Conductonaut for GPUs?

ubiquity · 4 Jun 2019 at 17:18

I am currently looking at getting a new thermal compound, and am intruiged by liquid metal options.

Appreciate that the difference in temperature is small compared to other, non-conductive alternatives, but at the price point it does sound appealing.

While I am confident enough to try this on a CPU, I have read different viewpoints with regards to replacing the stock compound on GPUs with liquid metal.

Anyone here using a liquid metal compound for their GPU?

Thanks,
Rich.

pmc25 · 7 Jun 2019 at 19:13

I always use Liquid Metal Ultra. Haven't tried Conductonaut yet, but people speak well of it, and it seems to marginally outperform Ultra in tests.

tyler_jrb · 7 Jun 2019 at 21:57

Watercooling or high end heatsink it does make a difference. Low-midrange air coolers not so much. Blower style coolers you would be talking next to nothing.

However i used to use it on GPU's both aircooled and watercooled but to me its not worth the pain of taping up the smd's and the time consuming application + the risk for a few degrees.

ubiquity · 8 Jun 2019 at 13:13

tyler_jrb said:
Watercooling or high end heatsink it does make a difference. Low-midrange air coolers not so much. Blower style coolers you would be talking next to nothing.

However i used to use it on GPU's both aircooled and watercooled but to me its not worth the pain of taping up the smd's and the time consuming application + the risk for a few degrees.

It would be for Alphacool Ice Wolf GPU Block... I am not sure I am confortable with my skills to try this on the GPU to start with, but after other posts on this forum regarding Kryonaught, not sure what the next best option is... Hydronaught?

pmc25 said:
I always use Liquid Metal Ultra. Haven't tried Conductonaut yet, but people speak well of it, and it seems to marginally outperform Ultra in tests.

Are you using this on your GPU, or just CPU?

pmc25 · 8 Jun 2019 at 15:55

ubiquity said:
Are you using this on your GPU, or just CPU?

Both. Though I should say the tests I've seen were on CPUs, and some were within margin of error, others very marginally better.

ubiquity · 8 Jun 2019 at 19:05

pmc25 said:
Both. Though I should say the tests I've seen were on CPUs, and some were within margin of error, others very marginally better.

Thanks, I am tempted to try the Ultra for both my CPU and GPU, but am worried I may end up messing up the GPU...

From what I saw the Ultra was tested as the top performer for AIO - https://www.tomshardware.com/reviews/thermal-paste-comparison,5108-8.html

Ross Thomson · 9 Jun 2019 at 17:57

no real point in using it on a GPU or IHS top.

ubiquity · 11 Jun 2019 at 09:29

Ross Thomson said:
no real point in using it on a GPU or IHS top.

On the GPU there is the added risk of a short circut, but on an ISH surely this gives you lower temps, for little capital outlay and minimal risk, no?

Apone70 · 11 Jun 2019 at 09:36

It is a risk using it on GPU's but as long as you use a liquid insulator then you should be fine. But is it worth it for maybe 5 degree's difference?

ubiquity · 11 Jun 2019 at 12:17

Hmm, risk vs reward

I I get a liquid metal for the CPU, then still tempted to attempt it for the GPU.

Did I see someone recommend nail polish as an insulator?!

Apone70 · 11 Jun 2019 at 12:22

Yeah nail polish is ok or just get some liquid insulation. Google it as i cant share a link on here

doyll · 11 Jun 2019 at 15:37

ubiquity said:
I am currently looking at getting a new thermal compound, and am intruiged by liquid metal options.

Appreciate that the difference in temperature is small compared to other, non-conductive alternatives, but at the price point it does sound appealing.

While I am confident enough to try this on a CPU, I have read different viewpoints with regards to replacing the stock compound on GPUs with liquid metal.

Anyone here using a liquid metal compound for their GPU?

Thanks,
Rich.

As others have said, there is more risk using liquid metal TIM, it is way more expensive, it is harder to get a good heat transfer print with and at best temps will only be a couple degrees cooler. That's hway many of us don't even consider using it.