[Xbitlabs] Swiftech H2O-320 vs. Thermalright Archon (really surprising results)

[996GT2]

Full article: http://www.xbitlabs.com/articles/coolers/display/swiftech-h20-x20-edge_8.html#sect0

The results are frankly quite shocking. I was expecting the watercooling system with a 3x120 Radiator to completely destroy the Archon, but it doesn't.

With the Apogee XTL waterblock, the water cooling system with fans @ 2140 RPM and pump @ 4500 RPM is only 8C better than an Archon. For 3x the cost, that's not such a major improvement at all. And at similar noise levels, the water cooling system is only about 2C better.

Anyone think that maybe xbit messed up their testing somehow? I still find the results hard to believe. I know the top air coolers do better than the sealed water systems like the H70, but it's really a stretch to think they're this close to a full-blown watercooling system with a 3x120 radiator and a powerful pump.

And the noise levels speak for themselves in favor of the air cooler:

 

[QuantumPion]

I have a similar setup (i7-920 @ 4 ghz, swiftech apogee xt, 360mm rad with 3 900 rpm fans, d5 pump) and my max load temp is like 55 degrees. Before I went to WC I had a similar tower air cooler and my max temp was around 75 degrees. A 20 degree difference is about right. According to that graph they only have a 10 degree improvement. That seems too low to me.

 

[Zargon]

maybe the system needed bled?

 

[Meghan54]

I wonder why XBit and the OP are focusing on results that include doubling the number of fans that the Archon comes with out of the box, and two better fans at that, while keeping the Swiftech system bone stock. Biased results?

Comparing the two cooling systems bone stock, you have a 13C difference, not 8C, which is a pretty big difference.

Then, the fans on the Swiftech system are suck completely. If XBit wanted to be fair, use a set of Gently Typhoon 1800rpm fans (which would both perform much better than the stock ones and also run quieter at the same time) like Swiftech has used during its own internal testing of items.

And don't be too surprised by the results. Decent radiator, third or fourth best cpu block will give decent results but not "top drawer" results. And as was suggested, if the loop wasn't purged well, air pockets could easily exist resulting in less than optimal performance.

Then add in the crap TIM that's included with the Swiftech system (certainly nowhere near the quality of TIM that Thermalright includes with the Archon) and you have a host of reasons the Swiftech system was only 13C better performing than the Archos.