[Fan problems] Asus HD7770

[plåtburken]

Heya!

I seen lots of people post here questions and etc about their AMD/ATI gpus, so I am gonna do it too, since I got a problem with my card.
Now the problem with my card is that it has some weird fan profile that I can't change at all, no matter what application I use.
The card auto adjusts itself, whenever it gets to 50 degrees it sets fan speed to 60% and as soon it goes down to it sets fans to 20% and then ramps up again just to cool it down and it does that over and over and over, none-stop!
Once the card reaches above 65 degrees the fan sets itself to almost 100% and it sounds like a hurricane.
I have no warranty left and not much money to spare. So wondering is there anyway to fix this issue? Like a Bios change?

 

[Liviu Cojocaru]

Heya!

I seen lots of people post here questions and etc about their AMD/ATI gpus, so I am gonna do it too, since I got a problem with my card.
Now the problem with my card is that it has some weird fan profile that I can't change at all, no matter what application I use.
The card auto adjusts itself, whenever it gets to 50 degrees it sets fan speed to 60% and as soon it goes down to it sets fans to 20% and then ramps up again just to cool it down and it does that over and over and over, none-stop!
Once the card reaches above 65 degrees the fan sets itself to almost 100% and it sounds like a hurricane.
I have no warranty left and not much money to spare. So wondering is there anyway to fix this issue? Like a Bios change?

What software have you tried to use this far? The BIOS change would be an option if you have the proper knowledge and skills ( this is a risky thing) another thing to do would be to get a better aftermarket cooler (again involves some skills to replace it)

 

[Mussels]

sounds like you need to dust out the heatsink and replace the thermal paste.

It gets hot, so the fan spins up
Overheats despite the fan spinning up, so it throttles the card
card is cooled down so fans slow down
Card then reverts to full speed, card starts to get hot again...

 

[plåtburken]

What software have you tried to use this far? The BIOS change would be an option if you have the proper knowledge and skills ( this is a risky thing) another thing to do would be to get a better aftermarket cooler (again involves some skills to replace it)

I been using MSI-Afterburner and Asus own GPUTweaker tool both work alright until the GPU hits 50 degrees after that I assume the internal fan profile hits in and takes over and overrides the software profiles and starts to make noise.

sounds like you need to dust out the heatsink and replace the thermal paste.

It gets hot, so the fan spins up
Overheats despite the fan spinning up, so it throttles the card
card is cooled down so fans slow down
Card then reverts to full speed, card starts to get hot again...

I have cleaned out the heatsink from dust with a compressed can of air and I have applied some none-conductive thermal paste as well as applied new thermal pads around the VRAM. The idle temps of the GPU is 22-24C, while doing light gaming, such as CSGO or Dota2 the GPU will never reach 50 degrees. However if I play games that are more demanding the GPU will hit a maximum of 66 degrees but the fans will spin almost at 100% which is insane and the GPU will not stop spinning after it cools down. I have to go to MSI afterburner to lower the fan down to 20-40% maxspeed, sometimes it won't work and I need to do a reboot.

 

[Liviu Cojocaru]

I been using MSI-Afterburner and Asus own GPUTweaker tool both work alright until the GPU hits 50 degrees after that I assume the internal fan profile hits in and takes over and overrides the software profiles and starts to make noise.

I have cleaned out the heatsink from dust with a compressed can of air and I have applied some none-conductive thermal paste as well as applied new thermal pads around the VRAM. The idle temps of the GPU is 22-24C, while doing light gaming, such as CSGO or Dota2 the GPU will never reach 50 degrees. However if I play games that are more demanding the GPU will hit a maximum of 66 degrees but the fans will spin almost at 100% which is insane and the GPU will not stop spinning after it cools down. I have to go to MSI afterburner to lower the fan down to 20-40% maxspeed, sometimes it won't work and I need to do a reboot.

Those are respectable temps, it seems that a custom BIOS would do the trick for you but that is a big risk. I would advise to get a better cooler instead.

 

[Laki89]

I saw few times when GPU is dying that it gets higher temps than usual
First sign of that is: Screen sometimes goes blank and following message you get: display driver stopped working and successfully recovered
When things get worse you get randomly blue screens
If that is a problem which i described, only solution is buying new GPU

 

[Mussels]

those temps do seem OK, unless something else on the card is overheating that doesnt have a temp sensor (VRMs?)

 

[plåtburken]

Those are respectable temps, it seems that a custom BIOS would do the trick for you but that is a big risk. I would advise to get a better cooler instead.

I would like to try the BIOS thing since I am not really in the current state to put money into anything outside of my bills.

I saw few times when GPU is dying that it gets higher temps than usual
First sign of that is: Screen sometimes goes blank and following message you get: display driver stopped working and successfully recovered
When things get worse you get randomly blue screens
If that is a problem which i described, only solution is buying new GPU

Nah I haven't suffered from any black or blank screens, nor have I suffered from BSOD or had display driver stopped working and successfully recovered. I am just suspecting that something else is up with the card.

those temps do seem OK, unless something else on the card is overheating that doesnt have a temp sensor (VRMs?)

