The Card Cooler - Page 4

Too “Cool” To Be True?
Based on the earlier points raised, The Card Cooler works best if a ventilated side-casing is available to provide enough external cool air in-take for circulation. This is especially so if you have a narrow casing like me.  Without this ventilation, air in-take may become stifled (insufficient air in-take forming the bottleneck) especially since The Card Cooler’s air in-take is situated just laterally underneath the side casing. In any case, it should still improve internal air circulation on a whole, rendering a drop in system temperature, albeit not as significant or optimal with holes present in the side-casing.

I also reckon that The Card Cooler manufacturers were spot-on in opting to blow air downwards towards the slots (as opposed to drawing hot air away) for three main reasons. Firstly, as discussed before, directing a focused stream of air onto the small surface area of the chip that is relatively distant from the fan, performs more rapid and stable cooling. Secondly, graphics chips are usually mounted with their own heatsink + fan combo, whose fan is blowing onto the heatsink itself (at least my GeForce is). Hence, directing air towards it would also assist in its airflow, but reverse ops may negate it. Thirdly, if The Card Cooler had conversely drawn air away from the card, its focused air out-take would have been hindered by the side-casing, rendering it highly inefficient. Unless you have amply drilled holes on your side casing to allow outflow, it would be far better to draw air from along and around the side casing and blow air towards the component instead. I guess the manufacturers must have foreseen that many would not possess holes in their side casings and opted for a sure-safe method instead.

With that in mind, let’s evaluate the temperature drop rendered by The Card Cooler.

As usual, my system specs first for reference:

  • P3 450 @ 558MHz (4.5 x 124MHz)
  • 128MB LGS RAM
  • 01 x 13GB Quantum CR Ultra-ATA66 HDD (5400 rpm); 01 x 10.1 GB Seagate Ultra-ATA66 HDD (5400rpm); 01 x 13.2 GB IBM Ultra-ATA66 HDD (7200rpm)
  • 01 x Creative GeForce Pro (32MB DDRAM onboard) using 3.62 reference Nvidia drivers; core/mem settings at 145/355;
  • 01 x Sony 6X DVD-ROM drive
  • SBLive!! Platinum (Liveware 3.0 drivers)
  • VirataLink 1025 ATM card
  • 17” Mitsubishi DiamondScan 70 monitor (0.28 dpi)
  • Win 98SE
  • Ambient Temperature of 25 degrees Celsius.
  • 300W Power Supply
  • 01 x Elan Vital chassis fan drawing air into casing;
  • Another 01 x Elan Vital chassis-sized fan + 01 x Cooltium chassis-sized fan (stacked on each other) placed after the preceding Elan Vital chassis fan to enhance air-flow & stream air past my PCI / AGP slots.  BTW: The Elan Vital & Cooltium fans are the most powerful I’ve ever encountered (highly recommended) for their size.
Yeah, call me “Paranoid”, but it works well!
With this setup, I don’t think it’ll get any cooler!
This is a nifty setup. Without these placeholders, my dual fan setup would “fall back” owing to the reaction force of the strong air current reaction they produce…Also, I can just slide the fans out whenever I require to change motherboard…

The combined setup appears pretty efficient with The Card Cooler drawing cooler air from its surroundings and directing this at my MB PCI / AGP slots, whilst the 2 stacked up fans blowing air perpendicularly away (past the slots), thereby improving circulation.

So, is it As Cool As It Gets?

Cooling Results
To gauge the effects of The Card Cooler on my GeForce’s temperature, I tested it with & without my existing stacked-fan setup. For all tests, the sole Elan Vital chassis fan drawing external air into my casing was retained. I also used a thermocouple and fastened that onto my GeForce heatsink to note its temperature variations at different stages. 

a. Temperature Findings:-

In order to stress the core, I ran a looping demo of Dagoth Moor Zoological Gardens at 1024x768/32bit. Note that from my previous benchmarking experiences, DMZ is probably one of the ultimate tests to manifest any hint of instability in GeForce ops. I also kept the core/mem at 145/355 (its optimal o/c speeds under my existing setup). 

