The fins of a 9700 tower heatsink get split into two sections to allow a 120mm fan to fit between them. The 9900 looks like the Zalman’s last kick at the 9000 series.
Postcript: Zalman CNPS9900 LED Unshrouded added Jan 9, 2009 (page 8)
Jan 2, 2009 by Lawrence Lee
| Product | Zalman CNPS9900 LED LGA775/K8 CPU Cooler |
| Manufacturer | Zalman Tech Co. |
| Street Price | ?? Probably US$70~90 |
POSTSCRIPT: Zalman CNPS9900 LED Unshrouded added Jan 9, 2009 (page
8)
Zalman has been sort of coasting since the then-unique and effective CNPS7000
CPU cooler. They made small alterations at first, making a
larger version with a bigger fan, then moved to translucent fans, to scatter
LED lighting. It wasn’t until the 9500
came along that a more interesting change came. The introduction of
heatpipes and the 90-degree rotation of the orientation into a sleek,
sexy, funnel-like tower heatsink breathed life into the CNPS name. From there
they again made minor alternations, increasing the fan size with the 9700
and then strangely decreasing it in the case of the 9300AT.
|
|
The CNPS9900 LED, which might be the last gasp of the 9000 series, takes a more
drastic approach. The cooler has been split down the middle and spread apart
to make room for a fan integrated into the center of the heatsink. It’s not a
new concept, having been implemented by notable manufactuers such as Scythe,
Thermaltake and Thermalright, but this is Zalman’s first attempt. The 9000 series
has been showing its age for quite sometime, not being able to compete with
today’s top heatsinks. Can the 9900 bring glory back to Zalman?
|
|
| Zalman CNPS9900 LED: Key Features (from the product web page) | |
| Feature & Brief | Our Comment |
| Does not generate noise or vibration in Silent Mode. | Zalman fans have never really lived up to this claim, though the absence of competition in the earlier years made this claim more plausible than it is today. |
| 100% copper heatsink with aerodynamically optimized “tunnel” design for maximum cooling efficiency. | OK. |
| Patented heatpipe design for cooling performance of up to 6 heatpipes with the use of just 3. | OK |
| Ultra-thin 0.2mm fins for minimized weight and significantly reduced airflow resistance. | For the record, the total fins area is just about the same as the CNPS 9700. |
| Ultra quiet 120mm PWM LED fan for automatic fan speed control according to the CPU’s temperature. | Let’s see if the new PWM fan helps Zalman achieve the first claim above. |
| Zalman CNPS9900 LED: Specifications (from the product web page) |
 |
PHYSICAL DETAILS
The CNPS9900 is in essence a CNPS9700 split into two sections.
It is Zalman’s first CPU cooler with a fan integrated into the center of the
body.
|
|
|
|
|
|
|
|
|
|
SHROUD/FAN REMOVAL
Zalman coolers are usually prime candidates for fan replacement.
They typically ship with loud ball-bearing fans that spin too fast. Getting
at the fan is fairly easy.
|
|
|
|
|
|
|
|
INSTALLATION
The most critical aspect of installation is for
the heatsink to be securely mounted. The more firmly it is installed, the
better the contact between the heatsink’s base and the CPU itself. It’s also
less likely to fall off. Ease of installation is also important — a simple
mounting scheme means less time spent installing, and a reduced likelihood
of screwing up.
|
|
|
|
|
|
|
|
TESTING
Before thermal testing, we take some basic physical measurements.
| Zalman CNPS9900 LED: Approximate Physical Measurements | |
| Weight | 790 g (heatsink alone) 820 g (including LGA775 mounting frame) |
| Fin thickness | 0.24 mm |
| Fin spacing | Varies |
| Vertical Clearance | 55 mm |
| Horizontal Overhang | 9 mm (measured from the edge of the heatsink to the top edge of our test motherboard’s PCB) |
| Comparison: Approximate Fin Thickness & Spacing | ||
| Heatsink | Fin Thickness | Fin Spacing |
| Zalman CNPS9300 AT | 0.17 mm | Varies |
| Zalman CNPS9900 LED | 0.24 mm | Varies |
| Scythe Ninja | 0.31 mm | 3.95 mm |
| Scythe Ninja 2 | 0.39 mm | 3.68 mm |
| Thermalright HR-01 Plus | 0.45 mm | 3.15 mm |
| Noctua NH-U12P | 0.44 mm | 2.63 mm |
| Noctua NH-C12P | 0.47 mm | 2.54 mm |
| Xigmatek HDT-S1283 | 0.33 mm | 1.96 mm |
| Zerotherm Zen FZ120 | 0.37 mm | 1.80 mm |
| Thermaltake SpinQ | 0.31 mm | 1.67 mm |
| Thermalright Ultra-120 | 0.45 mm | 1.42 mm |
Testing was done according to our
unique heatsink testing methodology, and the reference fan was profiled
using our standard fan testing
methodology. A quick summary of the components, tools, and procedures
follows below.
