News AM5 copper guard stops you from making a mess on your Ryzen CPU — also improves heat dissipation

[Admin]

Japanese PC accessory Ainex offers a copper AM5 plate guard to prevent you from messing up your PC build with too much thermal paste.

AM5 copper guard stops you from making a mess on your Ryzen CPU — also improves heat dissipation : Read more

 

[TechieTwo]

Apparently many folks missed the memo that TIM should be installed as thin as possible because it is only intended to fill in minute imperfections in the heatsink and CPU lid. The thicker the TIM the less heat transfer to the cooler.

 

[HardwiredWireless]

Well that's just lazy. If you don't have the skill to spread some paste neatly on a piece of metal you probably should stay away from building computers we're working with anything electronic for that matter.

 

[Alvar "Miles" Udell]

If you're applying TIM like this, you are not appling it correctly. The goal is to fill the microscopic grooves in the IHS and heatsink base, not frost it like a cake...

 

[Snarks Domain]

Or better yet, just switch to a Phase Change Material like Honeywell PTM7950 or similar.

It typically outperforms thermal paste and last an extremely long time.

I recommend cutting it about 3mm smaller on all 4 sides. Makes it easier to do 2nd peel. You'll still end up woth full coverage once it melts and flows out 80% of the material.

Now, running an actual contact (anti-warping) frame is something I can get behind, still with a PCM/PTM, or paste of your choice.

 

[CelicaGT]

If you're applying TIM like this, you are not appling it correctly. The goal is to fill the microscopic grooves in the IHS and heatsink base, not frost it like a cake...

Right?!

In my experience TIM is often over applied, and reduces cooling efficiency. We use this stuff at work on an industrial scale for 800VDC liquid cooled motor drives. A silk screen mask and roller are used to apply TIM to the cooling blocks in the optimal pattern and quantity. It's always way less than one might think, as in...a little more than a skiff. That's all. If that little skiff will transfer enough heat on a $75K drive then I'm sure it's enough for a wee little CPU.

 

[aliwalyd3]

Apparently many folks missed the memo that TIM should be installed as thin as possible because it is only intended to fill in minute imperfections in the heatsink and CPU lid. The thicker the TIM the less heat transfer to the cooler.

Wrong. The pressure from the cooler will ALWAYS push out any excess.

 

[aliwalyd3]

Right?!

In my experience TIM is often over applied, and reduces cooling efficiency. We use this stuff at work on an industrial scale for 800VDC liquid cooled motor drives. A silk screen mask and roller are used to apply TIM to the cooling blocks in the optimal pattern and quantity. It's always way less than one might think, as in...a little more than a skiff. That's all. If that little skiff will transfer enough heat on a $75K drive then I'm sure it's enough for a wee little CPU.

Completely incorrect. More will never reduce cooling efficiency. Any and all excess will always be pushed out the sides from the pressure applied by the cooler.

 

[DingusDog]

Is that article picture from the Verge PC guy? Way too much paste.

 

[DingusDog]

Completely incorrect. More will never reduce cooling efficiency. Any and all excess will always be pushed out the sides from the pressure applied by the cooler.

Great. And where does all that extra paste go? Down the sides of your socket that's where. Sure it's harmless unless using conductive TIM like liquid metal but why use that much in the first place?

 

[Amdlova]

Today I just squeeze the tube over cpu and it's done. Will work, will be hot 1 or 2°c will not kill anyone.

 

[Kashmir35]

I bought Noctua thermal paste recently and they advertised it for AM5 and indeed they had that kind of guard but it was made of plastic. Anyone with experience with that product, is plastic a good material to put there?

 

[usertests]

So people are pasting wrong. Maybe it's still a good idea to have a copper guard like this, to conduct a little more heat away from the awkwardly shaped AM5 CPUs?

 

[sftwn]

Great. And where does all that extra paste go? Down the sides of your socket that's where. Sure it's harmless unless using conductive TIM like liquid metal but why use that much in the first place?

Because you'd rather have too much than too little. Dry gaps are harbingers of heat pockets and the visual of overflowing compound is a positive sign that you didn't miss any. You wouldn't under-grease a bearing, so why would you under apply thermal compound?

This is one of those silly Internet arguments that is somehow worse than audiophiles claiming they can hear the crickets chirp outside the recording booth. Heatsinks and spreaders were made by geniuses who knew what they were doing to ensure the compound spreads evenly between the clamp faces. There have been countless tests that show under applying causes higher thermals for decades.

