Viewing page 3 of 3 pages.
Previous 1 2 3
TESTING
SPCR's standard heatsink testing method of a Panaflo 80mm low speed fan at 12, 7 and 5 volts. is not applicable here. The design calls for no fan to be used. There's no set way to mount one either, although it is easy to rig up if so desired.
Test platform
Intel P4-2.53 processor. Nominal power is 61.5W; may increase to ~75W if speed throttling doesn't stop it first. 71? C rated maximum junction temp.
Gigabyte GA-81RXP motherboard - Intel 845 chipset; on-die thermal diode monitoring. Modded with large passive heatsink on NB chip for fanless cooling.
Matrox G400 Max VGA card modified with Zalman chipset HS for fanless operation
256 MB DDRAM - PC2100 generic
Hitachi Deskstar 180GXP 30G hard drive (in Smart Drive silencing enclosure)
DigiDoc5 w/ thermal sensors
Antec SLK3700BQE case
Antec SmartPower 350S power supply (Single 80mm fan PSU supplied with SLK3700BQE case)
120mm case fan (1200 RPM max) supplied in SLK3700BQE case
Sunbeam Rheobus fan controller (to control 120mm case fan speed)
Arctic Silver Ceramique Thermal Compound
CPUBurn - Stress program used to load the CPU to 100%. It heats up the CPU a bit more than just about any other utility tried so far.
Motherboard Monitor 5 - Software utility used to monitor temperatures.
TEST RESULTS
1. On the test bench
The system was assembled out in the open with the motherboard flat on the test bench. No fans were anywhere near the CPU heatsink of motherboard. The measurements reflect MBM5's log after ~30 minutes of activity in each category of use.
|
Open Bench Test
|
| Conditions |
Idle
|
Normal use*
|
100% CPU load
|
| CPU Temp |
38?C
|
44~58?C
|
>68?C - freeze
|
*Normal use much as described by TS Heatronics: web surfing, email, word processing, and playing music or DVDs
The thermal overload under 100% load using CPUBurn occured in little over 10 minutes; the program crashed. As cautioned by the manufacturer, the NCU-1000 is not designed for this kind of stress. However, the straight up vertical position is also not ideal, as indicated by Silicon Acoustics and in the instruction manual provided by the manufacturer.
2. In the case, first try
The system was set up in the Antec SLK3700BQE case with the heatsink fins oriented for best vertical airflow, as per instructions. While the power supply does not employ a 120 mm fan, there is a 120 x 25 mm case fan on the back panel, mounted with soft rubber grommets. It is very quiet but can move a LOT of air; Antec does not specify any details about this fan but they did state that its rated speed at 12V is 1200 RPM. I'd guess the airflow at that speed to be at least 50 CFM, based on specs of other similar fans.
The front hard drive bays were left completely open for maximum airflow. The hard drive in the Smart Drive enclosure was placed on a piece of foam on the bottom of the case. Its noise contribution was negligible.
|
In-case Test #1
|
|
Case fan
|
CPU Temperature
|
|
Idle
|
Normal use*
|
100% CPU load
|
|
12V
|
38?C
|
40~45?C
|
>67?C - freeze
|
|
8V
|
38?C
|
41~46?C
|
>67?C - freeze
|
|
5V
|
44?C
|
41~48?C
|
>67?C - freeze
|
*Normal use much as described by TS Heatronics: web surfing, email, word processing, and playing music or DVDs.
Three things stand out here:
1) In normal use, the temperature is generally lower in the case than it is on the open test bench. This is probably because the 120mm case fan produces more airflow than just convection, and the cooling property of the HS is much better with it horizontal.
2) In normal use, the system is perfectly stable, no crashes or freezes. As noted, the maximum CPU temps vary a bit with case fan speed, but stay under 50C.
3) It still overheats with 100% CPU load -- which is not recommended by the manufacturer.
