[EOL] Arctic MX-5 is here!!Tests incoming! Completed. Now its MX-6 testing time!

[freeagent]

If comparing two sets of fresh applications of different pastes, I have a hard time believing 15c. That's like a cooler swap..

 

[nguyen]

If comparing two sets of fresh applications of different pastes, I have a hard time believing 15c. That's like a cooler swap..

Liquid metal on the inadequate 1080 cooler can drop temp like 10C

For Laptop it can be up to 20C

Of course this is comparing crappy stock TIM with Liquid Metal, with normal TIM the difference is less, and I wouldn't use LM on bare copper heatsinks on laptop. Although Asus do ship their 2021 laptop with LM pre-applied, even PS5 has LM.

 

[freeagent]

Lol I thought you were talking about TIM not LM.. duh

Sorry!

 

[nguyen]

Lol I thought you were talking about TIM not LM.. duh

Sorry!

Well Liquid Metal is literally thermal interface material .
Anyways the higher the operating temperature, the bigger difference that high quality TIM will make (that includes LM). So yeah save your most expensive TIM for GPU and Laptop, for CPU it doesn't matter so much.

 

[8BitZ80]

High quality thermal paste will only make discernable difference in worst case scenario, like GPU with poor stock cooler and laptop, where operating temperature is normally in the 80C range. Using good and highly viscous TIM in this case can shave off 5-15C easily.
But yeah on CPU it really doesn't make any difference though, unless you are deliding older Intel CPU.

A lot of users confuse operating temperature with the thermal transfer rate. The problem is that literally any cooling fan can be ramped down until the temperature is at 80C, so temperature does not always signify an extreme use case.

The reason why the temperature improvements are larger for GPU's and delidded silicon is because of the high thermal transfer rate (high TDP across small die area). That's where a high performing paste will make a big difference.

 

[Shrek]

I wish people would include graphite and Indium pads in the TIM comparisons.

 

[Space Lynx]

I wish people would include graphite and Indium pads in the TIM comparisons.

graphite always increased my temps by about 4 celsius, sometimes as high as 6 or 7... so i never bothered with that

 

[Shrek]

Paste it is then, and GD900 seems a winner if it does not age.

 

[ThrashZone]

I wish people would include graphite and Indium pads in the TIM comparisons.

Hi,
Yeah total waste on the IC Graphite thermal pads, I got a 90-90mm box and it was a waste of time and money for sure even the worst paste would do better or maybe even tooth paste would be better lol

 

[Shrek]

That's really good to know; saves a lot of time and trouble.

 

[8BitZ80]

Paste it is then, and GD900 seems a winner if it does not age.

Spoiler alert, it doesn't. The GD900 application on my 5600X is almost 6 months old and I'm still looking at around 75C under all-core load. I think reviewers running AIO water coolers were getting sort of mid-60's.

The real benefit of GD900 isn't daily use though, it's when you're constantly swapping out processors for testing. It'll take a long-ass time to go through 30g and it costs nothing, so you can basically experiment all you want.

 

[lexluthermiester]

The real benefit of GD900 isn't daily use though, it's when you're constantly swapping out processors for testing. It'll take a long-ass time to go through 30g and it costs nothing, so you can basically experiment all you want.

This is a very good point! Does it require time to set, like AS5 does?

 

[Mussels]

I'm not so sure one really needs a thermal conductivity much above say 5 W/meter K

If the power is 100 W, the area 3cm^2 and the thickness 0.05 mm then the temperature drop will be just over 1 C

I'd rather have a paste that lasts.

5800x would like to have a word with you about heat density

 

[8BitZ80]

This is a very good point! Does it require time to set, like AS5 does?

I did test that and couldn't find any obvious curing with GD900. On a quad core I recorded [86, 81, 86, 83] degrees with an old application and [86, 83, 86, 83] with a new one. Although keep in mind that I was replacing GD900 on both the heatspreader AND the die itself, so that would take into account the curing for multiple layers. I wouldn't be surprised if GD900 and MX-4 share similar chemistry.

 

[Space Lynx]

I still remember the first time using liquid metal on my gtx 1070 laptop... seeing that 20 celsius drop was insane...

but then i was an idiot and opened it up again to look and make sure it wasn't doing anything bad to the copper...

and that's when things went south... if I had just left it alone would have been perfect... eh.

it still works and all, but i never did go back to liquid metal.

 

[lexluthermiester]

seeing that 20 celsius drop was insane...

WOW! That IS a hellofa drop! The original TIM must have been both crap AND improperly applied!

 

[Space Lynx]

WOW! That IS a hellofa drop! The original TIM must have been both crap AND improperly applied!

liquid metal also benefits laptops more than desktops. i still don't understand the reasoning for that though

 

[nguyen]

liquid metal also benefits laptops more than desktops. i still don't understand the reasoning for that though

Because Thermal conductivity is not a constant figure, it varies with temperature. For Gallium (main component of LM) thermal conductivity increase with higher temperature, and the opposite for zinc oxide (normal paste).

So LM only works it magic for air cooling and high heat load and not so much for watercooling. On my watercooled 2080Ti, Conductonaut LM reduce temperature by 4C compare to Kryonaut (40C with Conductonaut and 44C with Kryonaut)

 

[Mussels]

heh i just helped a friend liquid metal his laptop and we saw about a 9C drop, but it STILL throttles with CPU and GPU on the same heatsink

 

[Space Lynx]

heh i just helped a friend liquid metal his laptop and we saw about a 9C drop, but it STILL throttles with CPU and GPU on the same heatsink

yeah, on my laptop in the BIOS I have all boosts turned off for CPU, i just let the CPU run at a lower wattage with 2.8ghz all 4 cores... and temps never break 65 celsius or so. and honestly games don't even seem to notice. if i leave BIOS at default, my CPU temps hits 95 celsius... lol

laptops are such crap design and always have been... its really sad

 

[freeagent]

I used to play NFSU on a Toshiba laptop with a P4 that loved to sit at 105c.. what a champ

 

[Falkentyne]

I wish people would include graphite and Indium pads in the TIM comparisons.

Indium pads (foil?), never used. If you mean indium foil, I only remember a few people saying they were going to try them but I never remember any results in the end. Just sporadic posts about it, and not interested in wasting more money myself on cooling experiments when I'm on fixed income.

Graphite pads: don't even bother if you are an overclocker! They're only good for stock. I tested the brand new Panasonic EYG-R and they were already a good 8C worse than Arctic MX-5 on a laptop GTX 1070 MXM card. On my overclocked 9900k, it reached 100C in one minute in Prime95 small FFT AVX disabled at 5 ghz. More than 10C worse than MX-5 (So it did worse on the higher wattage chip).

These pads are NOT suitable for overclocking at all. They perform about the same as a low end thermal paste. They are good for a "plant and forget" stock system however.

 

[8BitZ80]

liquid metal also benefits laptops more than desktops. i still don't understand the reasoning for that though

My understanding is that the low mounting pressure (i.e. thicker TIM) in laptops is where a higher thermal conductivity makes a big difference.

 

[lexluthermiester]

My understanding is that the low mounting pressure (i.e. thicker TIM) in laptops is where a higher thermal conductivity makes a big difference.

I would imagine that is a part of the equation.

 

[ThrashZone]

Hi,
After applying it looks like alien baby crap lol