Antec Mini P180: A micro-ATX P182

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

THERMAL ANALYSIS

Test Tools

Software Tools

  • SpeedFan 4.31, used to monitor the on-chip thermal sensor. This sensor is not calibrated, so results are not universally applicable.
  • 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.
  • ATITool utility to stress and monitor the graphics card.
  • 3DMark06 game simulation benchmark. (Mostly not used as the power load was not constant or high enough.)
  • Throttlewatch 2.01, used to monitor the throttling feature of the CPU to determine when overheating occurs.

TEST RESULTS

Ambient conditions were 21°C and 19 dBA. ATITool's "scan for artifacts" load was applied during load testing. The total AC power draw is about 5W lower than running it in default mode, but the stress on the GPU is higher, which means it gets hotter. GPU heat is more difficult for most systems to handle compared to CPU heat. Heatsinks for the latter are still in a higher state of development than for GPUs.

Configuration #1: With integrated graphics, standard fans -

This stress load was bound to be easily handled by just the big fan alone, set to low. The 120mm fan was left unplugged. In this "negative pressure" configuration (meaning air is forced to be sucked into the case via free vents as opposed to blown out), the only exhaust is through the 200mm fan, and all other holes in the top chamber become intake vents. The 120mm fan, for example, becomes an intake, as do all the other vents on the back panel.

Configuration #1: IGP, 200mm fan @ low
LOAD
AC
CPU
HDD
SPL
idle
65W
28°C
38°C
23 dBA@1m
CPUBurn x2
131W
48°C
38°C
ATITool + CPUBurn x2
133W
48°C
38°C

Never mind that this Scythe Ninja Plus Rev. B CPU heatsink sample didn't fare that well on the test bench. Here in the flow of the 200mm fan, it's just coasting. Note that the hard drive temperature remained constant at a very low temperature regardless of CPU load. This is expected as the "HDD" is upwind of the CPU, and the airflow across it is constant. The Enermax Modu82+ power supply fan did not budge beyond minimum speed. It was utterly inaudible, swamped by the noise of the big fan.

The overall noise signature was set by the big fan, first, then the WD hard drive in front. When the front door was open, the increased noise of the drive whirring could be heard from a meter away, although the overall SPL increased by only one or two decibels. The acoustic contribution of the Modu82+ was inaudible... and stayed inaudible throughout the testing.

Configuration #2: With ATI X1950XTX, stock fans -

Many combinations of fans and settings were tried. When both stock fans are turned on, it is still a "negative pressure" configuration. Air is blown out only through the fans; all other holes in the upper chamber are intakes.

Configuration #2: With ATI X1950XTX, stock fans
LOAD
20cm Fan
12cm Fan
AC
GPU
CPU
HDD
SPL dBA@1m
Idle
low
off
97W
42°C
31°C
39°C
23
CPUBurn x2
low
off
161W
42°C
50°C
39°C
23
low
low
162W
42°C
48°C
37°C
24
ATITool + CPUBurn x2
low
off
250W
86°C
58°C
39°C
23
low
low
251W
82°C
55°C
37°C
24

The jump in power caused by the ATI X1900XTX is dramatic, even at idle, where 32W more heat has to be dealt with in the system. Under CPU + GPU loads, power demand virtually doubled compared to Config #1. Our testing of the Enermax Modu82+ 625W PSU showed its efficiency to be 85.9% with 235W AC input, and 86.4% with 287W AC. Assuming 86% efficiency, the DC power demand with the ATI card is 216W, compared to ~110W with the integrated graphics (calculated based on the efficiency of the Modu82+ 625, about 82% at 133W AC input).

CPU temperature increased substantially at all settings. It remained well under 60°C at load, however, and never came even close to throttling, which we know (from previous experience with this CPU/board) doesn't occur until >70°C. The temperature of mid-80s°C is not that high for this GPU. We do not see any misbehavior with 99% of applications on this VGA card till the GPU temperature is well into the 90s and beyond.

Keeping the back 120mm fan on helps with cooling, as evinced by the small drops in all the monitored temperatures. At load, the heat of the air flowing out the 120mm fan can be easily felt. Whether the gain is worthwhile is a judgement call. Personally, I'd swap out the stock fan with a quieter one if I felt the need for its additional cooling.

Configuration #3: ATI X1950XTX, 200m fan, and 120mm intake fan -

Since running just the top 200mm fan gave temperature results that were safe but perhaps a bit high for some users' comfort, an intake fan was tried in the vent closest to the VGA card (in the lower HDD bay). The expected result was a significant reduction in GPU temperature, and perhaps a small drop in CPU temperature as well.

A Scythe Slipstream 120mm fan rated at 800 RPM was used. Initially, it was run at full speed (12V), but it was judged to be too audible. A damping resistor was used to pull the speed of the Scythe fan down just a bit, to 680 RPM, where its noise contribution dropped to below that of the WD hard drive just above it.

