Spire CoolGate SP450B0 Heatsink/Fan

Cooling
Viewing page 3 of 3 pages. Previous 1 2 3

ON THE TEST BENCH

With a new test bed comes a new benchmark cooler to compare to. From now on, I will be comparing new Socket 478 coolers to our unofficial standard, the Zalman CNPS7000A-Cu. The heat source will be an Intel Celeron 1.7 (Willamette core). Intel's TDP rates this chip's output at 63.5 Watts, while CPUHeat & CPUMSR calculates the Maximum Power output to be 79.8 Watts. Results will be provided based upon both methods of calculating heat output.


My new test bed: VIA P4PB 400 and Willamette at 1.7GHz.
Key Components in Socket 478 Test Bed
Intel Celeron 1.7 Willamette core -- TDP 63.5Watts, MP 79.8 Watts
VIA P4PB400 mainboard -- VIA P4X400 chipset, on-die CPU thermal diode monitoring
VGA GeForce2 MX400 PCI graphics adapter -- passively cooled (graciously provided by forum member, Trodas)
DDR SDRAM -- one stick, generic 256MB
Seagate ST3160023A -- 160GB PATA HDD, previously compared against Samsung SpinPoint in another article
Seasonic Super Silencer 300
Arctic Silver V thermal compound
CPUBurn load-induction software
Motherboard Monitor 5.3.7.0 system monitoring software
Extech True RMS MultiMeter 22-816 -- digital multimeter and thermal probe utilized for monitoring ambient temperature and accurate setting of
Zalman FanMate adjustable inline fan speed controllers with speed-reading passthrough feature (2; one set for 5 Volts, one set for 7 Volts)

TEST PROCEDURE

As with previous tests, all testing was performed open-air, rather than in an enclosure.

  • All heatsinks were cleaned prior to use.
  • Thermal interface material was allowed to properly set by cycling between completely off and full load several times for a minimum duration of 24 hours total.
  • All tests were run minimum of 25 minutes to allow stabilization of temperatures, with 20 minute cool down periods between test runs.
  • All tests were run three times to verify accuracy; any differing results were averaged before showing up here.
  • All tests were run with an ambient temperature of 24C. If the ambient climbed or dropped out of this range, testing was immediately stopped and did not start again until ambient returned to 24C.
  • Ambient temperature was measured at a position in front of and beneath the test bed several inches to ensure that test bed heat did not factor into the readings.
  • Acoustics tests were done with the fan all by itself in open air, as well as installed onto the heatsink, and both tests were done up close as well as 10' away on the other side of the room, at 5, 7 and 12 Volts.
  • All listening tests were conducted in the wee hours of July 26th, ensure lowest possible ambient noise level (I live in the suburbs on a major street).
  • Because the power supply of the test bed was audible several feet away, my quiet rig Sigma One was used for powering the fan during acoustic testing.
  • Once again, two separate FanMates were utilized, one set for 5V and one set for 7V, to maintain consistency of tests when switching from fan-alone to heatsink-and-fan acoustics testing.
  • Testing was also done to determine if the nature of the unit's acoustics changed at all with orientation (hub down, hub up or vertical) and it is confirmed that the Fanner Tech unit did not change acoustics with orientation.
  • Temperatures were read from the Willamette core's onboard diode, as reported by Motherboard Monitor. The thermistor on this test bed has been calibrated for accuracy using various methods; all numbers have had the necessary adjustments made.
  • Temp. Rise refers to the difference between the ambient temperature (24°C) and the CPU temperature.
  • °C/W refers to the rise in temperature (over ambient) for each watt of heat dissipated by the processor.

RESULTS

Because the cooler comes equipped only to run full speed, I utilized a pair of Zalman FanMate1 inline fan controllers to test the cooler at 5 Volts and 7 Volts. Using two separate controllers and keeping them preset to their respective voltages ensures consistency and accuracy between each of the three test runs. Once again, TDP was 63.5Watts, and MDP was 79.8 Watts

At 12 Volts, the fan is spinning at a tremendous 3800 rpm (compared to Zalman 7000AlCu 's 2600 rpm). Up close, it produced a high frequency wail signature sound, with very obvious wind whoosh, best described as a fffffffffffffffffffffff noise. When I moved to the other side of the room, the high frequency wail diminished quite a bit, and the air whoosh became the predominant effect, but the overall volume did not drop by much at this distance. Once installed onto the sink, the high frequency wail gets covered up by the intense wind noise, not unlike the sound of the fans set for high in your car's HVAC system. Moving way back, the sound signature remains the same, only lesser in volume.