Could be so, but I would really like to try fix a BIOS change on the internal fan profile setting.


Also this is the card I am having troubles with.
HD7770-DC-1GD5
https://www.asus.com/Graphics-Cards/HD7770DC1GD5/

 

[Mussels]

Have you considered hard wiring the fan to an external power source?

I've rigged GPU fans to 9V/7V in the past and left them at set speeds (if you go too low the temps rise, too high and its always loud - but its an option)

 

[plåtburken]

Have you considered hard wiring the fan to an external power source?

I've rigged GPU fans to 9V/7V in the past and left them at set speeds (if you go too low the temps rise, too high and its always loud - but its an option)

This sounds interesting, could try to wire it out of the case and add a potentiometer for lowering and raising the voltage.
If there is no option for flashing then I might look into this.

 

[jaggerwild]

Just use MSI afterburnner n crank up the fan profile, https://www.msi.com/page/afterburner

Also suggest as others have to replace the factory TIM if the warranty is void, as that means the card is a few years old by now. Always good to replace the TIM after the GPU gets a few years old.

 

[mstenholm]

Just use MSI afterburnner n crank up the fan profile, https://www.msi.com/page/afterburner

Also suggest as others have to replace the factory TIM if the warranty is void, as that means the card is a few years old by now. Always good to replace the TIM after the GPU gets a few years old.

OP did both. Read the whole thread....

 

[Laki89]

Nah I haven't suffered from any black or blank screens, nor have I suffered from BSOD or had display driver stopped working and successfully recovered. I am just suspecting that something else is up with the card.

That's good to hear
I have 2 questions:
you said:

I have applied some none-conductive thermal paste as well as applied new thermal pads around the VRAM

1. Those new thermal pads for VRAM - are they same size like old ones or they are fatter?
2. When you assembled graphics card after applying thermal paste - did you tight all screws around GPU with similar strength?
(also if you have 3 screws for GPU instead of 4 - that is huge problem)

 

[plåtburken]

That's good to hear
I have 2 questions:
you said:

1. Those new thermal pads for VRAM - are they same size like old ones or they are fatter?
2. When you assembled graphics card after applying thermal paste - did you tight all screws around GPU with similar strength?
(also if you have 3 screws for GPU instead of 4 - that is huge problem)

1) The new thermal pads for the VRAM are at the same size as the old ones
2) I assembled it by screwing cross-wise so, start at one end and then to the other side diagonally. First, a bit then fully screwed once every screw was placed. It has 4 screws.

 

[Laki89]

1) The new thermal pads for the VRAM are at the same size as the old ones
2) I assembled it by screwing cross-wise so, start at one end and then to the other side diagonally. First, a bit then fully screwed once every screw was placed. It has 4 screws.

You did it properly like a pro so problem is not on your side
I have 1 last tip for your problem because i had same problem
I think that problem is thermal taste, when thermal paste after some time starts to separate oil from itself
It's very small amount of oil but you can see it when you are applying paste on GPU or CPU
anyway that paste is useless because it conducts heat pretty badly

Now, you will probably ask: how to know did i have bad thermal paste (i am not sure that i saw any oil)?
Answer: easy, if you have bad thermal paste on idle temperature of the card is low , but when you start furmark test- Gpu temperature climbs almost instantly (a 2-3 seconds) to 65 or more steepen Celsius
If that is your case - than you should buy new thermal paste

 

[FreedomEclipse]

Sounds like you're not a fan of this problem OP

 

[PLAfiller]

May be a second hand aftermarket GPU cooler ? Something you can get on the cheap ? DBA.DK?

 

[plåtburken]

You did it properly like a pro so problem is not on your side
I have 1 last tip for your problem because i had same problem
I think that problem is thermal taste, when thermal paste after some time starts to separate oil from itself
It's very small amount of oil but you can see it when you are applying paste on GPU or CPU
anyway that paste is useless because it conducts heat pretty badly

Now, you will probably ask: how to know did i have bad thermal paste (i am not sure that i saw any oil)?
Answer: easy, if you have bad thermal paste on idle temperature of the card is low , but when you start furmark test- Gpu temperature climbs almost instantly (a 2-3 seconds) to 65 or more steepen Celsius
If that is your case - than you should buy new thermal paste

I the thermal paste is quite new. I bought it two weeks ago from now, I bought Arctic cooler mx-4. The GPU temp doesn't reach to 65 or more within seconds, it actually takes a few min and this is only when playing heavy titles such as, BF1 or Fallout4. Problem is that the GPU won't often even reach this temp since it will start the fans when it reaches 50 degrees and when it reaches that high, it starts to spin up the fans to around 50-60% and this is annoying as it goes up and down quickly. So you can imagine that the GPU temp will hover around 50 and below and the fan will spin over and over again. Worst part is that I can't control it with MSI afterburner or Asus GPU tweaker tool.

Sounds like you're not a fan of this problem OP

Haha, yeah.

