question for UK people: arctic silver paste versus antec silver paste?

I read on a US site that arctic silver is the thermal paste best used for CPUs, but Maplin don't seem to stock it but instead sell a silver paste with an antec logo on it (I'd link, but no copypaste or tabs, posting from Wii)

Anyone know if it's all the same, some types better or what? I was surprised at there not being much variety available

Go with Arctic Silver, plenty of other sites sell it.

What is the diff between Antec's one (also silver)?

Read this before agonising over different brands of thermal goop.

Thanks very comprehensive :)

There is a very good review of a lot of the of the different thermal pastes on the market at the link Here

Quote:

---

Originally Posted by schmunk

Read this before agonising over different brands of thermal goop.

---

2002? They use beef dripping and powdered egg :D

What a great artical
So what we need is a non-water based toothpaste

Quote:

---

Originally Posted by Lee @ SCAN

There is a very good review of a lot of the of the different thermal pastes on the market at the link Here

---

Thanks, much more up to date

Quote:

---

Originally Posted by Clunk

2002? They use beef dripping and powdered egg :D

---

Sorry, I forgot that physics changed in 2003. My mistake.

I remember hearing about some guy that used engineering grease, good to 1000 degrees or something ridiculous... he worked on an oil rig or something though.

Quote:

---

Originally Posted by schmunk

Read this before agonising over different brands of thermal goop.

---

Quote:

---

Originally Posted by Lee @ SCAN

There is a very good review of a lot of the of the different thermal pastes on the market at the link Here

---

I'm rather disappointed that neither of those even tried to explain what was happening, or could be happening, in any (great) detail. So I'll have a go :P

Heat transfer by conduction is governed by the following relationship:

Q = [ k.A.(T1-T2) ] / x

where

Q = heat flow (Watts)
k = thermal conductance (Watts per meter Kelvin)
T1 = temperature of hot side (Kelvin)
T2 = temperature of cold side (Kelvin) hot and cold refer to relative heat, not actual heat
x = thickness of medium heat is transfered through (meters)

In the case of a thermal interface, the chip is maintaining a temperature (T1) whilst producing heat (Q) which is travelling through an interface layer which occupies an area covering the chip/heatspreader (A) has a certain thickness (x) and has a certain thermal conductance (k). The heat passes though this to the base of the heatsink which is slightly cooler than the chip (T2). The difference in temperature (T1-T2) is what allows heat to flow through the interface, and as such, if you exchange a thermal paste for a worse one whilst keeping all the conditions the same, the chip will have to run hotter to compensate. Also if the interface area is much lower (e.g. covering an exposed chip rather than a heatspreader) then the differences between the thermal conductances of interface materials will be magnified.

So what does this mean?

Well it means that thermal resistance isn't the only thing that will affect the performance of a thermal grease. Thickness of the interface and the area it covers can have a great impact too, and both of these will be affected by variations in heat sink mounting, which is a phenomenon that has been observed by some reviewers, and even taken into account by some manufacturers who now advertise that their data takes this into account.

Examples:

Increased mounting pressure = more spreading of paste = lower thickness + higher area = reduced temperature difference required for a given heat load

Decreased mounting pressure = less spreading of paste = higher thickness + lower area = increased temperature difference required

Higher paste viscosity = less spreading with the same pressure = higher thickness + lower area = increased temperature difference required


Why did I type all this longwinded crap?

Because whilst the conclusions of these two articles are valid neither has even attempted to determine why their results are as they are. For instance:

Toothpaste isn't famed for its thermal properties, but it could quite eaily be the case that it spreads far more easily which resulted in an interface which, overall, matched those produced when you use specifically designed pastes.

IIRC Antec silver paste gives similar results to Arctic Silver 3.

I've got both the Antec stuff and Arctic Silver 5 and ill be damned if i can see any difference in temps.

I've just ordered the makings of a new box and got a bottle of this:

http://www.scan.co.uk/Products/Zalma...-inc-Cap-Brush)

It's easy to apply and is a thin film that just fills the voids - ideal.

No overclocking for me, though.

Ive put on literally HUNDREDS of heatsinks.... on loads of cpu's. The machine I'm running right now has had the heatsink on and off 5 times THIS WEEK as I try two different quads and a Duo out on this rig and my Dad's (in build) rig.

So I shall put my input ...er....in.

Arctic Silver 3 (and 5) is good stuff. I have also got OCZ paste, in a much bigger tube, of which I have lots left. I use it when cpu's are going on and off a lot to save my self the money. BUT the Arctic Silver 5 goes on for the final build. Always.

Why?

Cos when the heatsink comes off a year or more later, it's still doing a superb job. Cheaper ones (OCZ inc sadly) seem to give up and die after a long time.

I once used Copper Slip (copper grease for cars) and it was really good...for about 3 days. It smelled funny after that and kept crashing the pc.

When people spend a reasonable amount of dough on a good pc build, to be let down, long term, by crap thermal paste is like buying a cheap PSU. Not worth it.

Where's the balance of price vs effectiveness? Depends how often it's coming off and on. If it's going on to STAY, but the nice stuff :)

Yeah, I think what killed my old PC might've been the poor quality paste, it looked mostly dissolved/melted away when I pulled up the CPU, and the capacitors next to the CPU slot had brown marks/gunk on them like when AA batteries leak

It was a years old pentium 4 fujitsu-siemens premade PC, had been meaning to upgrade for ages anyway