Scythe Musashi ReviewThe Smith -
» Discuss this article (19)
I will test the Scythe Musashi by recording the GPU temperatures of the HD4850 at idle and load. They will be provided by the GPU-z 0.30 Utility. I will run ATI Tool Scan for artifacts function to load the GPU. Each temperature is measured thirty minutes after beginning the test, at idle and load. Also, I will use the thermal paste provided with the cooler, since the stock thermal interface is still in place on the stock cooler, which will be the comparison cooler. I will first run that reference design cooler at default speed, which is 35% duty cycle, and finally, for testing purposes, at 100%. However, I'm pretty sure nobody would want to run it at that speed for normal use, because it's so noisy, and as annoying as a crying baby. For the Musashi, I will run the test at the lowest speed the fan controller can provide me, and another time at the highest speed.
- Processor: Intel E6750 C2D
- Motherboard: Gigabyte X48-DQ6
- Memory: Mushkin XP2 Redline 8000 2 x 2GB 5-5-5-12
- Video Card: ASUS Radeon HD 4850 w/ Catalyst 8.12(Stock & 680/1050)
- Power Supply: Mushkin 800w Modular Power Supply
- Hard Drive: Seagate 7200.11 750GB SATA w/32MB Cache
- Optical Drive: Pioneer DVR-212D BK
- OS: Windows Vista Ultimate Edition SP1
- Ambient Temperature: 20.5 degrees Celcius
I'm very impressed with the temperatures the Scythe Musashi can provide. So here I repeat, on an HD4850 overclocked, the temperatures did not exceed 43C, at full fan speed. At idle, it went as low as 30C. There was almost no difference between the stock and overclocked temperatures. That's amazing. I was a little scared by the 100C and higher temperatures the stock cooler gave, but with the Musashi, your graphics card is in good hands. I would not even hesitate at all to install it on a voltmodded card.
One last thing I have to say is that I also connected both Scythe fans to the motherboard headers to be able to monitor their speed. If you remember, I said earlier that they were supposed to vary from 800 to 2000RPM. Well, one was not able to go lower than 1000RPM. That's more than twice the error margin of 10% written in the specifications. However, the other one was right on the specifications. They were also very silent at low speeds. The noise was easily masked by other fans in the testing setup.