May be a second hand aftermarket GPU cooler ? Something you can get on the cheap ? DBA.DK?

I haven't seen any cheap second hand aftermarket GPU cooler for my HD7770 and buying one online isn't cheap as I will have to pay for shipping(most of the time) and taxation.

 

[Laki89]

Last thing to try is AMD Overdrive - software included with graphics card driver
beside overclocking it gives user fan speed control on fixed (same) rpm all the time
You could set it on 40% or 50%, than start benchmarking your card while monitoring GPU temperatures (benchmark also include games that you mention before) and you are safe as long as temperatures are below 75 degrees Celsius

If that doesn't help, than only thing left to do is:

Have you considered hard wiring the fan to an external power source?

I've rigged GPU fans to 9V/7V in the past and left them at set speeds (if you go too low the temps rise, too high and its always loud - but its an option)

no more ideas /cry

 

[itsakjt]

Fan profiles can usually be controlled all the time from software (MSI Afterburner, ASUS GPUTweak and even AMD Control Center). I think the problem you are facing is due to having two similar software for the same thing that they might be conflicting with the settings. I would suggest you to keep either MSI Afterburner or ASUS GPUTweak and check the results. Also, go to AMD Control Center and check the settings there too. As I said, use any one software at a time.

Note: Even if you are not using more than one application but have them installed, some services can keep running in the background and might cause problems. Better to uninstall one of them.

With AMD Control Center, you cannot control the fan speed with temp changes but you can set it fixed by setting up a percentage.

Also, if you know what you are doing, GPU BIOS modding ain't that hard to do. You can take a backup of your factory BIOS and just edit the fan speed values using the appropriate BIOS editor (RBE should work for non-UEFI GPU BIOS). And even if you happen to do anything wrong, reflashing the card is easy if you have onboard VGA or processor graphics.
Based on your skills (from what I read here), I am confident enough that you will be able to mod the BIOS successfully to achieve your requirement if need be.
All the best.

 

[plåtburken]

Fan profiles can usually be controlled all the time from software (MSI Afterburner, ASUS GPUTweak and even AMD Control Center). I think the problem you are facing is due to having two similar software for the same thing that they might be conflicting with the settings. I would suggest you to keep either MSI Afterburner or ASUS GPUTweak and check the results. Also, go to AMD Control Center and check the settings there too. As I said, use any one software at a time.

Note: Even if you are not using more than one application but have them installed, some services can keep running in the background and might cause problems. Better to uninstall one of them.

With AMD Control Center, you cannot control the fan speed with temp changes but you can set it fixed by setting up a percentage.

Also, if you know what you are doing, GPU BIOS modding ain't that hard to do. You can take a backup of your factory BIOS and just edit the fan speed values using the appropriate BIOS editor (RBE should work for non-UEFI GPU BIOS). And even if you happen to do anything wrong, reflashing the card is easy if you have onboard VGA or processor graphics.
Based on your skills (from what I read here), I am confident enough that you will be able to mod the BIOS successfully to achieve your requirement if need be.
All the best.

Thanks for the good input.
I have only tried each application one at the time and uninstalled as I saw none of them working at all.
I checked out RBE but there is no information if it supports HD 7770 and the latest revision only seem to state about the HD 6000 series. My system have no onboard graphics card, so might be tricky to reflash incase something goes wrong.

 

[itsakjt]

Take a backup of your existing BIOS using GPU-Z. Then attempt to load that file in RbE. If it loads up and shows all the information, you are good to go.
Also, have you tried AMD Control Center? Does the manual fan control work from there?

EDIT: Also, was the fan profile always like this i.e. when the card was new? Or this is happening all of a sudden?

 

[AsRock]

Have you considered hard wiring the fan to an external power source?

I've rigged GPU fans to 9V/7V in the past and left them at set speeds (if you go too low the temps rise, too high and its always loud - but its an option)

This^, much better idea than taking the risk of flashing.

Depending on the orientation of the computer and how big the cpu cooler is you might be able to put a fan on the back\top of the card over the PCB, how ever make sure the fan is blowing toward the PCB as putting it the other way would risk the blades hitting parts on the PCB.

 

[plåtburken]

Take a backup of your existing BIOS using GPU-Z. Then attempt to load that file in RbE. If it loads up and shows all the information, you are good to go.
Also, have you tried AMD Control Center? Does the manual fan control work from there?

EDIT: Also, was the fan profile always like this i.e. when the card was new? Or this is happening all of a sudden?

I tried the AMD Control Center part, no as soon the GPU reaches 50C degrees the fans spin at 60% and the carousel of fans spinning up from 20 to 60% in just a matter of seconds as it tries to keep the GPU always at 50C degrees.
I will try the RBE part!

 

[itsakjt]

I tried the AMD Control Center part, no as soon the GPU reaches 50C degrees the fans spin at 60% and the carousel of fans spinning up from 20 to 60% in just a matter of seconds as it tries to keep the GPU always at 50C degrees.
I will try the RBE part!

Yes I get it. What I am saying is that are you manually able to control the fan speed using Global overdrive settings?