Zalman has been quick to cater to the needs of noise-conscious enthusiasts. The company offers a wide range of quiet cooling products, from audacious Reserator and monster CNPS7700 CPU coolers to passive GPU coolers like the ZM80. The latest addition to Zalman’s cooling lineup is the VF700 GPU cooler, which promises nearly silent noise levels and adequate cooling, even for GeForce 6800 Ultra-class graphics cards. Read on as we put Zalman’s new cooler through the wringer.
Zalman offers the VF700 in two flavors: the all-copper VF700-Cu and the copper/aluminum hybrid VF700-AlCu. Apart from the materials used in their construction, the coolers are essentially identical. However, there is a weight difference between the twothe 270g VF700-Cu weighs 50% more than the 180g VF700-AlCu.
The VF700’s fan-like design borrows heavily from Zalman’s recent processor coolers, although concessions have been made to squeeze the cooler onto a graphics card. At 91mm x 126mm x 30mm, the VF700 is definitely a double-wide design. Despite its size, the VF700 is compatible with a wide range of graphics cards. Zalman also maintains a list of incompatible cards.
Zalman uses a 75mm fan with dual ball bearings to cool the VF700. The fan comes with a three-pin power connector, so it won’t plug directly into a graphics card. Instead, Zalman supplies a four-pin Molex adapter that offers Silent and Normal fan speed settings. The Silent plug uses 5V power and spins the fan at 1,350 RPM, while the Normal setting uses 12V power and cranks the fan up to 2,650 RPM.
Flipping the VF700 reveals an impeccably smooth base. On the VF700-Cu, the entire heat sink is copper, including the base. On the VF700-AlCu, which is pictured above, the majority of the cooler is aluminum. The cooler’s copper component is a strip that runs down the middle of the cooler to make contact with a graphics chip.
The VF700 is more than just a graphics chip cooler, though. Zalman also claims that the fan’s design improves circulation and cools the entire graphics card, including its memory chips. A set of eight BGA memory chip heat sinks is even bundled with the cooler.
In addition to memory chip heat sinks, the VF700 also comes with all the mounting hardware you’ll need to secure it to a graphics card. The bundle also includes a tube of thermal compound, the Molex fan speed connector, and a case badge. Zalman even throws in a couple of extra screws and rubber grommets, just in case.
Although the VF700’s array of bundled screws, grommets, and brackets looks a little daunting, mounting the cooler is actually quite easy. Installation requires a screwdriver, but for a cooler of this size and weight, I’d prefer to use real metal screws rather than flaky plastic push pins anyway.
Depending on the graphics card, you may need a little extra clearance at the top of the card for the VF700’s cooling fins. Because of the way GeForce 6600 GT AGP cards are laid out, 30mm of clearance is required above the top of the graphics card.
Once again, keep in mind that the VF700 is most definitely a double-wide design. In fact, it may even obscure a second PCI slot, depending on the motherboard layout. At the very least, PCI cards shouldn’t be put right up next to the fan where they could obstruct air flow.
Fortunately, the VF700’s back bracket is short enough to avoid clearance issues on the other side of the card. That makes the VF700 much easier to accommodate than Zalman’s passive ZM80 GPU cooler, which slaps massive heat sinks on both sides of the graphics card.
The VF700 is only held in place by two contact points, which allows the heat sink to rotate slightly when installed. Still, the bracket is quite secure; you’d need to really torque the graphics cardlikely damaging the board in the processto break the cooler’s contact with the GPU. Rubber grommets prevent the mounting nipples from making direct contact with the PCB, and all screws are exactly the right length to prevent over-tightening.
Installing the VF700’s memory heat sinks is also a snap. The heat sinks come with double-sided tape, so you can just peel and stick. Be sure to clean the surface of the graphics card’s memory chips before installation, though.
All tests were run three times, and their results were averaged, using the following test systems.
|Processor||Pentium 4 3.4GHz Extreme Edition||Athlon 64 FX-51 2.2GHz|
|System bus||800MHz (200MHz quad-pumped)||HT 16-bit/800MHz downstream
HT 16-bit/800MHz upstream
|Motherboard||DFI LANParty 925X-T2||MSI K8T Master2|
|North bridge||Intel 925X MCH||VIA K8T800|
|South bridge||Intel ICH6R||VIA VT8237|
|Chipset drivers||Intel 184.108.40.2067||Hyperion 4.55|
|Memory size||1GB (2 DIMMs)||1GB (2 DIMMs)|
|Memory type||Micron DDR2 SDRAM at 533MHz||Corsair CM72SD512RLP-3200/S Registered PC3200 DDR SDRAM|
|CAS latency (CL)||3||3|
|RAS to CAS delay (tRCD)||3||3|
|RAS precharge (tRP)||3||3|
|Cycle time (tRAS)||8||8|
|Graphics||XFX GeForce 6600 GT with ForceWare 71.84 drivers||eVGA GeForce 6800 Ultra with ForceWare 71.84 drivers|
|Hard drives||Seagate Barracuda 7200.7 NCQ 160GB SATA||Seagate Barracuda 7200.7 160GB SATA|
|OS||Windows XP Professional|
|OS updates||Service Pack 2, DirectX 9.0C|
We’ve put together a couple of test systems for the VF700s. The first is built around a GeForce 6600 GT graphics card with factory overclocked memory running on an open test bed. The second system is equipped with a GeForce 6800 Ultra and running in an Antec Sonata enclosure. For each card, we tested with the stock cooler and the VF700-AlCu and VF700-Cu with their Silent and Normal fan speed settings.
