Please share your liquid metal success (or horror) stories

[Guest]

Got some Conductonaut incoming. The plan is to use it on CPU and GPU in the hope of lowering temps by a few degrees each. Both have Copper heat plates so corrosion shouldn't be an immediate concern.

 

Ryzen 2600X with a CM LiquidLite AIO.

Sapphire Pulse OC RX 5700XT stock cooler.

 

1- Have any of you used TG Conductonaut on either of the 2 above components?

2- Do you have any general success or horror stories to share about using liquid metal?

Thanks

[For Science!]

4 minutes ago, Shabba said:

 

Almost everybody i know that used liquid metal on bare copper has suffered from the liquid metal alloying with the copper over time and losing performance. I don't think it makes a tonne of sense to use liquid metal for either situation and would stick to high-performance thermal paste instead.

[Shimejii]

Not going to sugarcoat this, LM on a 2600x Is pretty fucking stupid. Its Not a hot cpu, so you dont need to use it. Do not use LM for Heatspreader to Cooling plate, its generally not a good time and especially if anything were to drip out.

 

Liquid metal cooling a GPU is not a good idea at all. The coolers are not designed for LM, and you have to do a lot for it to even make it safe to use on it. Generally you arent going to gain all that much performance, a few degrees is nice worth the risk to component here.

[Guest]

28 minutes ago, Shimejii said:

Not going to sugarcoat this, LM on a 2600x Is pretty fucking stupid. Its Not a hot cpu, so you dont need to use it. Do not use LM for Heatspreader to Cooling plate, its generally not a good time and especially if anything were to drip out.

 

Liquid metal cooling a GPU is not a good idea at all. The coolers are not designed for LM, and you have to do a lot for it to even make it safe to use on it. Generally you arent going to gain all that much performance, a few degrees is nice worth the risk to component here.

Thanks for your comments but I'm trying to get away from the anecdotes and get some real experience (as I asked in my 2 questons).

With that in mind do you have any first hand experience of using liquid metal on a 2600X and or any general experience of using liquid metal and what did you find?

 

In response to your specific point about how hot a 2600X gets depends partly on the OC and I run mine pretty high at 4.2 to 4.3GHz on all 6 cores so some temp reductions for the cost of £10 or so would be nice.

The GPU may or may not be designed for liquid metal but I am wondering if anyone out there has experience of using it on a GPU and what happened? Again if you have some experience here that'd be great to hear.

 

Some sources that gve me some encouragement are:

This test shows an 8c reduction on the stock cooler

Gamers Nexus says Copper is "for the most part fine" (direct quote).

 

I'm just really after some opinions from people who have done it.

 

Cheers

[Guest]

24 minutes ago, For Science! said:

Almost everybody i know that used liquid metal on bare copper has suffered from the liquid metal alloying with the copper over time and losing performance. I don't think it makes a tonne of sense to use liquid metal for either situation and would stick to high-performance thermal paste instead.

I am using Kryonaut at the moment, are there any pastes out there that perform like LM?

 

The people you know who use LM on copper. Did they just get staining or actual corrosion? And over what amount of time?

 

By the way my AIO was like £35 so it's not a big deal if it gets a bit of corrosion over the course of say a year or two. The GPU same, I'll likely upgrade it in the next 2 years so as long as there's no harm done in the near term I'm fine with it

[For Science!]

29 minutes ago, Shabba said:

I am using Kryonaut at the moment, are there any pastes out there that perform like LM?

 

The people you know who use LM on copper. Did they just get staining or actual corrosion? And over what amount of time?

 

By the way my AIO was like £35 so it's not a big deal if it gets a bit of corrosion over the course of say a year or two. The GPU same, I'll likely upgrade it in the next 2 years so as long as there's no harm done in the near term I'm fine with it

mainly staining, but it meant there wasn't enough liquid metal left over for good contact. I think usually people realize something is wrong within a year. I think kryonaut is a good choice and would personally leave it as is.

[Guest]

37 minutes ago, For Science! said:

mainly staining, but it meant there wasn't enough liquid metal left over for good contact. I think usually people realize something is wrong within a year. I think kryonaut is a good choice and would personally leave it as is.

from what I gather staining has no impact on performance so I'm okay with that

[For Science!]

Just now, Shabba said:

from what I gather staining has no impact on performance so I'm okay with that

Yes, but you will have to reapply the LM regularly until you saturate the interface, if you don't, the performance will be worst than paste.

[Guest]

3 minutes ago, For Science! said:

Yes, but you will have to reapply the LM regularly until you saturate the interface, if you don't, the performance will be worst than paste.

also okay with that just wanting to hear about peoples' experience

[Bad5ector]

I delidded my 7700k and replaced the garbage Intel TIM with some LM. Dropped my temps by 20 degrees almost while under load. 

