbing. As of typing this, it is resting at 61C. I have not changed a single thing, and all settings are untouched. Thinking it was possibly a fan control issue, I reinstalled all drivers to no avail.

Before I send this back to Diamond for an RMA, what would you recommend I do/look at? I've been physically examining the card inside my case and notice a faint scraping noise coming from the fan, which I am unsure whether or not was there before. It is devoid of any dust or debris, I like to keep my case impeccable.

I'm bummed, because I feel like Diamond will take their time to get another card out to me if it comes to it :(
Click to expand...

There are two options. You could A return it, or B you can open the card up and see if the fan wire isn't rubbing against the GPU fan. You can open the shell without exposing the GPU heatsink. Don't forget to unscrew the two scews in the back of the videocard before you unscrew the shell. Once you get the shell off see if the fan power wire is rubbing against it. I already opened my 7970s and replaced the TIM with artic silver 5. Granted I got uncomfortable at first when the GPU temps were alittle higher than I originall had it. But after a few weeks the temps are going down and I am reaching the peak curing time for AS5.
 
Turn off hardware acceleration of youtube (flash) and your browser. And do not follow the above disassembling instructions, all you need to do to inspect the fan is remove the plastic shroud which is held by 6 screws, 3 on either side of the card.

Added pic:
sy4q1.jpg
 
Last edited:
Lorien said:
Turn off hardware acceleration of youtube (flash) and your browser. And do not follow the above disassembling instructions, all you need to do to inspect the fan is remove the plastic shroud which is held by 6 screws, 3 on either side of the card.

Added pic:
sy4q1.jpg
Click to expand...

Thank you for this. Although I can't imagine why a wire would suddenly impede the fan from working without some form of movement.
 
Vibration from the fan most likely.. plus heat would soften up the cables so it would most easier from the vibration then.
 
Diamond is below average with their RMA. You should have read a little bit more about this company here and overclock.net. I am sorry to hear about your problems.
 
Nice i have 2 but still within 30 days to return them locally at Microcenter. Prob is that they dont have anything else in stock to exchange them with.. nor is there really anything else in stock out there.
 
Sadly, it looks like this card is going back. I removed the plastic shroud and examined the card, but found nothing out of the ordinary. There is a single wire connected to the "blower" type fan, and it did not impede it in any way. Also, the heatsink itself was perfectly normal, but I guess you wouldn't really be able to tell from a visual standpoint if there was something wrong with it.

It's not even that I think these temperatures are super high--my old 4870 idled at 80C. But a 30C jump within a day or two (35C-65C) rings some alarms.

EDIT:

The time has come. Save myself $50 and get another Diamond? Or spend $20 more and get the Asus DirectCU II?
 
Last edited:
I just wanted to update everyone.

I'm completely baffled as everything seems to have gone back to normal. I haven't changed a single variable in my setup but after 48 hours of erratic temperatures (idle ranging from 55-65C) the problem has seemingly corrected itself as I've now been idling at the correct ~35C for a few hours.

What could be the cause of this? The only thing I can think of is possibly the thermal paste setting in?
 
tonspike said:
I just wanted to update everyone.

I'm completely baffled as everything seems to have gone back to normal. I haven't changed a single variable in my setup but after 48 hours of erratic temperatures (idle ranging from 55-65C) the problem has seemingly corrected itself as I've now been idling at the correct ~35C for a few hours.

What could be the cause of this? The only thing I can think of is possibly the thermal paste setting in?
Click to expand...

For my setup, I sometimes notice my temperatures increase when it's idling when it really isn't.

There are 3 clocks (each have different voltages):

3D load (gaming)
2D load (desktop apps)
Idle

For my 6990 (rough estimate, I forgot the clocks since I haven't looked in a while):
830Mhz while gaming
~500Mhz when using windows media player or 2D apps that require GPU increase in clock speeds.
~150-200Mhz Idle when just websurfing and such.

In cases where the temperatures are not at 3D load and not at Idle means that there's a program or something that's stressing the video card causing 2D clock speeds.

I've had windows media player do this to me. After I close it, it remains open still in my processes menu. After I actually end the process, it drops to Idle clocks and my temps go back down. I would suggest running GPUz and keep an eye on memory and gpu clocks when you're doing basic tasks. Note the idle clocks. When you see your temps idling high again, check to see if the clocks are indeed still idle or in 2D mode.

This is strictly from my experiences only with 6990, 6970, and 5870.
 
Last edited:
thats strange. my MSI 7970 runs on the hot side too at stock voltage. Rather then deal with down time of sending it back ,exchanging it and so on. i just undervolted mine from 1.175 to 1.043 temps went from 90c+ (auto fan speed) under full load to 71C with only 38% fan speed manual
 
Blow out your card with some canned air. Dust, especially in a dusty place like OK, holds heat. Couldnt hurt.
 
You guys need to remove the stock heatsink and check the goop on that thing!

Out of 4 cards i removed the heatsinks on for the water blocks i noticed 1 that the goop was cooked/baked on the core! It was not gooy like the other 3 cards. I literally had to scrap the stuff off (chip at the dried up goop with my finger nail and alot of rubbing alcohol to get it off! So thats why it was getting much hotter then the other 3 cards AND it was lower voltage then the other 3 at 1.049v while the other cards where 1.112v and 1.174v. That card was the 2nd in the quad setup runing hotter then the first card!

God only knows how much longer that card would of lasted like that if i didnt notice.
 
The 7970 lowers voltage and fan speed on idle, before you go taking things apart and voiding your warranty I'd check the voltage level along side your temperatures. More voltage = more heat
 
Reinstall the drivers and see if the issue still persists. I'd say try booting a live cd and check to see if the idle temps are the same on different os'es which should help you narrow it down to hardware or software issue, but i'm not sure if it'll work since ati support for linux is abysmal.
 
You must log in or register to reply here.
Back
Top