Config #3: ATI X1950XTX, 200mm fan + 120mm Scythe intake fan
LOAD
20cm Fan
12cm Fan
AC
GPU
CPU
HDD
SPL dBA@1m
ATITool + CPUBurn x2
low
off
250W
86°C
58°C
39°C
23
low
680rpm
251W
78°C
55°C
39°C
23.5

The greatest improvement was seen in the GPU temperature. The CPU ran a bit cooler as well. Overall SPL was hardly affected, and subjectively, there was no audible penalty for running the Scythe Slipstream 120mm fan at this speed. It sounded the same with the Scythe fan on or off.

Configuration #4: With the Big Fan on Med -

Some readers will undoubtedly want to know how much cooling improves with the Big Fan on medium. Let's not keep you in the dark. The first line in the table is Config #3.

Config #3: ATI X1950XTX, 200mm fan + 120mm Scythe intake fan
LOAD
20cm Fan
12cm Fan
AC
GPU
CPU
HDD
SPL dBA@1m
ATITool + CPUBurn x2
low
680rpm
251W
78°C
55°C
39°C
23.5
med
680rpm
251W
78°C
52°C
35°C
28
med
880rpm
251W
75°C
52°C
35°C
30
low
880rpm
251W
78°C
54°C
38°C
27
med
880/low*
252W
73°C
51°C
35°C
31

Turning the big fan to medium led to a 3°C drop on the CPU and a 4°C drop on the HDD, but the GPU temperature did not change.

So, the front fan was turned up to full speed. That effected a 3°C drop in the GPU temperature, and no other change. Slowing the big fan down to low while keeping the front fan at full speed pushed the GPU temp back up to where we began, and the CPU and HDD both up by 2~3°C.

Finally the last test has an asterisk in the "12cm Fan" column — it's trying to indicate that the stock back panel fan was also turned on. The difference in noise between this setting and the middle one (big fan on med, front Scythe fan on full) is very minor, hard to hear. There is a measurable improvement in cooling.

When set to medium speed, the big fan's air turbulence noise (whooosh) dominates, along with a bit of rattling from it at close distance. It's not an unpleasant sound, but considerably higher than the lab's ambient acoustics, higher than most home offices' ambient level. Some folks won't mind it; there are no obvious tonal nasties.

The front fan works against some degree of impedance when the door is closed. This may be one reason that the overall noise seems to increase higher than expected when the Scythe fan in front is running at its full 880 RPM speed (according to SpeedFan and the motherboard monitoring).

MISC NOTES

1. HDD location - You may recall from a previous photo that the WD drive was mounted in the top HDD bay, in the slot closest to the left (main) side panel. I wondered whether putting it in the middle, allowing air to flow more easily around both of the drive's sides would help keep it cooler. The change was made, and with just the big fan on low, the temperature did drop in idle, from 39°C down to 37°C.

2. Dust buildup will occur despite the filters. Because the fans are set up in a "negative pressure" configuration, air is pulled in from all holes in the case other than the exhaust fans. While the front vents are filtered, there are many holes on the back panel that are not. This means you still need to pay some attention to cleaning out the system once in a while, depending on environment and positioning.

3. The effects of the main front door, the slatted intake vent covers, and dust filters were all explored while the system was under full CPU/GPU load (ATITool + CPUBurn x2), in Configuration #3 above. The results are summarized below.

Thermal Impact of Doors and Covers
Main Door
Vent Covers
Filters
GPU
CPU
HDD
closed
closed
on
78°C
55°C
39°C
open
closed
on
74°C
53°C
37°C
open
open
on
72°C
52°C
37°C
open
open
off
71°C
51°C
36°C
closed
open
on
75°C
53°C
37°C

Opening the main front door had the biggest temperature effect, especially on the video card. This option also had the biggest impact on noise; it rose just a little over one dBA@1m, maybe two, but the subjective increase in HDD noise was significant. This is important if the system is placed on the desk.

The inner vent covers had a smaller effect, and the filters made hardly any difference at all. All of these effects will vary with the amount of air you're trying to push through the intake vents; i.e., with the number of fans used, and the speed they're set to.

The best compromise of acoustics, cooling and cosmetic appeal is probably removing the vent covers, while keeping the front door closed. This is the last combination detailed above. There are some significant improvements in cooling, without any acoustic penalty, and the filters do help block dust. Because the internal steel chassis is painted black, the appearance is not severely degraded when the door is open; it's still monolithic black.


It looks OK to me with the vent covers removed.

5. The power supply seemed to get plenty of airflow through the bottom facing vent to keep itself cool under the test loads. Its fan never sped up. This may not hold true for all PSUs; the Enermax Modu82+ is exceptional in its acoustics. Do not place this PC directly on a carpeted floor. On a carpet, it must be raised up a bit to make sure there's clearance enough so the bottom facing vent doesn't get blocked. Also, if the case is placed on the floor, check and clean that bottom filter often.

6. An alternative design idea: Since the lower chamber intake vent is on the bottom, why not just extend the vent to directly under the PSU, position it "right side up" and allow it to draw air directly from the outside? Then the partition wall could be eliminated without any effect on overall cooling, a couple of inches of could have been shaved off the height, and the overall price could have been reduced substantially as well.



Previous 1 2 3 4 5 6 7 Next

Cases|Damping - Article Index
Help support this site, buy the Antec Mini P180 from one of our affiliate retailers!