Key Components in Heatsink Test Platform:
- Intel
Pentium D 950 Presler core. TDP of 130W; under our test load, it measures
78W including efficiency losses in the VRMs. - ASUS
P5LD2-VM motherboard. A basic microATX board with integrated graphics
and plenty of room around the CPU socket. - Samsung MP0402H
40GB 2.5″ notebook drive - 1
GB stick of Corsair XMS2 DDR2 memory. - FSP Zen 300W
fanless power supply. - Arctic Silver
Lumière: Special fast-curing thermal interface material, designed
specifically for test labs. - Nexus 120 fan (part of our standard testing methodology; used when
possible with heatsinks that fit 120x25mm fans)
| Nexus 120 fan measurements | ||
| Voltage | Noise | RPM |
| 12V | 16 dBA@1m | 1100 RPM |
| 9V | 13 dBA@1m | 890 RPM |
| 7V | 12 dBA@1m | 720 RPM |
| 5V | 11 dBA@1m | 530 RPM |
Measurement and Analysis Tools
- Seasonic
Power Angel for measuring AC power at the wall to ensure that the
heat output remains consistent. - Custom-built, four-channel variable DC power supply, used to regulate
the fan speed during the test. - PC-based spectrum analyzer:
SpectraPlus with ACO Pacific mic and M-Audio digital
audio interfaces. - Anechoic chamber
with ambient level of 11 dBA or lower - Various other tools for testing fans, as documented in our
standard fan testing methodology. - SpeedFan
4.32, used to monitor the on-chip thermal sensor. This sensor is not
calibrated, so results are not universally applicable. - CPUBurn
P6, used to stress the CPU heavily, generating more heat than most
real applications. Two instances are used to ensure that both cores are stressed. - Throttlewatch
2.01, used to monitor the throttling feature of the CPU to determine
when overheating occurs.
Load testing was accomplished using CPUBurn to stress the processor, and the
graph function in SpeedFan was used to make sure that the load temperature was
stable for at least ten minutes. The stock fan was tested at various voltages
to represent a good cross-section of its airflow and noise performance.
TEST RESULTS
The tests were performed with a voltage controller, which is our standard procedure. We’ll grant that a PWM controller might be able to run the fan slower and quieter, but at the expense of cooling performance.
| Zalman CNPS9900 LED | ||||
|---|---|---|---|---|
| Fan Voltage | SPL @1m | Temp | °C Rise | °C/W |
| 12V | 39 dBA | 32°C | 11 | 0.27 |
| 9V | 32 dBA | 34°C | 13 | 0.28 |
| 7V | 24 dBA | 36°C | 15 | 0.31 |
| 5V | 16 dBA | 43°C | 22 | 0.35 |
| Load Temp: CPUBurn for ~10 mins. °C Rise: Temperature rise above ambient (21°C) at load. °C/W: based on the amount of heat dissipated by the CPU (measured 78W); lower is better. | ||||
Fan @ 12V: Thermal performance was excellent at 11°C above ambient —
the best result we’ve ever recorded. Of course a Zalman CPU fan at 12V is a
sonic force to be reckoned, creating a symphony of turbulence and whine that sounds
like a swarm of tiny, very threatened bees.
Fan @ 9V: The CPU temperature increased by only 2°C. However, the fan
droned and was very breezy with plenty of turbulence. At 32 dBA it was still
unacceptably loud.
Fan @ 7V: Performance suffered by 2°C once again. Our ears detected a
bit of hum and whine, but the bulk of the noise was the result of turbulence.
Though the SPL was 24 dBA, it actually didn’t sound half bad. We would have
prefered this to be the fan’s maximum speed.
Fan @ 5V: The temperature skyrocked by an additional 7°C — the CNPS9900
suffers dramatically when the airflow is significantly cut. The noise level
however was excellent at only 16 dBA. The fan exhibited a slight low-pitched
hum, but other than that it was fairly smooth.
