Thermalright SI-128: Evolution of a Past Master

TESTING

On the test bench...

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. Under load, it measures 78W minus the efficiency losses in the VRMs.
• ASUS P5LD2-VM motherboard. A basic microATX board with integrated graphics and plenty of room around the CPU socket.
• Hitachi Deskstar 7K80 80GB SATA hard 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.

Test Tools

• Seasonic Power Angel for measuring AC power at the wall to ensure that the heat output remains consistent.
• Custom-built, four-channel variable-speed fan controller, used to regulate the fan speed during the test.
• Bruel & Kjaer (B&K) model 2203 Sound Level Meter. Used to accurately measure noise down to 20 dBA and below.
• Various other tools for testing fans, as documented in our standard fan testing methodology.

Software Tools

• SpeedFan 4.31, used to monitor the on-chip thermal sensor. This sensor is not calibrated, so results are not universally applicable; however,
• CPUBurn P6, used to stress the CPU heavily, generating more heat that most realistic loads. 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.

Noise measurements were made with the fan powered from the lab variable DC power supply while the rest of the system was off to ensure that system noise did not skew the measurements.

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. Every fan was tested at four voltages: 5V, 7V, 9V, and 12V, representing a full cross-section of the fan's airflow and noise performance.

The ambient conditions during testing were 19 dBA and 20°C.

TEST RESULTS

The SI-128 performed acceptably well at every level except 5V, when the temperature began to jump up. A cooler processor (including most Core 2 Duo chips) would probably be fine with good system airflow, but hotter chips will probably require more airflow.

Besides, there is little acoustic difference between 5V and 7V when using our reference fan, so there is very little benefit to running the fan this slowly anyway. At 7V, the cooling performance was good enough for the majority of mainstream processors. Bleeding-edge overclocking chips like the QX6800 or the FX-74 will probably require more airflow — but these are exceptions, not the norm.

Performance at higher airflow levels continued to improve temperatures a bit, but the improvement came at the expense of increased noise from our reference fan. Given the level of performance already achieved at 7V, there's probably not much point in running an increased fan speed.

VS. THE HEAVYWEIGHTS

A comparison against the current high-end heatsinks and Thermalright's old XP-120 confirms the SI-128's position. For most of the lower-airflow points, the SI-128 performed no better than the original XP-120 that is two generations old.

The one point where the SI-128 did perform significantly better was with the fan at 12V, which suggests that the SI-128 was held back by our slow reference fan. The densely spaced fins are not ideal for high performance with low airflow, and it's quite likely that we did not have enough airflow to reach the SI-128's "sweet spot" during our testing.

Some readers might balk at the above comments and ask why a higher airflow fan was not used. The answer is that in our heatsink reviews, we're always interested not in ultimate cooling performance, but performance at low noise levels. With a high speed fan, the SI-128 may provide the best cooling of all the heatsinks mentioned, but this is irrelevant for SPCR, because the cost will be too high a level of noise.

NOISE RECORDINGS IN MP3 FORMAT

Reference 120mm fan: 5V-7V-9V-12V, 5s Ambient between levels: One Meter, One Foot

Comparatives:

Scythe Infinity: 5V-7V-9V-12V, 5s Ambient between levels: One Meter, One Foot

Arctic Cooling Alpine 64: 5V-7V-9V-12V, 5s Ambient between levels: One Meter, One Foot

Scythe Mine w/ stock fan: 5V-7V-9V-12V, 5s Ambient between levels: One Meter, One Foot

Thermaltake Big Typhoon: 5V-7V-9V-12V, 5s Ambient between levels: One Meter, One Foot

FINAL CONCLUSIONS

With a properly chosen fan, the SI-128 is capable of cooling most mainstream processors very quietly. The K8 clip is elegant and quite easy to use; the issues with 775 socket mounting have more to do with Intel's choices than Thermalright's.

However, in a market where the top-dog processors consume no more than 65W and most heatsinks are designed to cope with 130W loads, the low noise, low airflow performance is no longer as impressive as it once might have been. The fact is that there's no shortage of stiff competition, especially at US$50. When the cost of the fan is taken into account, the SI-128 doesn't look like a great deal.

The heatsink market as a whole has evolved past the point where performance is the deciding factor. The SI-128 has good build quality and a convenient K8 mounting system, but that may not be quite enough to set it apart in a crowded marketplace. It's quite good — like a dozen other heatsinks we've reviewed.

As noted in the introduction, it will probably find most of its buyer in the home theater market, where price is no object and tall, high-rise heatsinks are counted out by their size. The SI-128 is just short enough to fit in a HTPC case that supports full-size expansion cards. And, given how poor the airflow is in many HTPC cases, the SI-128 may have a real advantage over the more compact heatsinks out there.

Enthusiasts, overclockers and gamers who don't mind noise may also have fun using a high speed fan with the SI-128. There's little doubt it will perform better with a higher speed fan.

Much thanks to Thermalright for the SI-128 sample.

* * *

Articles of Related Interest

* * *

Comment on this article in our Forums