We used the following versions of our test applications:
The test systems’ Windows desktop was set at 1280×1024 in 32-bit color at an 85Hz screen refresh rate. Vertical refresh sync (vsync) was disabled for all tests. Game tests were run with high quality detail levels at 1024×768 with 4x antialiasing and 8x anisotropic filtering.
All the tests and methods we employed are publicly available and reproducible. If you have questions about our methods, hit our forums to talk with us about them.
Noise levels were measured with an Extech 407727 Digital Sound Level Meter placed one inch from the graphics card and out of the path of direct air flow. For this test, we also included a passively-cooled graphics card to illustrate the system’s base noise levels. Since none of the cards feature load-dependent GPU fans, tests were conducted idle.
The VF700s’ Silent fan speed setting is nearly that, producing system noise levels less than one decibel higher than with our passively-cooled graphics cards. With both graphics cards, the Silent and even Normal fan speed settings offer significantly lower noise levels than stock cooling.
So the VF700s are quiet, but can they adequately cool our graphics cards? To find out, we used the temperature monitoring function within NVIDIA’s ForceWare graphics drivers to track GPU temperatures at idle and after several runs through our trhaze DOOM 3 demo at 1024×768 with 4x antialiasing and 8x anisotropic filtering. The ForceWare drivers also report ambient results that measure temperatures around the graphics card. Tests were conducted with DOOM 3 running in a window to ensure that temperature readings were collected when the card was under full load. Temperatures stabilized after three to four runs through the timedemo, although we did six full runs with each configuration, just to be sure.
With the GeForce 6600 GT, the VF700s have no problem delivering lower GPU and ambient temperatures than the stock cooler. The VF700-Cu has a slight edge over the AlCu, but the results are very close.
The VF700s don’t fare quite as well with our GeForce 6800 Ultra. In Silent mode, both coolers produce higher GPU temperatures at idle and under load, although only by a few degrees. In Normal mode, only the VF700-Cu produces lower GPU temperatures than the stock cooler under load.
Looking at ambient temperatures, the VF700s do marginally better than the stock cooler with their Normal fan speed settings, but not in Silent mode.
For kicks, we also tested whether the VF700s had any impact on the overclockability of our GeForce 6600 GT and GeForce 6800 Ultra graphics cards. We used the “Coolbits” registry tweak to unlock the ForceWare drivers’ built-in auto overclocking tool and determine the highest stable GPU and memory clock speed for each card. By default, our GeForce 6600 GT runs at 500MHz core and 1200MHz memory (normal GeForce 6600 GTs run their memory at 1000MHz, but our card is factory overclocked), while the GeForce 6800 Ultra runs at 425MHz core and 1100MHz memory.
Unless you’re the type to get excited about an extra 10MHz, our memory overclocking results don’t show much advantage for the VF700s.
The VF700s don’t yield significantly higher GPU core overclocks, either.
The VF700s certainly deliver on their promise of lower noise levels. In Silent mode, they’re barely louder than passive cooling, an impressive feat for a fan spinning at over 1,000 RPM. The Normal fan speed makes a little more noise, but it’s still much quieter than both of the stock coolers we tested.
Of course, noise levels aren’t the only performance metric for GPU coolers; temperature levels are also important. With our GeForce 6600 GT, the VF700s easily managed much lower temperatures than the stock cooler. Our GeForce 6800 Ultra results aren’t as stellar, though. Only the VF700-Cu managed lower GPU temperatures than the stock cooler under load, and then only with the Normal fan speed setting. Still, even the worst VF700 performance only produced GPU temperatures 10% higher than the stock cooler. I can certainly live with that given the coolers’ relative silence.
The VF700-AlCu and VF700-Cu sell for $27 and $31, respectively. If you’re into quiet computing, that’s pretty affordable, especially since the coolers could conceivably last through a graphics card upgrade or two. If I had to pick one, I’d go with the VF700-Cu. The VF700-AlCu is certainly competent, especially with a GeForce 6600 GT, but for only $4 more, the VF700-Cu offers an all-copper design that improves performance just enough to justify the price gap.