Scythe Samurai SCSM-1000 CPU Cooler

Cooling
Viewing page 3 of 4 pages. Previous 1 2 3 4 Next

ON THE TEST BENCH

I compared the Samurai to my sort-of-reference Zalman 7000AlCu heatsink, as well as the stock Intel P4 heatsink. I used the same processor that I've been using in all my heatsink tests, a Pentium 4, 2.4C. Yes, it may not put out as much heat as a new Prescott, but using this processor will allow me to directly compare the performance of the new Samurai against all the other high performance P4 heatsinks that I've tested during the past year.

I also tested the Samurai with the SPCR Reference 80mm Panaflo L1A. This will let us directly compare the cooling ability of this heatsink against other heatsinks that we have tested with the Reference fan, as per the basic SPCR heatsink testing philosophy of keeping everything the same (including the fan) and changing on the heatsink. The Samurai fan and the Panaflo L1A were both tested blowing down on the HS as well as blowing up.


Ye Olde Test Setup

Key Components in P4 HS Test platform:

The P4 HS test platform is an open system not enclosed in a case.

Intel P4-2.4C Northwood core - Maximum power is 66.2W.
Intel 875PBZ motherboard - Intel 875P Chipset; on-die CPU thermal diode monitoring
ATI Radeon 7500 passively cooled video card (AGP)
Mushkin PC3200 Level II - 2 x 256MB DDRAM
Seagate 80GB Barracuda IV hard drive
Seasonic SuperSilencer 400W (rev A1) PSU
Arctic Silver Ceramique Thermal Compound
Two-level metal platform with rubber damping feet. Motherboard on top; other components below.
CPUBurn processor stress software
Motherboard Monitor 5.3.4.0 software to track CPU temperature and fan speed

Each heatsink was cleaned and installed on the test system as per the manufacturer's and Arctic Silver's instructions. Prime95 was then run for 8 hours to verify system stability and cure the Ceramique. The system was then shut down and not restarted until the next morning when the actual testing was done. The system was allowed to cool between tests for 30 minutes. Each test was run for 30 minutes even though all temperatures generally stabilized within 15 to 20 minutes.

Each heatsink was retested three times on consecutive mornings to check to the consistency of the results. All results were within 1-2°C of each other and the average readings are included in the charts.

Ambient temperature was measured at 71-72°F (22°C) over the entire series of tests. No tests were run unless the ambient temperature was at that reference level.

* All temperatures in degrees Celsius.
* Diode: Reading from P4-2.4C CPU diode via Motherboard Monitor.
* Temp Rise refers to the difference between ambient temperature and the diode reading. .
* °C/W refers to the °C of temperature rise per watt of heat dissipated by the CPU.

RESULTS

The included Samurai fan controller is adjustable from a low of 1500 rpm to a high of 3350 rpm. I tested it at both extremes as well as at the middle of its range at 2475 rpm.

Full Speed Fan Results
Heatsink
airflow direction
idle
load
°C rise
°C/W MP
°C/W TDP
Samurai - stock fan
up
26°C
40°C
18
0.26
0.27
down
27°C
42°C
20
0.27
0.30
Samurai - Panaflo
up
27°C
43°C
21
0.28
0.32
down
26°C
47°C
25
0.33
0.38
Zalman 7000
down
24°C
40°C
18
0.26
0.27
Stock Intel HSF
down
25°C
46°C
24
0.34
0.36
°C rise refers to the rise in temperature over the ambient at load.

At the maximum fan speed the stock Samurai cooled quite well but was also pretty noisy. It sounded a lot like the stock Intel HSF at full speed: A lot of air turbulence noise and a very audible whine. There's no way this could be used in a quiet system, but it sure did cool well.

At full speed the Intel fan sounded as bad as the Samurai fan but it didn't cool as well. The Zalman at 12V was certainly audible, but not nearly as objectionable as the two other fans. At full fan speed, the Zalman cools as well as the Samurai but its noise is lower in level and pitch, but still not acceptably quiet by SPCR standards.

Next I swapped out the Samurai fan for the 80mm Panaflo L1A, running at 12V. It didn't cool quite as good as the stock fan, but it was also noticeably quieter. It had less air turbulence noise and the electrical/mechanical noise was less objectionable as well.

After running the tests with the Samurai fans in their default "blow up" position I flipped them over and reran the tests with them blowing down onto the heatsink in the more normal style. Both the Samurai and the Panaflo fans performed a few degrees worse in this orientation, showing that the Samurais' design does work best with the fan blowing up.

