CoolerMaster Vortex Dream Heatsink/Fan

Cooling
Viewing page 3 of 4 pages. Previous 1 2 3 4 Next
TEST PLATFORM & PROCEDURE

The test bed has changed slightly since my last review, with the power supply modified to use an 80mm Panaflo L1A for extremely low noise, and the 160GB Seagate 7200.7 replaced with a 40GB 7200.7. These changes reduce the test bed's noise level so I can perform listening tests more easily. I have also moved the test bed to a location with more consistent ambient temperatures.


Further evolved test bed: PSU and HDD are both quieter,
and stability of ambient temperatures has also improved.

Key Components in Socket 478 Test Bed

All testing was performed with the equipment in open-air rather than in an enclosure.

Intel Celeron 1.7 Willamette core -- Intel's official Thermal Design Power (TDP) spec is 63.5 Watts; the more realistic Maximum Power (MP) as calculated by CPUHeat & CPUMSR, is 79.8Watts.
VIA P4PB400 mainboard -- VIA P4X400 chipset, on-die CPU thermal diode monitoring
eVGA GeForce2 MX400 PCI graphics adapter -- passively cooled (graciously provided by forum member, Trodas)
Generic DDR SDRAM -- one stick, 256MB
Seagate ST340014A -- 7200rpm, 3.5", 40GB PATA HDD
SeaSonic Super Silencer 300 -- stock fan swapped for 80mm Panaflo L1A
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 5V, one set for 7V)

All heatsinks were cleaned and 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; they were run through three times to verify accuracy; any differing results were averaged before showing up here.

All testing was conducted with an ambient temperature of 23C. If the ambient climbed or dropped out of this range, testing was immediately stopped and did not start again until ambient returned to 23C. 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.

  • Diode: The temperature of the Willamette core onboard diode, as reported by Motherboard Monitor. The temperature readings from the testbed have been calibrated to remove motherboard-induced reporting errors
  • Temp. Rise refers to the difference between the ambient temperature (23°C) and the temperature as reported.
  • °C/W refers to the rise in temperature (over ambient) for each watt of heat dissipated by the processor. The CPU power figures used for the calculations:
    • TDP (Thermal Design Power) = 63.5W
    • MP (Maximum Power) = 79.8W
  • Noise: Reported in dBA @ 1 meter as measured by Mike Chin on his heatsink/fan samples.

TEST RESULTS

Curiously, the voltage provided to the fan by its speed controller could not be accurately measured. There were two reasons:

1) The proximity of my hand near the fan wires or the rheostat itself causes the fan voltage to drop. The slowdown could be heard as a change in the pitch of the fan's noise and seen as a reduction in monitored RPM.

2) Applying the voltmeter's probe on the + contact also caused a drop in voltage.

So, the tests were performed with the stock speed control set at full speed in conjunction with a Zalman FanMate. (Two fanmates were used, actually: One set for 5V and the second set for 7V.) Also included are the results with a Panaflo 80L1A. Results for the similarly priced Spire CoolGate and the performance benchmark Zalman 7000A-Cu are included for comparison.

Test 1: Stock Fan at Various Voltages

Stock Fan w/Fanmate1 Controller
Fan Speed
Idle
Load
°C Rise
°C/W MP
°C/W TDP
Noise
(dBA/1m)
12V: 4600 RPM
29°C
44°C
21
0.26
0.33
48
7V: 3200 RPM
29°C
45°C
22
0.28
0.35
43
5V: 2600 RPM
29°C
46°C
23
0.29
0.36
34

At full power, the Vortex Dream's fan is blazing away at 4600 RPM. The fan is extremely noisy, measured at 48 dBA/1m on Mike Chin's sample. He says it is similar to the noise produced by the Scythe FC-50 stock fan at 12V. The performance is very good but the noise is totally unacceptable.

Interestingly, as the fan speed is slowed all the way down to 5V (to a speed just 56% of max), the cooling performance worsens by only 2°C while the noise drops by 14 dBA, or about 1.5 times quieter. All that huffing and puffing really doesn't amount to much: It's clear that with any given heatsink design, there is a certain optimum airflow level for high cooling-to-noise performance ratio. Beyond that point, additional airflow provides only marginal gains at the cost of major noise.

5V Comparisons

Since the noise performance is unacceptable at all but the 5V level, let's see how it compares at that voltage against some other heatsinks we've reviewed:

Comparison with Stock Fans at 5V
Heatsink
RPM
Load
°C Rise
°C/W MP
°C/W TDP
Noise
(dBA/1m)
Vortex Dream
2600
46°C
23
0.29
0.36
34
Spire CoolGate
1700
54°C
30
0.38
0.47
-
Zalman 7000A-Cu
1350
48°C
24
0.30
0.38
23

The fact that the Vortex Dream can nose past the excellent Zalman 7000 seems quite a feat... until you consider the noise. The Vortex Dream fan is spinning a lot faster and making more than twice the noise. A few SPCR members might be OK with this level of noise, and there are probably lots of overclockers who would think this inaudible, but for most people interested in a quiet PC, it is still too noisy.

With Built-in Control at Minimum

It turns out that simply turning the attached speed control to min (with it connected directly to the fan header on the motherboard) gives a voltage lower than the 5V min of the Fanmate, and it makes the fan spin slower and quieter: Mike measured 29 dBA/1m at this setting which is at least below the 30 dBA/1m level we define as quiet. (Editor's Note: It's still 12 dBA higher -- more than twice the noise -- than the 16 dBA claimed by CoolerMaster; did they put the sound level meter on the next floor?)

Then, in order to see how much quieter the fan could go, the built-in speed control was turned down while plugged into the Fanmate set to min. After some experimentation, it was determined that 1600 RPM was about the lowest speed at which a safe cooling performance could be obtained. The two results are tabulated below, along with a comparison against the Zalman 7000Cu at 5V.

Stock Fan at or lower than 2600 RPM (5V)
Heatsink
RPM
Load
°C Rise
°C/W MP
°C/W TDP
Noise
(dBA/1m)
VD - Fanmate1 at Min; Built-in Control at max (5V)
2600
46°C
23
0.29
0.36
34
VD - Built-in Control at Min
2350
48°C
23
0.31
0.39
29
VD - Built-in Control at 1 o'clock; Fanmate1 at Min
1600
56°C
33
0.41
0.52
22
Zalman 7000A-Cu at 5V
1350
48°C
24
0.30
0.38
23

As you can see in the table above, at roughly the same low noise level of 22-23 dBA/1m, the Vortex Dream is a hot 8°C behind the Zalman. The 33°C temp rise is at the limit of safe cooling for the CPU. The noise of the fan at this speed has some buzzing, humming and chuffing, but it's all at a low enough level that none of it is really consequential. It will be masked by other component noises or damped by the case.



Previous 1 2 3 4 Next

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