Review: Arctic Cooling VGA Silencer

Cooling
Viewing page 2 of 2 pages. Previous 1 2

COOLING PERFORMANCE

Reviews by other hardware sites

Here are some words from a very good roundup review of six different ATI-9800 GPU-based video cards conducted recently (Sept 2003) by OC Prices Hardware. One of the cards examined was a HIS Excalibur IceQ 9800 Pro 128MB, which comes equipped with this very same cooler, as shown in the photo below. Of course, the labeling is HIS, but the cooler is clearly the same. (Arctic Cooling lists this review on its site.)

To quote Ben Williams who wrote the OC Prices Hardware review,

"...I was pleasently (sic) surprised by almost complete silence... Overclocking the IceQ however obtained staggering results. In silent mode we were able to get the core of the Excalibur up to an impressive 462MHz. This is already some 21MHz quicker than the previous leader. Jacking the fan speed up to full allowed us to push the core yet further to 474 MHz without any visible anomalies being evident. This is the highest stock speed we have seen from a 9800 Pro without water or phase-change cooling, a staggering 94MHz quicker than the stock card... the HIS Excalibur 9800 pro is without doubt the overclocking champion of the test... the overclockability of these cards seems to be directly proportional to the quality of the coolers on the card."

Not surprisingly, Ben concludes that the VGA Silencer-equipped IceQ "stands head and shoulders above the competition in terms of overclockability." HIS lists other reviews that concur with this assessment about the efficiency of the cooler on the Excalibur IceQ 9800 Pro.

My Own Tests with the Sapphire 9500

Unlike Brett Wasserman's GeForce4 Ti4200 card that was used for the Zalman HP-80A VGA heatpipe cooler, I did not have the luxury of a thermal dioded embedded in the GPU. ATI only just made a temp diode available in their newest and greatest card, the 9800 XP. Previous 9xxx ATI cards do not have a thermal diode.

A case could be made for the use of temperature probes to measure the heatsink temperatures and make comparisons, but that's an inaccurate way of judging the actual core temp, which would have to be guessed at from the temperature of the heatsink.

I'm of the opinion that the temperature of the GPU core is largely irrelevant. For us techno-geeks who want to know every minute bit of data possible, that is a hard pill to swallow, but what really matters is stability in the most tortuous conditions: If it doesn't crash, then it's cool enough. And if it does get too hot, the card will cause the system to hang or revert to software rendering before any damage can be done.

The idea of an open air testbed was abandoned, since it is only inside the case that the problems with conventional VGA cooling becomes apparent.

For comparison, my store-bought sample of the other silent-minded VGA cooling champ, Zalman HP-80A VGA heatpipe cooler, was tested in the same way. The fanless HP-80A is what I have been using for quite a while now, having yanked off the original small whiny fan soon after I acquired the 9500.


The double-sided Zalman competition.

System Test Bed for Arctic Cooling VGA Silencer:

* Fong-Kai FK-330 Case, with a Panaflo 120mm L1A@5v exhausting, no intake fans (The Fong Kai case is new, it is very good and a review is coming!)
* Abit NF7 motherboard, with latest BIOS revisions. Modified with passively cooled NB heatsink.
* AMD XP2100+ Thoroughbred B CPU, stock speed and voltage (1.65v, 62.1 Watts max)
* Thermalright SLK-800 heatsink with Panaflo 80mm L1A@ 8v
* 1GB of PC2100 RAM.
* Seagate Barracuda IV hard drive.
* Fortron Aurora 300W PSU, with 120mm fan set at lowest setting
* Temperatures read with Motherboard Monitor 5.
* Sapphire 9500 graphics card

Note: The Sapphire 9500 card is run as a 9800 Pro, with modified Catalyst 3.8 drivers. (My monitor didn't even explode once!) For all appearances, the card behaves exactly like a 9800 Pro -- with lower clock and RAM speeds. All of the "features" of the 9700/9800Pro are available and fully functional.

1) Temperature as an indicator of Cooling Performance

This was considered relevant because of the AC VGA Silencer's Direct Heat Exhaust System feature: Does it exhaust enough hot air out of the case at the low fan speed (the only one most SPCR enthusiasts would use) to make any difference?

3DMark03 (build 330) was used as a stress test. Although a lot electronic ink has been spilled regarding the usefulness of synthetic benchmarks like 3DMark, it remains an easy and effective tool for comparative testing. Its benchmark score is really only useful for a single card in a singular system.

For testing the effects of the two different VGA coolers on the overall case environment, 3DMark03 was run in a continuous loop for an hour, with the before / after case, CPU and hard drive recorded. Ambient room temperature was 23° C.

Temperatures at standard clock speed: 275/270 MHz, GPU/RAM

Zalman HP-80
Case
CPU
HDD
Idle
26° C
30° C
32° C
After 3DMark2003 for 1 hr.
28° C
41° C
33° C

*

AC VGA Silencer LOW
Case
CPU
HDD
Idle
26° C
30° C
32° C
After 3DMark2003 for 1 hr.
26° C
39° C
32° C

While a 2° C improvement isn't earth shattering, it is significant in the new Fong Kai case, which I've set up with a 120mm exhaust fan. In combination with an amazingly open front intake vent, it has excellent airflow. In a case with less effective airflow, I think the temperature difference would be much higher.

2) Overclocking as an indicator of cooling performance: The Sapphire 9500 card was overclocked to its maximum stable speed with each cooler. This test is often used by reviewers, but the results should not be taken too seriously. There are simply too many variables involved in overclocking to put all the blame on the cooler for the limits of overclockability. Since neither of these coolers has any memory cooling component, only the core speed was adjusted during these tests.

I started at the default speed using RivaTuner by Guru3D, then ran the game tests in 3DMark03, increasing clock speed by 10 MHz after each pass until the card froze or crashed.

The maximum overclock speed reached:

Stock Sappphire 9500 cooler
325 MHz
Zalman HP-80A
330 MHz
AC-VGA Silencer (fan on Low or High)
385 MHz

As cautioned above, the overclocking results are open to some interpretation:

A. The fact that the AC-VGA-S reached much higher clock speed suggests that the HP-80's results were limited by its cooling ability. Yet if that were true, increasing the fan speed on the AC-S should have raised the limit, at least slightly. It did not, however.

B. This might be a more plausible explanation: The clock speed is limited by cooling with the Zalman HP-80, and limited by the GPU core itself in the VGA Silencer, in which case, no amount of extra cooling power would improve the clock speed, at least with this specific card.

2. NOISE

I am in complete agreement with MikeC on the noise. He got it right in the preview, which I reiterate here almost word for word.

At the low (1200 rpm) fan setting:

Subjectively, it is about as quiet as a Panaflo 80L fan at 6V. In other words, it is essentially inaudible from 1 meter distance in a room of 18 dBA ambient noise. Up close, it has a smooth character. There may be a trace of buzz or clicking but the level is so low it is irrelevant. The noise could not be measured because at 1 meter, it was below the ambient 18 dBA noise level of the room . I suspect it was well below.

At the high (2400 rpm) fan setting:

Subjectively, it is somewhat louder than the Panaflo 80L fan at 12V. It is easily heard a couple of meters away. The character of the noise is benign; it does not have any particular annoyances like nasty whining of buzzing. It is mostly the whooshing of wind turbulence noise. It is much preferable to the whiny, buzzy, high-pitched noise of 30-60 mm fans typically used on VGA cards. Mounted on a VGA card installed in a case, the fan noise would likely be muffled; it is possible that vibration-borne noise could occur. (But this is a possibility with any fan of this size or larger spinning at this speed inside a case.) The noise measured 31 dBA at 1 meter distance.

CONCLUSION

Arctic Cooling keeps up their recent run of rock-solid quiet cooling performers for a bargain basement price. As seen in the case temp tests, the Silencer basically takes the VGA card out of the case cooling equation. Having a high powered VGA card has minimal effect on case temps with the Silencer installed. In a smaller case with poorer ventilation, having it on may actually reduce case temps. Given its excellent CFM/noise ratio, the Arctic Cooling VGA cooler might even be useful with a low powered VGA card as a replacement for a case fan!

Compatibility is a bit more limited with the Arctic Cooling VGA Silencer than with the Zalman HP-80. It only fits the ATI 9500-9800 series, while the Zalman fits on both nVidia and ATI cards. Plus, there is the new Zalman 80C-HP, which can be fitted with a 80mm fan. (A review on this product is ALSO on the way!) Admittedly the fan on this new Zalman does not blow the hot air out of the case, but it's cooling performance is surely better than without a fan. Of course, there is a cost issue: The new Zalman cooler + fan is about double that of the Arctic Cooling unit.

Good design, low noise, and the attractive suggested retail price of just US$22 makes the VGA Silencer well worth a try for ATI Radeon 9500-9800 card owners who'd like to drop the noise without sacrificing stability and cooling. (At time of publication, the VGA Silencer is definitely available from sample supplier Frontier PC in Vancouver, BC, Canada. SVC in San Jose may also have stock.)

My final words: Considering that the AC-VGA Silencer isn't audible over the other fans, cools the case, OC's higher and even does 2D work perfectly fine with the fan unplugged altogether... It's staying in my system!

Much thanks to Arctic Cooling and Frontier PC for the VGA Silencer samples.

* * * * *

Discuss this article in the SPCR Forums.



Previous 1 2

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