The cooler also passed some vibrations through the board to our testing platform.
The amount of vibration varied depended on how much of the backplate made contact with the
platform surface. Properly installed in a case there should be enough clearance
to make this a non-issue.
| Comparables | |||||
| Zalman CNPS9900 LED | Zalman CNPS9300 AT | Thermaltake SpinQ | |||
| SPL @1m | °C Rise | SPL @1m | °C Rise | SPL @1m | °C Rise |
| 32 dBA | 13 | 30 dBA | 17 | 30 dBA | 21 |
| 24 dBA | 15 | 25 dBA | 18 | 27 dBA | 22 |
| 21 dBA | 23 | 21 dBA | 24 | ||
| 16 dBA | 22 | 16 dBA | 27 | ||
Compared to the Zalman CNPS9300
AT, the CNPS9900 is a significant improvement, posting better numbers at
more or less equivalent noise levels. At 5V, the CNPS9900 matches the CNPS9300’s
performance but at a much better SPL — 16 dBA vs. 21 dBA. It is also superior
to the Thermaltake SpinQ.
| CNPS9900 vs. Comparables @ 16 dBA | |
| Heatsink | °C Rise |
| Thermalright SI-128 | 21 |
| TT Big Typhoon VX | 21 |
| Zalman CNPS9900 | 22 |
| TT Big Typhoon | 24 |
| TT SpinQ | 27 |
Unfortunately as a quiet cooler, the Zalman CNPS9900 is poor. When the
fan is undervolted to 5V it generates 16 dBA, the same noise level as our reference
Nexus 120mm at 12V. Of all the coolers we’ve tested with this fan in the past
two years, the CNPS9900 manages to only beat one: the Thermaltake
Big Typhoon.
| CNPS9900 vs. Top Coolers @ 16 dBA | |
| Heatsink | °C Rise |
| Thermalright U120E | 12 |
| Thermalright HR-01+ | 13 |
| Xigmatek HDT-S1283 | 13 |
| Noctua NH-U12P | 14 |
| Zalman CNPS9900 | 22 |
Compared to the best heatsinks we’ve tested the CNPS9900 is inferior by about
8-10°C. That’s a very large gap.
MP3 SOUND RECORDINGS
These recordings were made with a high
resolution, lab quality, digital recording system inside SPCR’s
own 11 dBA ambient anechoic chamber, then converted to LAME 128kbps
encoded MP3s. We’ve listened long and hard to ensure there is no audible degradation
from the original WAV files to these MP3s. They represent a quick snapshot of
what we heard during the review.
These recordings are intended to give you an idea of how the product sounds
in actual use — one meter is a reasonable typical distance between a computer
or computer component and your ear. The recording contains stretches of ambient
noise that you can use to judge the relative loudness of the subject. Be aware
that very quiet subjects may not be audible — if we couldn’t hear it from
one meter, chances are we couldn’t record it either!
The recording starts with 10 second segments of room ambience, then the fan
at various levels. For the most realistic results, set the volume so that
the starting ambient level is just barely audible, then don’t change the volume
setting again.
- Zalman
CNPS9900 LED at one meter
— 5V (16 dBA@1m)
— 7V (24 dBA@1m)
— 9V (32 dBA@1m)
— 12V (39 dBA@1m)
Reference Comparatives
- Thermaltake
SpinQ at one meter
— 5V (16 dBA@1m)
— 7V (21 dBA@1m)
— 9V (27 dBA@1m)
— 12V (30 dBA@1m)
- Zalman
CNPS9300 AT at one meter
— 5V on foam (18 dBA@1m)
— 5V (21 dBA@1m)
— 7V (25 dBA@1m)
— 9V (30 dBA@1m)
— 12V (37 dBA@1m)
- Nexus
120mm Real Silent Case fan at one meter
— 5V (11 dBA@1m)
— 7V (12 dBA@1m)
— 9V (13 dBA@1m)
— 12V (16 dBA@1m)
FINAL THOUGHTS
The CNPS9900 LED is Zalman’s best performing heatsink to date. It is the most
successful of the many adaptations of the original flower
heatsink design, but it’s still a long way from the best heatsinks unless the fan is cranked all the way up. It’s a big, lumbering
giant, that pound for pound and dollar for dollar, just does not keep up with the competition.