Fan at 12V
Heatsink
Idle
Load
°C Rise
°C/W MP
°C/W TDP
CoolGate
31°C
44°C
20
0.25
0.31
7000AlCu
29°C
45°C
21
0.26
0.31
°C Rise refers to the rise in temperature over the ambient (in this case, 24°C) at full load.
At 7 Volts, the fan spins at a more reasonable 2500 rpm (1800 rpm for Z-7000AlCu). Listening to just the fan, the sound signature dropped to a somewhat lower pitched drone, with some mid-high frequency buzzing. From across the room, the drone noise dropped, but the buzzing did not drop as much; in effect, it appears to get louder ( the drone is no longer covering it up as much). Once the fan was installed on the heatsink, the drone remained, but it combined with air turbulence to drown out the buzz noise up close. From farther away, it's really the mid-high frequency drone that shows through most. Hard-mounting the fan seems to reduce the buzz effect somewhat.
Fan at 7V
Heatsink
Idle
Load
°C Rise
°C/W MP
°C/W TDP
CoolGate
32°C
47°C
23
0.29
0.36
7000AlCu
29°C
46°C
22
0.28
0.35
At 5 Volts, CoolGate's fan slowed to 1700 rpm (with the 7000AlCu now doing 1350 rpm). Sitting all alone on some foam and up close, the fan produces a mid-frequency drone, with a buzz that sounds similar to the sound of a ratchet; moving back 10 feet makes the buzz more pronounced than the drone. Mounted to the heatsink, there isn't much of an increase in air turbulence noise, but the buzz diminishes quite a bit, leaving a pure drone with just the slightest clicking at short range. Moving to longer range, the only sound is minor droning and nothing else, but it is still quite audible.
Fan at 5V
Heatsink
Idle
Load
°C Rise
°C/W MP
°C/W TDP
CoolGate
33°C
54°C
30
0.38
0.47
7000AlCu
30°C
48°C
24
0.30
0.38

ANALYSIS and CONCLUSIONS

Spire's CoolGate is easily capable of cooling any chip that our benchmark Zalman CNPS7000A-AlCu can tackle at 12 Volts. Acoustics, unfortunately, just isn't very good with the stock fan. Because a 70mm fan is used, installing a larger, quieter fan, such as an 80mm Panaflo L1A, is not a simple swap. The CoolGate is very affordable, however, and with a bit of modification may serve its owner well.

The vertical, lower-air-impedance design of the heatsink allows a clean, straight-through air path. Paired with a motherboard whose CPU socket is oriented in the right way, it allows a crafty user to design a duct directly from the cooler to the case's rear exhaust fan. With enough airflow from the exhaust case fan, this arrangement may allow the user to remove the stock HS fan all together. Such a setup has the advantage of keeping the warm air from the CPU from floating into the PSU, and causing the PSU fan to ramp up from increased thermals.

With the low price, some people may consider this bit of work worthwhile. Many of these possibilities aren't options on conventional HS designs that blow air downward. The heated air from conventional HS can only spread throughout the case, with some going out the case exhaust and some into the PSU. The main challenge is to find a suitable quiet 70mm fan.

A couple recommendations for Spire would be to utilize a copper base, and also to mount the fan with a 15 to 25mm deep shroud, to distance it from the heatsink fins and reduce the dead zone as well as air turbulence noise. In the CoolGate, it is the proximity of the fan (right against the high-fin-density sink) that causes the majority of the air turbulence noise.

The Good

Hey, it's cheap. $25 is a heck of a lot less than a Hyper 6, SP-94 or even a Zalman 7000A-AlCu.

With the right mainboard, this cooler has lots of fun modding potential, including direct-to-exhaust ducting.

The light weight and compact design is very safe and highly compatible with a greater majority of installations, even in many tighter microATX cases that need to travel a great deal.

Tool-free mounting system is a joy to use.

The Bad

A great deal more difficult to use a non-stock fan than some other coolers, a few of which are similar in price.

A fan speed reduction device would been nice to include, like its same-priced brother, the Spire CoolWave.

The cooler, while capable, seems to rely on high airflow for its performance, most likely due to a combination of high fin density and the all-aluminum construction. The dual heatpipes can only do so much to help.

The Ugly

The stock, 70x70x15mm fan needs serious help. Were it easier to install an alternate fan, this wouldn't be such a big deal, but that's simply not the case here, and the sound quality of the stock fan is unacceptable.

Many thanks to Spire for providing this CoolGate sample.

* * *

Discuss this review in the SPCR Forum.



Previous 1 2 3

Cooling - Article Index
Help support this site, buy from one of our affiliate retailers!
Search: