facebook rss twitter

Heatspreaders: What difference do they make?

by Steve Kerrison on 8 February 2006, 22:21

Tags: Corsair

Quick Link: HEXUS.net/qaeqm

Add to My Vault: x

For the last few years, RAM, particularly that at the enthusiast end of the spectrum, has featured heat spreaders. The question being posed by many is whether they actually do anything.

One probably should wonder whether a heatspreader makes a RAM module any cooler, and if it does, what good does that do? How does a heatspreader affect the overclocking capabilities of a module. This is something which Corsair has recently investigated, and they have just provided us with the results of their testing.

Corsair's tests, which were internal, involved taking a RAM module purchased online, in this case an OCZ 512MiB PC3500 BH-5 based module, with a 'gold' XTC heatspreader, and testing it with the original heatspreader, two of their own, and no heatspreader whatsoever. Using a DFI Lanparty nF4 Ultra-D motherboard and Athlon 64 X2 3800+ with a lowered multiplier, they tested the module with the heatspreaders to find the highest stable overclock in each state.

OCZ's XTC heatspreader is a mesh style spreader with a honeycomb of holes, whereas Corsair have two types, a stamped aluminium spreader and a cast aluminium spreader. So what did Corsair find?

Firstly, without a heatspreader, they couldn't get past 250MHz. With the XTC heatspreader and their stamped spreader, they could get another 3MHz higher; 253MHz. Finally, with the cast aluminium spreader, 254MHz was achieved.

Now, first of all, the difference between these overclocks is hardly significant. So perhaps one thing we'd take from this is that heatspreaders make some difference, but not that much, really. Secondly, OCZ's XTC spreader performs the same as Corsair's 'solid' heatspreader, and of course Corsair seem to have a spreader that in their tests outperforms the XTC. The writeup documenting the tests has the following conclusion:

It was clear from the tests that the use of a heat spreader on the module increases overclockability in all cases. The XMS Pro provided the highest overclock; we believe this is due to the increased surface area provided by its increased size and ridged design. The Mesh heatspreader and the stock Corsair XMS Platinum heatspreader performed identically. One can conjecture that the additional surface area provided by the mesh approximately compensates for the reduction of the mass of the heat spreader.

One thing Corsair failed to provide was temperature readings. It would be interesting to see how hot the modules get with the different RAM spreaders. Also, how do heatspreaders affect lifespan? Perhaps heatspreaders are more important for letting a module run for longer at a higher voltage and clock rate.

Given that Corsair used OCZ's module and XTC spreader for comparison, we put to them the outcome of Corsair's testing, and they were good enough to provide a response.

"It is our policy to devote our resources to engineering superior products for enthusiasts, not address what competitors are doing to the press community," stated Ryan Peterson, CEO, OCZ Technology, Inc. “We will continue to focus our efforts on constantly improving our own products to fulfill the true needs of our customers.”

“OCZ is all about finding ways to introduce better solutions for consumers that demand the highest level of performance” said Alex Mei, VP of Marketing, OCZ Technology, Inc. “Rather than rely on our own internal testing we prefer to let the product speak for itself, and numerous independent media reviewers have already found the XTC results very positive when it comes to cooling.”

We expect thoughts from readers on the test results will be as mixed as those from Corsair and OCZ. Is it about time we did some heatspreader testing in the HEXUS.labs, or is there nothing there worth testing for? Do let us know what you think.



HEXUS Forums :: 11 Comments

Login with Forum Account

Don't have an account? Register today!
I think we've known this for a while. And considering it's an internal test, it's not necessarily perfectly unbiased, I wouldn't say. It certainly doesn't have statistically significant resultant benefits.

Despite this, I bought (cheap) RAM modules with heatspreaders for two reasons.

1. They look better, nice coloured metal bits always looks better.
2. They're easier to fit and play with when they have a protective covering, I don't worry about breaking the thing.
I would consider OCZ and Corsair to be competitors, if I owned a company, I would never allow a press release saying our competitor was atleast as good as us. So I would take anything , other than a truly independant review, with a pinch of salt.

That said, heat spreaders in general have been shown to provide some performance gain, I can understand this - they will have a larger surface area/chip. And they do look a lot better/safer.

The only thing I have against them, is they remove the fragility feeling of ram modules meaning they now feel so safe that you THINK you can throw them across the room and and jump up and down on them they will still work. When obviously they probably won't. But I don't play tennis with my modules, just look at them in my motherboard, so all is happy.

Nox
Since heatspreaders doesn't actually touch the chip, but relying on the heat-conducting adhesive tape… If you're not stressing the ram it can reduce thermal stress of the ram, but if you're pushing the ram real hard (i.e. 3.3V or DDR600 kind of stuff) And the thermal gradient of the adhesive tape just hinder cooling so much….The chips may be at 60'C and the heatspreader is only 45'C…Blowing a 12cm fan wouldn't help… Thats why many G.Skill modules didn't come with a heat spreader…
Full disclosure, I work for Corsair, so you can say that I'm biased now and yell mean things at me if you'd like.

How would air provide better thermal transfer than even heat tape?

Almost any material that contacts the IC conducts heat away from it better than air could.

It's the same principle with a CPU. If you have a hot processor, you wouldn't just throw a high speed fan on it and hope that worked, right? You'd put a heatsink on it first to provide more surface area. Once that was done, the fan is working in conjunction with the heatsink to move heat away from the cpu and into the air in a much more efficient manner.

Since heatspreaders work in a similar (yet not identical) way, the theory stands that regardless of the manner of adhesive, a heatspreader should be more effective than no heatspreader. I can't say why G.Skill don't include heatspreaders on some of their older modules, but you'll notice many of their newer modules do.

Of course, as implied by their name, the heatspreader just takes heat from each individual IC and spreads all that heat out over a larger surface area to facilitate in cooling. As mentioned in the article, much of the IC heat is actually transferred through the pins and into the ground plane of the PCB, so there's only so much a standard heatspreader can do.
Yes please - I would love Hexus to do some proper analysis on this kind of thing.

Personally I think no heatspreaders at all would be best in terms of temperature. Yes you transfer heat away for short while, but then you immediately store the heat in the heatspreader capacity which would then stop further transfer away from the chip. Most spreaders are smooth and shiney, and I can't believe the heat transfer from this kind of surface to air would be higher than from non-heatspread-covered sticks.

However, heatspreaders do look cool, and it's something you can feel good about paying extra money for, which is completely what the ram market is all about. That's being cashed in on more and more these days which is why newer g.skill sticks have them, and why companies like Corsair and OCZ use heatspreaders as a marketing thing.

The added protection while handling is also a bonus.