The CNPS9900 fan, like others from Zalmans, is poor acoustically and spins far too fast for
our liking. The placement of the fan in the center of the cooler is dubious
— we’ve never found such designs to be very good. The large gap between the
fan and the rest of the heatsink only exacerbates problem. The plastic shroud also ruins the CNPS9900’s aesthetics
— it looks like it’s going for a frivolous whiplash lawsuit.
Continuing to recycle the same heatsink with various tweaks and tricks can
only garner so much improvement. And while Zalman continues to modify the
CNPS9000 series, heatsinks from Thermalright, Noctua, and Xigmatek are handing
them their lunch. Yes, Zalman, your CNPS coolers were revolutionary — classic
products on the cutting edge of silent computing. For that, we thank you from the bottom
of our silent hearts. But the way you’ve dragged the design through time with
various disguises makes us think of Weekend At Bernie’s. Time to
move on.
| Zalman CNPS9900 LED | |
| PROS * Secure mounting systems | CONS * Fan too loud |
Our thanks to Zalman
Tech Co.
for the CNPS9900 LED heatsink sample.
* * *
POSTSCRIPT, Jan 9, 2009: Zalman CNPS9900 LED, Unshrouded (overleaf,
on page 8)
Articles of Related Interest
ThermalTake SpinQ: Unique Blower-fan Heatsink
Thermolab
Baram Blows In
Scythe
Ninja 2: Tweaking a Classic
Zalman
9300AT: Not me too, but me again
Noctua
NH-U12P Tower Cooler
Zerotherm
Zen FZ120 CPU Cooler
* * *
Discuss
this article in the SPCR forums.
POSTSCRIPT: Zalman CNPS9900 LED Unshrouded
Jan 9, 2009 by Lawrence Lee
After some discussion with Zalman, we found it prudent
to test the 9900 LED without the offending shroud to see what effect it would
have on performance and acoustics. With the fan exposed on the sides, the fan
has better access to fresh air and there is more space for heat to dissipate.
There was also the possibility that the shroud increased vibration.
|
|
| Zalman CNPS9900 LED | ||||
|---|---|---|---|---|
| Fan Voltage | With Shroud | Without Shroud | ||
| SPL @1m | °C Rise | SPL @1m | °C Rise | |
| 12V | 39 dBA | 11 | 34 dBA | 9 |
| 9V | 32 dBA | 13 | 29 dBA | 10 |
| 7V | 24 dBA | 15 | 23 dBA | 12 |
| 5V | 16 dBA | 22 | 17 dBA | 16 |
| Load Temp: CPUBurn for ~10 mins. °C Rise: Temperature rise above ambient (21°C) at load. °C/W: based on the amount of heat dissipated by the CPU (measured 78W); lower is better. | ||||
As it turns out, the unshrouded heatsink performed significantly better. Between
12V and 7V there was a 2-3°C improvement, but more importantly the CPU was
6°C cooler at 5V. With the fan puttering at low speed, it was difficult
to exhaust hot air through the fins of the front side of the cooler. Removing
the head-band allowed heat to escape in all directions.
Acoustics were also significantly improved at 12V and 9V. The character of
the fan was unchanged but without the plastic encapsulating it, it sounded less
‘hollow’ — an effect that was amplified at higher fan speeds. There was
barely a measurable SPL difference at 7V and 5V. The shroud was actually fitted
very snuggly all around so there wasn’t any noticeable change in the amount
of vibration.
| CNPS9900 vs. Top Coolers @ 16 dBA | |
| Heatsink | °C Rise |
| Thermalright U120E | 12 |
| Thermalright HR-01+ | 13 |
| Xigmatek HDT-S1283 | 13 |
| Noctua NH-U12P | 14 |
| Zalman CNPS9900 (Unshrouded, 17 dBA) | 16 |
| Zalman CNPS9900 | 22 |
The modified CNPS9900 at 5V perfomed much better, coming within a few degrees
of the heatsinks we consider to be elite (at approximately the same noise level).
We’re glad to see the CNPS9900 improve so much with a simple modification,
but we believe a few more tweaks could make it an even better cooler. Without
the shroud, it is a possible to use a bigger and therefore more efficient fan,
preferably one that sounds smoother than Zalman typically uses. We also recommend
narrowing the gap between the fan and the two sides of the heatsink. This can
be done by making the fins larger, bending the heatpipes so they’re closer together,
or by simply utilizing a thicker fan.
With these new findings there may be some life left in the nines yet. Perhaps
we’ll see a CNPS9950 in the near future?
