Shuttle's Smallest Yet: XPC X100

TESTING

Our testing examined the effectiveness of the cooling system and how much noise it made. Noise, power consumption, and CPU temperature were all measured in three different states: Idle, CPU load, and CPU and GPU load together. The following tools were used:

• B&K model 1613 sound level meter
• Extech Power Analyzer / Data Logger 380803 to measure power consumption
• Core Temp to measure CPU temperature. When the two cores reported different temperatures, the higher temperature was reported.
• CPUBurn to stress the CPU.
• ATI Tool 0.25 Beta 15 to stress the GPU.

The ambient conditions at the time of testing were 18 dBA@1m and 22°C.

At idle, the X100 consumed 42W — low, but surprisingly high for a system built of laptop parts. Engaging Speedstep by switching the Power Option in Windows from Home / Office Desk to Minimal Power Management provided a worthwhile drop of 6W, down to 36W. By way of comparison, Apple's iMac, which contains almost identical hardware, consumed 46W at idle — including the power required by the LCD monitor.

The temperatures at idle, with or without Speedstep, seemed quite high — normally we would expect a notebook processor to hover much closer to ambient. However, two factors are worth paying attention to:

1. Core Temp report the value of an on-chip digital temperature sensor, not the thermal diode that is reported by most motherboards. Intel says this digital temperature sensor is more accurate than thermal diodes, which rely on motherboard-embedded circuitry to translate the output into degrees. According to Core Temp, the CPU will not throttle until the thermal sensor reports 85°C — significantly higher than what we would normally expect.
2. Until the temperature hit 62°C, the fan did not run at all.

Intuitively, leaving the fan off until it is absolutely needed seems like a good idea. And, when the fan was off, the X100 was indeed very quiet: The only source of noise was the hard drive, which emitted a hollow whoosh that was easy to ignore. However, because the cooler was not quite good enough to prevent thermal rise even at idle, the fan came on every five minutes or so. The result was a somewhat inconsistent noise character with the fan cycling on and off, with a small burst every time the fan started up. The fan was not loud when running, and added only a little background hum to the noise character. It would have been preferable to leave the fan running at this level all the time instead of having the fan draw attention to itself every time it started. However, the difference with the fan on and off was quite small, both subjectively and in measured sound pressure level.

Unfortunately, attempting to tweak the fan controller in the BIOS was unsuccessful, as there was very little BIOS to tweak. Aside from boot order, there were no options to speak of — it was less flexible than my beige-box P-133 from 1996!

The qualitative noise character was judged to be better under load — it was no louder than when the fan was running in idle, and it stopped cycling on and off. The core temperature rose to 67°C, well under the 85°C threshold for throttling. The system stayed perfectly stable under load (it did freeze once in idle, though), so it seems that 67°C is cool enough.

Throwing the VGA card into the mix increased the power draw by only three watts, but those three watts were significant: The additional heat made the system audibly noisier as the fan increased in speed to compensate. Changes in fan speed were slow and gradual, and, even at its loudest, the system was still quiet enough for use in an entertainment center.

Realistically speaking, very few people will ever run both the VGA card and the main processor at full load for long periods of time, so perhaps the increase in noise is not so important. However, the system could be quite sensitive to its surroundings. Poor airflow or a slightly higher ambient temperature could easily force the fan to speed up at an earlier point.

The peak power consumption of 61W was quite respectable; it was in line with what we've seen from other systems that use notebook processors.

One final note: The hard drive was highly audible and intrusive when seeking. Seagate's recent drives have a reputation for sharp, noisy seeks, and the DB35 was no exception. The rigid mounting system didn't help; vibrations from the seek caused the whole case to resonate.

SHUTTLE X100 VS. APPLE iMAC & SHUTTLE SD11G5

Apple's Core Duo-based iMac and Shuttle's SD11G5 barebones system make good points of reference for the X100. All use notebook processors to keep heat down, and all were fairly quiet on our test bench. The exact details of the hardware are not quite the same, but all should provide roughly the same level of performance. Detailed information about the iMac and the SD11G5 can be found in their respective reviews.

In terms of acoustics, the iMac is the best of the bunch. It's quieter at both idle and load, and resonance from the hard drive or exhaust fans is not an issue. It managed to damp the 3.5" HDD noise well enough that only from the back could the drive be identified as a specific source of noise.

The X100, on the other hand, is the noisiest of the bunch. The undamped, full size hard drive limits it too much to compete with the notebook drive that we used in our SD11G5 testing. With the same drive, we would expect the two Shuttle systems to be fairly similar acoustically, although the larger 92mm fan in the SD11G5 is probably smoother sounding than the little 60mm fan in the X100.

Despite the numbers, the iMac is the most power efficient of the bunch; its power readings include the built-in LCD monitor, which consumes some 25W.

MP3 RECORDINGS

Shuttle X100 — Idle (Fan off): 23 dBA@1m: One Meter, One Foot

Shuttle X100 — Idle / Load (Fan on): 25 dBA@1m: One Meter, One Foot

Shuttle X100 — HDD Seek noise: One Meter, One Foot

COMPARATIVES

Antec NSK3300, Config 1 (System Fan @ L): 24 dBA@1m: One Meter, One Foot

Antec NSK3300, Config 2 (Rear Fan swapped to Nexus @ 5V): 23 dBA@1m: One Meter, One Foot

Lian Li PC-101, Config 1 (No Intake Fan): 24 dBA@1m: One Meter, One Foot

Lian Li PC-101, Config 2 (Intake Fan @ 5V): 26 dBA@1m: One Meter, One Foot

CONCLUSIONS

The X100's small, trendy form factor and the less-is-more approach to its design are also notable in Apple's Mac Mini. The similarity extends to the technical details, too: The Core Duo processor and the use of mobile parts are both features of Apple's Mac Mini and iMac lines. (Editor's Note: That was true until the beginning of this month; the iMac has now moved to Core 2 Duo processors exclusively.)

While the X100 is certainly one of the best integrated PC we've seen, it's still not quite as refined or as quiet as the iMac.

The X100 will do better as a media center than a home PC. Its styling and small size make it perfect for the living room, and the 21~24 dBA@1m noise level is still quieter than many PVR or cable boxes. Windows MCE 2005 is steadily gaining popularity, and it doesn't cost too much extra. However, the X100 doesn't come with a remote control or a VFD. An external remote control can be added, but having an infrared receiver hanging off the front panel may ruin the aesthetic appeal of the device for some users.

Ultimately, the X100 should find its own niche. Although it borrows much from Apple, it's far from a clone, and there is plenty to like about its small form factor and quiet performance. 24 dBA@1m out of the box is pretty darn good for a complete off-the-shelf system, even if it can't complete with the quietest systems that enthusiasts build from for themselves.

Much thanks to Shuttle for supplying the X100 sample for us to review.

*

SPCR Articles of Related Interest

Apple iMac w/Intel Core Duo: A User's Review
17" iMac: The Official SPCR Review
Shuttle SD11G5: Pentium M SFF PC

* * *

Discuss this article in the Silent PC Review Forums.