(For °C/W - TDP calculations, Intel's TDP of 66W was used. For °C/W - MP calculations, CPUHeat & CPUMSR Projects' estimate of 75W was used)

Mid-speed Fan Results
Heatsink
airflow direction
idle
load
°C rise
°C/W MP
°C/W TDP
Samurai - stock fan
up
26°C
44°C
22
0.29
0.33
down
28°C
47°C
25
0.33
0.38
Samurai - Panaflo
up
29°C
58°C
36
0.48
0.54
down
27°C
57°C
35
0.47
0.53
Zalman 7000
down
24°C
41°C
19
0.27
0.29
Stock Intel HSF
down
28°C
55°C
33
0.47
0.50

Next up were the mid-speed fan tests. I set the stock fan speed to 2475 rpm, the middle of the rpm range. The Panaflo, Zalman and Intel fans were tested at 7 volts. At this setting the Samurai fan was much quieter than at full speed. The whine was still fairly noticeable but was at a lower level. The air noise was also quieter but was still easy to hear from one meter away. A slight amount of clicking started to become audible as the fan noise and electro/mechanical noise started to recede into the background. I was impressed by it's cooling to noise ratio at this middle speed range.

The 7 volt Samurai fan sounds a lot like an 80mm Panaflo L1A at 12 volts. This level of noise would be acceptable in a system running around 30-35dBA/1m. The Zalman still worked very well at 7V and its noise was pretty unobtrusive, consisting of a slight bit of air noise and a quiet clicking. The Intel HSF was already starting to breath pretty hard at this voltage and still had a bit of whine and air noise.

The Samurai did not coo very well with the Panaflo L1A at 7 volts, blowing up or down. It's just not moving enough air at 7 volts to cool effectively. It's the quietest of all four fans at this mid-speed setting.

Repeating the tests with the Samurai fan in the blow down mode gave results almost exactly like the 12 volt tests, with the blow up orientation a few degrees cooler.

Low Speed Fan Results
Heatsink
airflow direction
idle
load
°C rise
°C/W MP
°C/W TDP
Samurai - stock fan
up
28°C
51°C
29
0.39
0.44
down
28°C
53°C
31
0.41
0.47
Samurai - Panaflo
up
28°C
74°C
52
0.69
0.79
down
28°C
72°C
50
0.67
0.76
Zalman 7000
down
26°C
45°C
23
0.33
0.35
Stock Intel HSF
down
28°C
64°C
42
0.60
0.63

At 5V the stock Samurai fan still cooled decently and was very quiet. There was just the slightest bit of air noise and a very slight amount of clicking that sounded more like very quiet static than actual clicking. I was impressed with its cooling-to-noise ratio at this minimum speed. Of course the good 'ol Zalman was still working just fine down at 5V. Its cooling performance was still very good and it's noise level was very quiet. I could barely hear any wind noise and just the slightest bit of very quiet clicking from about .5-1 meter away. The Intel HSF is basically useless at 5V. Load temps were pretty high and its whine and hum were louder than any of the other fans in this test.

Switching from the Samurai fan to the Panaflo at 5 volts was a basically worthless endeavor. The Panaflo at 5 volts gave load temps in the low to mid 70°C range, which is unacceptable. It is the quietest of all the tested fans at their lowest settings, but much quieter than the stock Samurai fan. The Samurai fan at min speed is nearly as quiet as the Panaflo but cools over 20°C better.

Scythe claims that their Samurai cooler will cool "all Pentium 4 processors" but I only tested it with the (relatively) moderate output of the 2.4C processor. We can however, use the "°C/W" numbers to extrapolate the theoretical performance of the Samurai to any other P4 processor, including the 100+ watt P4 3.4GHz Prescott. According to Intel's "TDP" specs, the 3.4 Prescott puts out 103W. The higher "MP" power rating for the 3.4Ghz Prescott is 115 watts so we'll use those numbers to calculate the estimated maximum CPU temperature for a 3.4E running 2xCPUBurn:

Extrapolated Samurai temps for a 103W, 3.4GHz Prescott using the "TDP" power rating

Fan RPM
C/W
°C rise
load
3350 (max)
0.27
28
50°C
2475 (mid)
0.33
34
56°C
1500 (min)
0.44
45
67°C

Extrapolated Samurai temps for a 115W, 3.4GHz Prescott using the "MP" power rating

Fan RPM
C/W
°C rise
load
3350 (max)
0.26
30
52°C
2475 (mid)
0.29
33
55°C
1500 (min)
0.39
45
67°C

Depending on how much faith you have in these calculations, it does look like the Samurai is capable of cooling the hottest running current P4 CPU, at least down to its mid-speed fan setting. Whether this noise level is acceptable depends on your choice of system hardware, ambient noise level and personal noise tolerance.

I mentioned earlier was that there's no good way to determine whether the mounting tension is correct. I initially tightened them down as much as I felt comfortable. I found that my original setting was fine; tightening it any more made no difference in load temps. I was able to loosen it up enough to give higher temps, but it was seemed fairly easy to tell that this was not enough pressure. Hopefully the average builder will be able to determine the correct amount of pressure, since the mounting instructions are rather vague.



Previous 1 2 3 4 Next

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