As a prelude to each setup, I would do a fresh boot and wait for about 10 mins for the system temp to equalize upon boot-up into the desktop, before I measured the start temp and commenced the demo. I would then check on the MAX function of the thermocouple after another 30 mins (if it hadn’t hung already), and note the peak temperature attained during the demo. 


Dagoth Moor Zoological Gardens (1024x768/32bit) Looping demo; GeForce DDR @ 145/355; Ambient of 25 deg C; All temperature measurements taken from heatsink of GeForce

Test Setup Starting Temp (deg C) Max Temp (deg C) Remarks
Card Cooler + Twin Fan setup 29.3 31.8 No lockups
Card Cooler alone 29.5 31.8 No lockups
Twin Fan setup alone 30.8 34.1 No lockups
No fans at all 40.0 43.4 Locked up at 43.4 deg C during 5th cycle of demo


Dagoth Moor Zoological Gardens (1024x768/32bit) Looping demo; GeForce DDR @ 145/355; Ambient of 25 deg C; All temperature measurements taken from heatsink of GeForce

Test Setup Starting Temp (deg C) Max Temp (deg C) Remarks
Card Cooler + Twin Fan setup 33.0 35.0 No lockups
Card Cooler alone 33.7 34.9 No lockups
Twin Fan setup alone 35.5 39.0 No lockups
No fans at all 43.8 48.4 Locked up at 48.4 deg C during 2nd cycle of demo

As one can infer by comparing the 2nd and 3rd rows, just having the Card Cooler alone offers a decent 2-5 deg C drop in temperature over my original dual fan setup. This is further compounded if I hadn’t had my existing dual-fan setup (ie. 4th row), where it offers an amazing 10-13 deg C drop in temperature! Incidentally, it is also apparent that coupling my twin fan setup with the Card Cooler only offers incremental cooling. And that the benefits derived from a plain Card Cooler is far greater than my dual fan setup alone. So despite the stifled air-intake under a closed casing setup, The Card Cooler still manages to deliver a significant drop in temperature, especially for setups without existing forms of cooling.

Additionally, one may inquire why the system hung at such a large temperature disparity between closed and open environments (48.4 deg C for closed casing; 43.4 deg C for open casing) during the final setup without fans. Although I don’t have a confirmed explanation to this, I believe at higher temperatures (say > 40 deg C), its operational stability also seems largely dependent on prolonged strain placed upon it. Hence, although the card did lock-up at a lower temp of 43.4 deg C (open casing), it was only after 5 cycles did that occur. In contrast, the card locked up at 48.4 deg C for the closed casing setup, just prior to completing its 2nd demo cycle. And in this case, the high temperature of 48.4 deg C would have formed a larger attribute to its lockup. 

b.  Overclocking Results:-

Already convinced with The Card Cooler’s usefulness in initiating temperature savings, I proceeded to test its effects on overclocking.  Here I ran it through the DMZ demo as before (but in benchmark mode) under increasing clock / mem settings.  My results as follows:

Under Dagoth Moor Zoological (DMZ) benchmark (1024x768/32bit; Vsynch OFF), I managed to achieve 48.5 fps utilizing a 150/355 mem setting without lockups.  However, trying to overclock to 155 core or 355 mem speeds were unfruitful.  Nevertheless, this still represents a slight improvement in stability for the core, as my previous DMZ tests (see my earlier DDR review) would fail consistently upon a core setting of 150.

As a side-note, I did manage to achieve 155/370 without lockups during benchmarking Q3Demo 1.09 with The Card Cooler installed.  However, I guess Q3Demo doesn’t stress the card as much as DMZ does, and offers a weaker basis for studying true overclocking stability.

So is it a COOL enuff product to get?

< Previous

Next >