Antec ISK 300-65 Mini-ITX Case

Cases|Damping
Viewing page 7 of 8 pages. Previous 1 2 3 4 5 6 7 8 Next

TEST RESULTS - still continued

C. With Noctua 80mm fan: Arctic Cooling Alpine Pro heatsink, fan removed

It was time to try a different heatsink. This time, the Arctic Cooling Alpine 7 Pro was called into service. With the integral fan it stands too tall to fit, so the fan was removed. Load testing was done again. Alas, the results were only a couple of degree better than the with the fanless stock Intel cooler. Time to put a fan on the heatsink.

Noctua NF-R8 fan Case Fan + AC Alpine 7 Pro, fanless
Fan
dBA@0.6m
CPU
GPU
7V
19
88°C
68°C
9V
25
78°C
65°C
24°C ambient

D. With Noctua 80mm fan: Stock Intel Heatsink, fan at very low speed

Since it was about the only heatsink/fan on hand that would fit, a stock Intel socket 775 cooler was installed, with fan. It is a PWM fan, and the Zotac board has a PWM CPU fan header, so it was plugged into the board. The BIOS fan control was used to select the lowest fan speed. On this particular Intel heatsink/fan, that speed turned out to be 460rpm. It barely made any sound at this slow speed, but instinct suggested it would be enough to keep the CPU temperature under control.


Minor variant stock Intel heatsink/fan: The fan is frameless.

Noctua NF-R8 Case Fan + Intel stock cooler @ 460rpm
Fan
dBA@0.6m
CPU
GPU
5V
17
70°C
72°C
7V
20
67°C
69°C
24°C ambient


There were no obvious tonalities in the sound. Note the peak near 20kHz; this is from the power adapter, which was placed right next to the case on the desk. It may be audible to young people with excellent hearing. Simply placing the power adapter on the floor makes the sound immeasurable from the microphone position.

Finally, the component temperatures fell to a safe range. The noise level was quite modest. With the Noctua fan set at 5V, the overall noise was quiet enough at the seated user position to be acceptable over extended periods even for individuals with low noise tolerance.

E. With Two Noctua 80mm fan: Stock Intel Heatsink, fan at very low speed

It was the last thing left to try, in the absence of more suitable heatsinks, and deemed worth doing by the SPCR team of proofreaders and subeditors. Sigh. Installing the second fan with anti-vibration plugs was tricky and tedious due to the lack of space to pull the plugs through. In the end, only one pair of the plugs was used, on opposite corners. The other corners were damped and secured with some double-sided sticky foam tape. A Y-splitter was used to parallel-wire the second fan with the first to the Zalman Fanmate controller. That extra fan would pull up to another 1.32W which seemed small enough not to be an issue for the power supply.


Two Noctua fans in place. You can see the mess of wires on the left, including the Y-adapter. It still all fit not too badly though.

Two Noctua NF-R8 Fans + Intel Stock cooler @460rpm
Fans
dBA@0.6m
CPU
GPU
5V
21
67°C
54°C
24°C ambient

So, as expected, the increased airflow of two fans lead to lower temperatures all around. The drop in GPU temperature was nearly 20°C, which suggests that the second fan may have changed the airflow path across the motherboard and exposed the NB chip heatsink to greater airflow. The maximum AC power of the system also dropped by 3W, which suggests higher efficiency of power components (VRMs on the motherboard as well as the DC/DC converter) due to improved cooling.

The cost in noise at the 0.6m seated user psotion was about 4 dBA, from 17 dBA to 21 dBA. Subjectively, it was still unobtrusive and very quiet, though no one could say it's inaudible. Putting the unit behind the monitor brought the noise down to 17 dBA again. This would be a useful strategy for those seeking minimal noise from any small desktop computer.

POWER DETAILS

Upon request, Antec provided detailed specifications of the power adapter and the DC/DC board used in the ISK-300. The power adapter is made by Delta Electronics, who has been working with Antec on other power supply products in the past year or so. Curiously, the DC/DC converter turns out to be made by the ODM (original design manufacturing) Product Division of Gigabyte. The most salient specs of the two power components are summarized below.