3. In the case, 2nd try
I decided it was worth trying the cooling mounted the "wrong way" with airflow more optimized to go front-to-back, since the back panel fan was pulling a lot of air through. See the setup below.
|
In-case Test #2
|
|
Case fan
|
CPU Temperature
|
|
Idle
|
Normal use*
|
100% CPU load
|
|
12V
|
38?C
|
39~42?C
|
>67?C - freeze
|
|
8V
|
38?C
|
40~44?C
|
>67?C - freeze
|
|
5V
|
40?C
|
40~47?C
|
>67?C - freeze
|
Temperatures actually improved a bit during normal use but CPUBurn still crashed. It might have taken a few minutes longer for the crash to occur though. So it seems that because of the greater airflow through the back case fan rather than the power supply, this fin orientation provides slightly better cooling.
3. In the case, 3rd try: With 120mm PSU
Well, I didn't give up. Instead, I pressed a Seasonic Super Tornado 300 into service. This is a power supply with a 120mm fan, as recommended by Silicon Acoustics. That fan is capable of blowing something like 70 CFM, but the voltage controller is set up to keep the fan speed very low till above the ~120W DC draw. This system was unlikely to pull that much power, so the PSU fan probably would not spin up much and thus not provide enough airflow for the Zen. We would see.
The Heatlane Zen had to be rotated again. Pull the cooler off, remove the copper base, rotate 90 degrees. Then reattach the base, and install the cooler back on.
Now with those huge 120mm fans at right angles to each other within a couple of inches from the Zen, I began to wonder about turbulence effects. I could picture the air swirling like eddies in a stream, just spinning the heat round and round but going nowhere. Hmm... Banish the image!
Back to the temperature measurements.
|
In-case Test #3
|
|
Case fan
|
CPU Temperature
|
|
Idle
|
Normal use*
|
100% CPU load
|
|
12V
|
38?C
|
39~42?C
|
>67?C - freeze
|
|
8V
|
38?C
|
39~43?C
|
>67?C - freeze
|
|
5V
|
39?C
|
39~44?C
|
>67?C - freeze
|
The normal use temperatures are best with this setup, despite my misgiving about potential nasty interactions between the airflow of the 120mm fans. With the case fan at 5V, this setup is very quiet, measuring just under 25 dBA / 1m (in a 22 dBA room ambient).
CONCLUSIONS
The manufacturer's application notes make it abundantly clear the the Heatlane Zen cooler, when run as intended without any direct cooling fan, is not designed to cool a P4 CPU that is being stress tested with a constant 100% load. This was verified in the testing with many configurations.
When used exactly as directed -- web surfing, email, word processing, and playing music or DVDs -- the Heatlane Zen NC-1000 kept the P4- 2.53 processor well cooled and perfectly stable with virtually any combination of power supply and case fan speed in the test setup. Excellent case airflow should be an objective for all PC silencers, but those who wish to use the Zen (fanless) will want to pay particular attention to good case airflow.
In thinking through about the ideal use for the Zen, I have to admit that Silicon Acoustics' suggestion of a 120mm fan PSU makes sense. But to that recommendation, I would add that it is ideal with a P4 that has relatively lower heat dissipation. In other words, I am not in complete agreement about the "up to P4-2.8" (>68W) recommendation for this cooler. Perhaps if the case cooling provides 50 cfm across the heatsink fins? The P4-2.53 (61W) used in the test may represent something close to the practical high limit of this cooler, if it is run fanless.
If it is provided forced airflow across its fins, I have little doubt that the Zen could easily handle >3.0G CPUs. It would no longer be fanless, but that is a somewhat dubious distinction. Given its size and the proximity of the CPU to the PSU fan and/or case exhaust fan, there will almost always be some forced airflow across its fins in any typical system.
One thing I would strongly recommend against is to use this heatsink with a hot P4 in a fanless case. As long as you use it within its limits and intended role, the Zen will be a cool and noiseless addition in your quiet system. Overclockers note: Good case airflow along with, say, 60 CFM directly across these fins, ideally with a 120mm fan, will probably allow you the highest overclock possible with any P4.
All in all, the Scythe / Heatlane Zen NCU-1000 Cooler is a very inventive heatsink for silent operation.
Our thanks to Scythe for the Heatlane Zen NCU-1000 sample and their kind support.
* * * * *
Comment on this article in our Forums.
| Help support this site, buy from one of our affiliate retailers! |
|
