Optimized CPU Cooling with Top-Down Heatsinks

The staff at Benchmark Reviews have an article posted today that covers optimized CPU cooling with top-down heatsinks. Suck? Blow? Read the review to find out.

Not all that long ago we tested the Cooler Master Geminii S524 Ver 2 CPU Cooler, and noted how well it performed compared to many other high-end heatsinks. We looked at the temperatures, measured the cooling performance, compared it to the original GeminII S524 Heatsink, and then gave our rating of the product. What we didn’t do is discuss the benefits of a horizontal heatsink, and how they might actually help you reach an even higher overclock on your computer system.

With most modern cases providing plenty of vents and fan mounts, wouldn't the airflow within the case provide adequate cooling for the components surrounding the CPU on the motherboard? I fail to see where the top-down approach is advantageous in a properly configured chassis, unless it is a particularly small form factor, like one of those NCASE M1s or similar where you are relying on the airflow from the CPU fan almost exclusively to cool those surrounding components.

The article is a hypothesis. Experimentation is needed to reach a conclusion.

Full Disclosure: The product sample used in this article has been provided by Cooler Master.

Click to expand...

Without actual benchmark results, this "article" reads more like a shill piece for Cooler Master. Show me where this configuration comes close to beating any recent Noctua tower cooler and I'll STFU.

I was expecting some actual testing... none was done. Fail.

Now as far as top down coolers... oh yeah, let's blow the heated air back onto the board, RAM, and CPU socket area. That is going to improve cooling - NOT!

Putting the fan in a pull type configuration will help, but you are still not going to be getting very much cool air movement to help cool the RAM and surrounding area.

I'll take my tower coolers and actual cool air from case fans to cool the RAM and surrounding area.

It's the old Alpha Heat Sync debate again.

I was disappointed that they didn't provide any numbers, that implies that they aren't very good and the rest of the article is likely a fluff piece. I'm not suggesting anything, but that is how it looks.

The cooler looks similar to the one I had on my Q6600, a Silverstone NT06 and I assume the advantage of this is the same as that was. The fan cools the board so you don't actually need any case fans (yes, they're nice, but not needed).

This was a dumb article, the reason to use coolers this this are for small cases HTPC setups etc.

Noctua makes coolers like the NH-C14S for this purpose. It is meant to push air down through the heatsink and towards the motherboard. This is the way it works best, because you want t push air through the heatsink not pull and you want to make sure air is getting to the board components.

Modern tower heatsinks are obviously better, but if you have to use a low profile you want the air flow pushing down. I don't know exactly why that works better and you might even find some cases where it doesn't , but those are probably cases that don't require this kind of cooler in the first place. My guess would be that it fights the airflow that should normally reach the motherboard components while also doing a poorer job of cooling the CPU which leads to higher temps all around.

I still believe vertically oriented heatpupes work better even with wicks and microchannels.

Top down coolers are better because your motherboard usually relies on the air flow to cool VRMs. Hence top down coolers are better.

Ashbringer said:

Top down coolers are better because your motherboard usually relies on the air flow to cool VRMs. Hence top down coolers are better.

Click to expand...

The only time this helps at all is if there is absolutely no other airflow to those components.

In general, trying to cool the other components with hot air from the CPU cooler is not the right way to go about it.

To be fair, it's not the first time they publish an article with an open ended question or hypothesis like that, and it hasn't always been to semi blatantly push a product (that they already reviewed).

Clicked thru a couple of their old thermal paste and application articles, kinda interesting but they still left more stuff unanswered... Seemed like some of the dual line (parallel or crossed) application methods could be a viable alternative to the center pea style.

The TIM/paste roundup article goes at length about how real world HSF testing is the worst way to tease out the differences between them and proposes alternative test methods, but then concludes readers won't care and discards them.

Anyway, case airflow should be enough for VRM considering how little surface area they have, but a lot of cases have pretty poor airflow across the CPU area with many only having an exhaust at best.

I just really dislike the standard ATX S shaped airflow path.

no real world tests, only fun theories.

put money with the numbers, all CPU coolers from years gone by were top-down.
Ever since the introduction of tower CPU coolers, CPU temps have gone down, noise has gone down, all is well.

the only case for top-down coolers is when you don't have a case that's high enough to fit a tower cooler (HTPC cases come to mind)

in all other cases (pun intended) the tower cooler will offer superior cooling and overall lower case temperatures.

The key to good cooling is flowing cool air to components.
To do this we have to flow all air warmed up in the process of cooling component out of the case without it mixing with any of the cool air flowing to components.
This is near impossible to do with downflow coolers.
Add to this the problem of GPU coolers that dump their heated exhaust in all directions (like downflow coolers do) and we end up with a fuster cluck / fubar of epic proportions.