Antec ISK 300-65 AC/DC power adapter Specs
Input 90 ~ 264VAC, 50/60Hz
1.5A Max at 65W load, 100VAC input
Output 19V (18.4~ 20V), 3.42A
Efficiency More than 85% average efficiency of 25%, 50%, 75% and 100% load at 115/230 VAC, measured at the end of DC cable.
Output Ripple & Noise <350 mV peak to peak (tested at 25° C).
Protection Over Voltage, Over Current, Short Circuit, Over Temperature
Temperature Operating range: 0 to +40° C
Case Temp. Rise: 45° C at 40° C Ambient
Antec ISK 300-65 DC/DC Converter Specs
Input 19V, 65W max.
Output 5V/ 7A Max.
12V/5A Max.
3V/6A Max.
5V_SB/ 1A Max.
-12V/ 0.1A Max.

During thermal/load testing, the AC power consumption at full load started at 60W and increased slowly to a maximum of 66W. With the coolest configuration tested, the maximum power dropped slightly to 63W. This shows that the motherboard VRM and/or the DC/DC converter efficiency drops as they get hotter. It may also be the case that both were red-lined by the system's power demand when the system was not cooled adequately. The second fan caused no clearly measurable increase in power; the increase was probably within the margin of error of the power meter (under 1W).

At idle, power draw was 30W, which is quite low, and it suggests very high efficiency, very close to the picoPSU with 80W or 110W power brick. In previous testing, same motherboard and components drew 35W at idle and 65W at full load with a 400W 80+ Seasonic power supply.

During full load testing, the power brick became quite warm, but never too hot to touch. The brick would likely get hotter with higher ambient room temperature than in our lab, where it never rises above ~25° C. The DC/DC board temperature was not checked. These devices typically have >90% efficiency, so what little heat it produces would be far less than any of the other parts within the case, namely the CPU and the motherboard components.

When the computer was turned on and working, the power brick emitted a low level of ultra high frequency noise (near 20kHz) which should not bother any people. It's too low at level and too high a frequency. We cannot comment what effect it might have on dogs and bats.

When the computer was turned off or put in standby mode with the power adapter still plugged in, the buzzing noise moved down in frequency to the midband and became much more audible. It could be just heard a meter away. While this can be disturbing, it's quite common with power bricks. Most of the time, such noise is not heard because power bricks are usually stuffed behind a desk or other other furniture, usually on the floor.

MISC NOTES

1. The system is quieter, subjectively, when placed vertically on the stand than when set horizontally on its rubber feet. In the latter position, there is greater mechanical coupling between the case and the table, leading to a slightly higher level of hum. This was difficult to measure, but fairly easy to hear.

2. The HDD activity LED connector spans 3 pins. Our test motherboard's HDD LED pins are next to each other, however, so it was not used. It's a simple fix; just cut the connector in two or remove one of the wires from the connector, and move it next to the other wire.

3. If more room over the motherboard is desired, it's possible to remove the back end of the HDD tray frame by popping out two rivets. Two screws still hold the tray to the front of the chassis, and this is probably enough to keep the tray stable even with a slim optical drive and a 2.5" drive. No guarantee — we did not try it.

4. The CPU used in our test setup is rated for 65W TDP, which is the same for all Intel dual core C2D processors today. We would probably have obtained better cooling results with a 45W TDP AMD dual-core, but a suitable AMD mini-ITX board was not on hand. There's little question that the Zotac IONITX with dual-core Atom 330 processor would be easily cooled in the system, especially with a fan directly on that board's heatsink.

5. Finally, we still think a more ideal heatsink — one that occupies the entire space and height available with large fins surface area not too tightly spaced — would allow good cooling of the components with just one quiet 80mm case fan.



Previous 1 2 3 4 5 6 7 8 Next

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