Even when I worked in aerospace, the engineering guides recommended and MIL specs required applying until compound leaked out of the edges of the clamp. Even in R&D where I worked, if you didn't have overflow, it could be flagged as a defect. I promise we didn't have some secret contract to buy as much thermal compound as possible.

Liquid metal is an exception, but honestly nobody really needs it except in really tight spaces like laptops and slim profiles. In those cases, they sell pre-portioned applicators with dams to prevent overflow.

 

[COLGeek]

Completely incorrect. More will never reduce cooling efficiency. Any and all excess will always be pushed out the sides from the pressure applied by the cooler.

Well, you assume the user tightens the cooler fully/properly. Many don't. They are afraid of over-tightening, then too much creates problems.

 

[sftwn]

Well, you assume the user tightens the cooler fully/properly. Many don't. They are afraid of over-tightening, then too much creates problems.

I'd still rather have too much in this case than too little. An uneven thermal interface is still more efficient at heat conduction than air.

 

[COLGeek]

I'd still rather have too much in this case than too little. An uneven thermal interface is still more efficient at heat conduction than air.

Considering that air pockets/gaps could remain when under-tightened, I think not.

All would agree, I think, that it best to get it right from the start. It isn't that difficult. More is not always better.

Noctua includes a plastic "shield" with some coolers that does the same as this article's subject, it only deals with the goop reduction, not with any cooling properties.

 

[sftwn]

Considering that air pockets/gaps could remain when under-tightened, I think not.

I think yes. Too much is still better in this case. Even if you didn't apply it in one continuous blob in the middle (which prevents this entirely), think for a second: What is a dry spot? It's a big air pocket. It's better to have compound with air pockets than no compound at all.

Again, it comes down to the PC building community inventing these myths that no other industry who uses thermal compound has. I have no idea where it came from, but it's super silly.

Edit: Since it came to mind. I've worked with devices with such massive and fast heat ramps that it required special thermal potting compounds that had to be sealed under vacuum. The columns of air bubbles that could form, even at sub-mm thicknesses, could cause large gradients across the surfaces and warp them. CPUs and GPUs don't get that hot, aren't large enough to have that problem, and don't dynamically ramp up and down as quick. If you whipped air with a stand mixer into the thermal compound and used it, as long as you used enough to overflow, you'd still get better thermal performance than if you dry mounted a cooler.

 

[COLGeek]

I think yes. Too much is still better in this case. Even if you didn't apply it in one continuous blob in the middle (which prevents this entirely), think for a second: What is a dry spot? It's a big air pocket. It's better to have compound with air pockets than no compound at all.

Again, it comes down to the PC building community inventing these myths that no other industry who uses thermal compound has. I have no idea where it came from, but it's super silly.

Edit: Since it came to mind. I've worked with devices with such massive and fast heat ramps that it required special thermal potting compounds that had to be sealed under vacuum. The columns of air bubbles that could form, even at sub-mm thicknesses, could cause large gradients across the surfaces and warp them. CPUs and GPUs don't get that hot, aren't large enough to have that problem, and don't dynamically ramp up and down as quick. If you whipped air with a stand mixer into the thermal compound and used it, as long as you used enough to overflow, you'd still get better thermal performance than if you dry mounted a cooler.

Aside from semantics and talking passed other. Too much or too little is obviously a concern, with negative effects.

The issue of too much and an under-tightened cooler will result in overheating/throttling. Not really a debate.

Your points have been made, but I think the context of response may have been lost.

 

[HardwiredWireless]

Or better yet, just switch to a Phase Change Material like Honeywell PTM7950 or similar.

It typically outperforms thermal paste and last an extremely long time.

I recommend cutting it about 3mm smaller on all 4 sides. Makes it easier to do 2nd peel. You'll still end up woth full coverage once it melts and flows out 80% of the material.

Now, running an actual contact (anti-warping) frame is something I can get behind, still with a PCM/PTM, or paste of your cho

Or better yet, just switch to a Phase Change Material like Honeywell PTM7950 or similar.

It typically outperforms thermal paste and last an extremely long time.

I recommend cutting it about 3mm smaller on all 4 sides. Makes it easier to do 2nd peel. You'll still end up woth full coverage once it melts and flows out 80% of the material.

Now, running an actual contact (anti-warping) frame is something I can get behind, still with a PCM/PTM, or paste of your choice.

Or a graphene thermal transfer pad.