MaximumPC 2005 05

(Dariusz) #1

OUR TESTING


METHODS
Each kit was installed
in a Cooler Master
Cavalier ATX mid-
tower case, using an
Abit AA8XE Fatal1ty
motherboard and an
Intel LGA775 3.6GHz
P4 processor. (All
kits tested work with both Intel and
AMD CPU models.) We evaluated each
kit based on four criteria:

Cooling Performance
Once the kit was installed, we measured
the CPU temp at idle. We next ran CPU
Burn-in ( http://users.bigpond.net.au/cpuburn/ )
for one hour with a 100-percent load
on the CPU, and then measured its
temperature at the end of the cycle.
To measure the effectiveness of water
versus air cooling, we compared
the water-cooling temperatures with
those from a stock Intel heatsink/fan
(HSF) and, alternately, a high-velocity
aftermarket fan from Thermaltake—the
Jungle 512.

Ease of Installation
This is a necessarily subjective
measurement, but it’s a crucial one
nonetheless. We evaluated each kit
for the quality of the manufacturer’s
documentation, how well the kit’s parts
were labeled, and how easy it was to
install the sucka. This could very well
be the deciding factor when choosing
among kits that perform similarly.

Overclocking Potential
Because water cooling facilitates lower
overall temperatures than air cooling,
it’s often used by overclockers to reach
astronomic clock speeds. We tested
each kit for its overclocking ability and
scored some impressive results, but we
did encounter some limitations related
to the CPU and motherboard we used.
In other words, your mileage may vary.

Noise Level
One of the big advantages of water—
versus air—cooling is that it doesn’t
require as many notoriously noisy
fans. Still, most kits include a fan to
draw air over the radiator, and not all
of the kits include enough water blocks
for all the components you’ll want
to cool. Once we installed a kit, we
measured its sound output from
18 inches away using a decibel meter.

50 MA XIMUMPC MAY 2005


Swiftech H2O-120


The Swiftech H2O-120 kit stands out in this
roundup because of its modular design:
You can purchase a CPU-only confi guration
today, and then buy additional water
blocks and radiators to expand the system
according to your needs. The kit we tested
included water blocks for the CPU and GPU,
a 120mm radiator with a fan, a coolant
reservoir that mounts in a 5.25-inch bay,
and a 12V pump. The kit uses 1/2-i nch
tubing with “cool-sleeve” coils to prevent
the formation of kinks that would impede
water fl ow, and it uses green UV-reactive
(aka “glow-in-the-dark”) coolant that’s
intended to be mixed with distilled water.
One very nice touch is that the 120mm
fan/radiator assembly can be mounted on a
wide range of exhaust-fan mounts, so it will
work with just about any ATX enclosure.
The Swiftech unit is fabulous hardware
that performs very well, though installing
it is a bit of an adventure that requires
extreme levels of patience and diligence.
And the problem stems from the very
modularity we praise the kit for: Rather than
provide a step-by-step walk-through for the
particular kit you’ve purchased, the H2O-120
comes with installation instructions for the
entire catalog of Swiftech parts.
As much as we appreciate that Swiftech
provides a printed manual—not all the
vendors in this roundup do—we don’t
appreciate having to wade through 41
pages searching for the nuggets of hard
information we need to install something.
The Swiftech kit warns you—perhaps
a little too late—that the H2O-120 is “for
expert users,” and they’re not kidding.
Installation of the kit requires a lot of
patience and care, making the Swiftech kit
easy to install for those that know what
they’re doing, but intimidating to the
average Joe who is looking to install his
fi rst kit. In our opinion, a simple overhaul of
the manual would have made this a fi ne kit
for all parties.

TEST RESULTS
Cooling performance
The H20-120 performed quite well,
registering the second-lowest temperature
delta from idle to full load. Under full load
it was only hotter than the stock HSF at idle
by four degrees Centigrade.

Ease of installation
This kit wasn’t that hard to install, but the
41-page manual would be very intimidating
to water-cooling newcomers. All the
information you need is in there, but you
have to hunt to fi nd it.

Overclocking
This was the best overclocker in the bunch,
allowing us to ratchet up our 3.6GHz P4 all
the way to 4.12GHz. Not bad.

Noise output
Because the 120mm fan mounts on the
outside of the case, noise output increased
from the stock HSF by four decibels, but we
think that’s an acceptable trade-off for the
overclocking prowess.

MA XIMUMPC VERDICT (^8)
$260, http://www.swiftnets.com
The Swiftech VGA water block
prior to installation.
The internal reservoir is made from
molded plastic and conveniently
mounts in a 5.25-inch bay. To add
more water, you just pull it out,
unscrew the cap, and fill ‘er up.
The H20-120 uses “cool-sleeve” coils
throughout its interior and exterior to
prevent the tubing from kinking; kinked
tubes will impede water flow and
severely reduce cooling efficiency.

Free download pdf