3DVelocity would like to
thank Bacata,
France and especially Jean-Christophe Agobert for their
help and courtesy in providing this HSF for review.
Continued :
Time to take a closer
look at how these puppies are constructed. Although the base
didn't give much away, removing the fan and taking a look
at the top reveals all.
The 60 fins are created
from 30 individual sheets of copper which are clamped at the
centre between those two Aluminium blocks. To create the extra
bulk at the centre small sheets are inserted between the main
ones.

Here's a closer look.
You can see how even the ends of the smaller Copper "spacer"
sheets have had the ends turned to help keep the fins evenly
spaced. I sincerely hope this is an automated process coz
I for one wouldn't fancy putting these together all day.

Even with the copper core
removed from its plastic housing the fins feel amazingly solid
and although the thing looks pretty flimsy in the pics it
would need some deliberate force to bend things out of shape.


Installation :
Fitting the 5000+ is an
absolute breeze, or a complete nightmare depending on how
you view it. For the P4 owner (both of you :o), things are
just about as easy as it comes. Either your case will have
threaded holes or captive nuts behind the holes on the motherboard
into which you screw, or your motherboard will have been supplied
with some form of adapter or bracket to sit at the back of
your board which will serve the same purpose. You will though
need to swith to the longer screws supplied with the 5000+
and not use the ones it has in when you open the box.
The AMD user may also
be lucky and have threaded holes behind the holes in the motherboard,
but it's far more likely you'll have to resort to the slightly
more labour intensive method we're about to cover.
The first step (take a
deep breath) is to remove your motherboard. Okay this sounds
a bit drastic but you should remember not everybody swaps
and changes HSF's the way we reviewers, tweakers and hardware
junkies do, so for most people this will be a once only procedure.
What seems a shame is that the computer newbie will be out
of the frame unless he/she can persuade a friend or pay a
pro to do it.
Once the board is removed,
you then need to fit the four threaded posts onto the board
through the four holes situated around your socket. I've called
them "threaded posts" unlike Zalman who refer to
them as "nipples" not because I get the giggles
every time I type the word "nipple", but because
they just don't strike me as being "nipples", therefor
from now on they're posts. Good, glad that's sorted out!
In one of the bags you'll
find four screws and a variety of plastic washers. The size
of washer you use depends on the size of hole in your board,
but either way you fit one washer either side of the hole
and then screw the post securely down into place.

Once all fours posts are
in place, you're ready slap your board back into the case
and fit the 5000+.
All that's required here
is to sit the cooler onto the four posts and screw it down.
The four screws in the actual Zalman are fitted with springs
which apply a downward force as you tighten them. You are
best tightening them a little at a time rather than tightening
each one fully as this will even out the pressure on the die.
The great thing is that you know there's no way you're going
to chip a corner off the slug and end up on a mercy dash to
the nearest PC store to replace it (been there? I know I have!)

Getting the correct pressure
is also a doddle, you simply keep turning until the head of
the screw is flush with the plastic casing. Couldn't be simpler!

Repeat for all four screws
and there you have it, one Zalman CNPS5000+ all installed
and waiting to go.
Now before somebody with
too much time on their hands takes a look at the picture below
and decides there's no CPU fitted, you're right! I used a
spacer to replace the CPU so I could photograph this series
of pictures on an Abit KT7 RAID I have in the workshop. I
didn't want to completely remove the motherboard from the
test system just for the photos so I did it this way instead.
When it comes to supplying
the unit with some juice, you have to decide which (if any)
of the two resistors you're going to fit. I'm sure I don't
need to explain that the greater the resistance the slower
the fan and the less efficient the cooling. Fan speeds vary
from 5600 RPM at full tilt to 3200 RPM in low noise mode (90mm
resistor/lead) and 2000 RPM in silent mode (60mm resistor/lead).
You simply need to decide if your priority is performance
or low operating noise. One concern is that I measured somewhere
around 4.5 volts with the larger 100 OHM resistor in line,
yet Zalman's own specs suggest the 60mm fan they fit is rated
as having a 5 volt start requirement. For this reason alone
I think I'd tend to stick with the 55 ohm (90mm) lead.

Testing :
System Spec -
DFI AK75EC Motherboard
AMD Athlon Tbird 1GHz @ 1GHz
256MB PC133 CAS2
Radeon 64MB DDR VIVO
All temperatures taken using on board thermistor.
Idle temps are just that,
30 minutes sat doing sweet nothing with all background tasks
closed, while load temps are taken after 30 minutes of looped
3DMark2001 running at the same time as Prime95. Averages are
taken of 3 runs.
To even out the playing
field a little, I tested the Thermosonic ThermoEngine both
with its stock low profile fan, and with the Zalman fan installed.
Arctic Silver (original)
thermal grease was used in all tests.
| HSF |
Ambient |
CPU Idle |
CPU Load |
| Zalman
CNPS5000+ Normal Mode |
23 |
31 |
41 |
| Zalman
CNPS5000+ Quiet Mode |
23 |
31 |
46 |
| Zalman
CNPS5000+ Silent Mode |
23 |
32 |
51 |
| Thermosonic
ThermoEngine, stock fan |
23 |
31 |
45 |
| Thermosonic
ThermoEngine,Zalman fan |
23 |
30 |
41 |
| Thermaltake
Super Orb |
23 |
34 |
47 |
Conclusion :
What can I tell ya? I
was pleasantly surprised to see the 5000+ keep pace with the
ThermoEngine under load when they used the same fan.
The noise levels with
the fan unthrottled are far from quiet, but they're not really
obtrusive and they're still better than most. Fitting the
"Quiet Mode" resistor offered significantly better
noise levels but at the expense of cooling power. This isn't
really an option for anyone attempting an aggressive overclock,
but for normal use it will be a nice compromise. The "Silent
Mode" resistor delivered pretty much that. The sound
levels were barely perceptible, but cooling performance became
dire and for the reasons I mentioned earlier about the low
voltage I think I'd avoid this one.
The CNPS5000+ is an impressive
progression in Zalman's cooling technology and while the $49.00
price tag isn't exactly budget, this is a classy piece of
hardware that offers flexible performance and solid build
quality in an unusual and innovative way. The method of mounting
makes this ideal for anyone who has damaged the lugs on their
socket, who owns a P4 or who wants the reassurance of having
a good looking piece of Copper staring back at them every
time they open their case (not that you can see that much
of it when it's fitted).
Zalman's design means
the CNPS5000+ offers an incredible 1060 square centimetres
of surface area, which kind of makes you wonder why it didn't
well and truly trounce the ThermoEngine, and I'm beginning
to wonder if the laws of physics have brought us close to
the "brick wall" of conventional air cooling technology.
I have read that reversing
the fan on this unit offers a significant improvement in cooling,
so I tried switching from its stock exhaust position (air
is drawn in at the base of the heat sink and exhausted out
of the top) to the intake (air is drawn in at the top and
blown downwards into the heat sink) position, and in my particular
setup the readings were all but identical. I actually expected
them to be slightly higher because the plastic shroud, normally
only used with the fan in the exhaust position, can't be removed.
All in all this is an
excellent cooler that I'm sure will find a home in many a
system. It's not as whacky as their 2000 and 3000 series coolers,
it's not as light as the ThermoEngine, it's not as cheap as
the Themaltake Volcano II and it's not as heavy as a Hedgehog,
but for anyone wanting an impressive looking HSF with a solid
performance it's right up there.
I must again thank Bacata
for supplying this unit for test, and for those who don't
know who Bacata are, why not pop over and take a look.
