Thermal Pad Necessary?

[PingPonguin]

So I have an Asus GTX 1070 Ti Cerberus, and I saw that there are only thermal pads on the MOSFETs and they contact the heatsink. However, there are no thermal pads on the capacitors, the VRAM,  chokes/inductors (the block things), nor the backplate. Does that matter if I am not overclocking the 1070 Ti Cerberus? And if I do overclock it, does it make a difference in longevity or allow for slightly better performance? Thanks in advance for help!

 

[Jurrunio]

1 minute ago, PingPonguin said:

capacitors, the VRAM,  chokes/inductors (the block things), nor the backplate

these things don't really need thermal pads, as long as airflow from the fans can still reach them. Of course it's a good idea to put thermal pad yourself so there's contact between them and the heatsink, but it's not necessary.

[-rascal-]

Just now, PingPonguin said:

So I have an Asus GTX 1070 Ti Cerberus, and I saw that there are only thermal pads on the MOSFETs and they contact the heatsink. However, there are no thermal pads on the capacitors, the VRAM,  chokes/inductors (the block things), nor the backplate. Does that matter if I am not overclocking the 1070 Ti Cerberus? And if I do overclock it, does it make a difference in longevity or allow for slightly better performance? Thanks in advance for help!

 

 

You don't need and thermal pads on the capacitors.

Some GPUs have thermal on the chokes, while others do not -- they should be adequate from the cooling fan blowing air over them.

 

Thermal pads (or heatsinks) on the VRAM chips ...that will vary.

I find lower tier GPUs do not have any cooling over those -- just cooled by the fan blowing air down over them.

[PingPonguin]

Just now, Jurrunio said:

these things don't really need thermal pads, as long as airflow from the fans can still reach them. Of course it's a good idea to put thermal pad yourself so there's contact between them and the heatsink, but it's not necessary.

But if I do, will it make a difference in the longevity of those components or better performance/overclocking potential?

[Jurrunio]

1 minute ago, PingPonguin said:

But if I do, will it make a difference in the longevity of those components or better performance/overclocking potential?

lifespan: inductors - none, caps & VRAM - yes, but long enough even without pads, backplate - how can you degrade a backplate?

 

performance: inductor - none. It's just a wrapped up wire. VRAM - yes and no, 25-50MHz if one's really well cooled compared to one that's not, but that doesnt show anything noticeable in games. caps - no. Backplate, no because even a backplateless graphics card don't do better than with a backplate.

[-rascal-]

7 minutes ago, PingPonguin said:

But if I do, will it make a difference in the longevity of those components or better performance/overclocking potential?

 

Longevity?

The GPU, as it is now, should last longer after it becomes obsolete.

 

Better overclocking, and potentially better stability?

Probably.

 

I bought a used ASUS ROG R9-280X Platinum Matrix (I bought it as a show piece).

The owner reported graphical glitches and artifacts when the GPU is under heavy use.

Did some digging around online, and pointed out the VRAM chips is the cause -- they are getting too hot (heavily overclocked from factory).

Bought some copper heatsinks and mounted them on the affect VRAM chips, and haven't seen the artifacts since.

[PingPonguin]

1 minute ago, Jurrunio said:

lifespan: inductors - none, caps & VRAM - yes, but long enough even without pads, backplate - how can you degrade a backplate?

 

performance: inductor - none. It's just a wrapped up wire. VRAM - yes and no, 25-50MHz if one's really well cooled compared to one that's not, but that doesnt show anything noticeable in games. caps - no. Backplate, no because even a backplateless graphics card don't do better than with a backplate.

Oh I should have clarified about putting thermal pads on the other side of the PCB so it would sink to the backplate. Thanks for the help though!

[KarathKasun]

Lifespan of caps goes up exponentially with lower temps.

 

At 105c they are good for ~3000 hours.

At 90c they are good for ~12000 hours.

At 65c they are good for ~200000 hours.

[PingPonguin]

10 minutes ago, KarathKasun said:

Lifespan of caps goes up exponentially with lower temps.

 

At 105c they are good for ~3000 hours.

At 90c they are good for ~12000 hours.

At 65c they are good for ~200000 hours.

How much of a difference in temps do you think there would be for:
1. just air going it like it is now
2. the caps have a thermal pad and are heat sinked to the fin stacks?

[KarathKasun]

11 minutes ago, PingPonguin said:

How much of a difference in temps do you think there would be for:
1. just air going it like it is now
2. the caps have a thermal pad and are heat sinked to the fin stacks?

You would need to install some K-Type thermal probes to check TBH.

[Skander1345]

2 hours ago, PingPonguin said:

How much of a difference in temps do you think there would be for:
1. just air going it like it is now
2. the caps have a thermal pad and are heat sinked to the fin stacks?

Depends. The heatsink might get hotter than the caps themselves because of the VRM/other components. You might end up driving their temps higher.

 

Caps don't really generate too much heat, I would think that passive cooling with residual airflow over them and no transfer medium is more than sufficient.

 

You'll need a probe and test before/after for a definitive answer.