 

AIDA64 stress test before LM, was high 80s creeping into 90s with custom water loop an and fans on full. 

 

AIDA64 stress test after delid and LM, mid 60s. I could even get away with not cranking fans up to full and just let my default fan curve so it's thing. 

 

Not sure if this experience helps you any for your application (2600x) as I don't think you're planning on delidding your chip. Not sure how much benefit you'll see with LM over a more standard TIM in this scenario. 

[Guest]

Well I YOLOd it because WGAF it's only money. I put conductonaut on my Sapphire Pulse OC RX5700XT, didn't even bother nail varnishing the bits around the sides.

 

It booted even though I managed to squirt conductonaut all over the bloody FAN SHROUD AND DINING TABLE place and it took me ages to clean it up (so that's a win straight away) The initial results are very very interesting.

 

Idle temps are down by about 10c to 15c on average I have even seen it dip as low as 29c with the fans spining at inaudible levels which wasn't even possible previously with fans on 100% at idle!

 

Valley Bench

 

I have now run this a few times and the results always within margin of error. This was just the most recent one which was not actually the best.

 

During the valley test, same valley settings, same radeon settings the max temp reached with Conductonaut was 63c on the GPU versus 77c with Kryonaut and the max Hotspot temp was 74c versus 109c with Kryonaut!

 

The valley score also improved on Conductonaut scoring 4638 with max FPS 204.7 versus 4473 an max FPS of 185.0 on Arctic MX-4

 

Side note

 

Memory and VDDC VR temps were up slightly from 76c to 82c and 59c to 69c respectively but there could be 2 possible reaons for that.

1- The GPU fans never went above 1590RPM the entire Valley run where as previously on the thermal paste they went to near 3000RPM due to the stock fan curve.

2- Perhaps the thermal pads aren't making as good contact with the rad and backplate now that it has been taken apart and put back together.

 

Some graphics

 

 

 

[For Science!]

3 hours ago, Shabba said:

Well I YOLOd it because WGAF it's only money. I put conductonaut on my Sapphire Pulse OC RX5700XT, didn't even bother nail varnishing the bits around the sides.

 

It booted even though I managed to squirt conductonaut all over the bloody FAN SHROUD AND DINING TABLE place and it took me ages to clean it up (so that's a win straight away) The initial results are very very interesting.

 

Idle temps are down by about 10c to 15c on average I have even seen it dip as low as 29c with the fans spining at inaudible levels which wasn't even possible previously with fans on 100% at idle!

 

Valley Bench

 

I have now run this a few times and the results always within margin of error. This was just the most recent one which was not actually the best.

 

During the valley test, same valley settings, same radeon settings the max temp reached with Conductonaut was 63c on the GPU versus 77c with Kryonaut and the max Hotspot temp was 74c versus 109c with Kryonaut!

 

The valley score also improved on Conductonaut scoring 4638 with max FPS 204.7 versus 4473 an max FPS of 185.0 on Arctic MX-4

 

Side note

 

Memory and VDDC VR temps were up slightly from 76c to 82c and 59c to 69c respectively but there could be 2 possible reaons for that.

1- The GPU fans never went above 1590RPM the entire Valley run where as previously on the thermal paste they went to near 3000RPM due to the stock fan curve.

2- Perhaps the thermal pads aren't making as good contact with the rad and backplate now that it has been taken apart and put back together.

 

Some graphics

 

 

 

Just for my curiosity, For the previous kryonaut application on the gpu, how did you do apply it? (Pea, Line, etc).

[Guest]

1 hour ago, For Science! said:

Just for my curiosity, For the previous kryonaut application on the gpu, how did you do apply it? (Pea, Line, etc).

I can't remember if I was using kryonaut or mx4 because I had done some testing on pastes but I'm pretty sure it was kryonaut on my gpu.

I applied same as I always do, pea size in middle.

When I removed the paste before applying conductonaut the application seemed to have been perfect, full coverage of the die with a tiny bit of excess over the edges so I'm sure the previous paste application had been good.

Also I should mention that for the conductonaut application I put the conductonaut onto the cue tip and applied it onto the die like that. I then repeated on the heatsink copper plate ending up with two very thin but consistent layers of liquid metal

[Neggy-Z]

4 hours ago, Shabba said:

Also I should mention that for the conductonaut application I put the conductonaut onto the cue tip and applied it onto the die like that. I then repeated on the heatsink copper plate ending up with two very thin but consistent layers of liquid metal

That's probably the easiest way to do it, how i've always done it anyway.

If the heatsink is nickel plated then it should last a good couple of years.

Not sure how conductonaut ages but i used, and left, coollaboratory liquid pro on my laptops cpu/gpu for a little over 3 years, when i finally noticed temps starting to go get worse i replaced it with kryonaut and temps still weren't as good as the aged dried out LM.