Xigmatek’s direct heatpipe-to-heatsource technology finds its way to VGA cooling. The big, aptly named, 4-pipe Battle-Axe looks like a serious contender for cooling the hottest video cards. Does it do it well… and quietly?
Xigmatek, a German company founded in 2005, carries several product lines, including power supplies, and both liquid and air cooling solutions. They have been a relative unknown until recently, but the success of their HDT line of CPU heatsinks has propelled them into the limelight. Many enthusiasts are now familiar with their work, and they’re eager for more.
GPU cooling is a smaller market, but no less competitive. There are several popular products, including the Thermaltake Duorb, Thermalright HR-03, Zalman VF1000 LED, and the Arctic Cooling Accelero S1 that dominate the space. The high-performing S1 in particular made us wonder if there was room for further improvement for VGA cooling without resorting to exotic methods. Xigmatek seems to think so, and the Battle-Axe VD964 is their solution. It’s a formidable product on paper, armed with four direct touch heatpipes and two thick 92mm fans. It’s hard to imagine that its performance will be anything but spectacular.
The Battle-Axe box sitting next to our test card with an Accelero S1 installed.
nVIDIA® GeForce® 8800 / 7900 / 7800 Series ATI Radeon™ X1950 / X1900 / X1800 Series
6mm x 4pcs
92(W) x 92(H) x 25(D) mm
2200 R.P.M. (±10%)
The Battle-Axe is a large, dual fan cooler which takes up three extra slots, one for the heatsink itself and two for the 25mm thick fans. It is only slightly wider than a standard, full-sized video card, so it should fit in narrower cases.
The Battle-Axe gets its name from the design of the metallic fan casing. The metal frame is secured to the fans using four screws a piece, and the fans in-turn are mounted to the heatsink at their extremities using rubber fan isolators.
The underside in profile.
At first glance, the underside of the Battle-Axe seemed fairly pedestrian until we noticed that a good portion of the heatsink’s fins are closed. The airflow generated by the two fans blowing down on these sections won’t go all the way through and instead will be diverted to the sides. A through-and-through design no doubt would work much better, especially with low airflow.
The base is typical of Xigmatek’s direct touch coolers. The heatpipes at the base have been flattened smooth but there are shallow groves where they meet the mounting plate.
Heatpipe-mounting plate connection.
PHYSICAL DETAILS CON’T
Viewed from the side, the fins are rather minuscule. Directly below the center of the fans, the contours make the overall surface area much lower than if it had a straight, uniform design. On the fan side, they are slightly concave, presumably to reduce turbulence.
The fans are soft mounted to the heatsink only at the extreme left and right. At the center they are held by the metal frame.
The fin spacing is fairly tight, making it a poor choice for a passive cooler. The overall construction suggests that Xigmatek is counting on the heatpipes and fans to do most of the work.
The fan isolators on each side attach just as they do on other Xigmatek heatsinks. The isolators have slits in the middle, allowing them to be inserted through the edges of the fins.
A view of the base.
The fans are held on by only two fan isolators a piece. However there is room for more at the center and Xigmatek provides four extra isolators. Why they are not pre-installed in the first place is a mystery.
The fans have dimensions of 92 x 92 x 25 mm and are rated to draw 0.17A and spin at 2200 RPM. With these two beasts blowing onto the heatsink, we have no doubt the Battle-Axe will be an excellent performer.
The Battle-Axe comes with mounting hardware, low-profile ramsinks, a heatsink for the NVIO chip found on some Geforce 8800 series cards, four extra rubber isolators, a dual 3-pin fan adapter, and thermal compound.
Accessories and mounting equipment.
Xigmatek supplies six standard BGA sized ramsinks and two larger ones that are long enough to cover three chips on their own. The thermal adhesive on the ramsinks unfortunately do not stick very well and fell off several times when we tried to apply them to our test card. If you wish to attempt to use them, we recommend cleaning the memory chips carefully so that they are dust-free and heating up the adhesive before application to help facilitate bonding.
Mounting arms installed.
The installation procedure is fairly simple. The mounting clips are screwed into the base, and four hex-screws are secured to the appropriate holes.
Back of the card.
On the back of the card, four spring-loaded nuts attach to the hex-screws protruding from the other side, in essence, capping them.
It seems the mounting hardware is somewhat fragile. We slightly over-tightened one of the nuts (despite the fact that it has a spring to prevent this), and were rewarded with the sound of ricocheting metal. A portion of the one of the hex-screws snapped off inside the nut. To keep proper distance on that corner of the mount we improvised, using a small piece of plastic with a hole inside it, approximately the same height as the fat portion of the hex-screw. It was taped on so it wouldn’t move about (see picture below).
A second nut/screw set became stuck together and effectively became one piece (unbinding them would have probably ended badly) which is why the picture below only shows two hex-screws. The Zalman VF series VGA coolers use a similar installation system, but unlike the Battle-Axe, their screws are thicker with coarser threads and the use of thumbscrews keeps fumbling to a minimum. The Battle-Axe’s mounting equipment is flimsy and fidgety in comparison.
Thermal compound footprint.
GPU INTERFACE DETAIL
Also note the small footprint the thermal compound made in the photo above. Only the center two heatpipes made direct contact with the bare 190mm square core of our ATI X1950XTX GPU. As the surrounding mounting plate is not really designed to transfer hea, the usefulness of four heatpipes is questionable, at least for this size of GPU core. Xigmatek may get better results using three heatpipes, or two thicker heatpipes instead.
Examining the base very closely with a micrometer, we determined that a GPU can be up to 215mm square and still not make contact with the two outside heatpipes in the base. For a GPU core to make contact with about 1/3 of the two outside heatpipes in the base, it would have to be 250mm sq. For all four heatpipes in the base to make complete contact, the GPU would have to be 330mm wide.
What all this means is that the Battle-Axe does not interface as effectively with the smaller cores that result from the most modern microprocessor fabrication technology because of the relatively wide spacing of the heatpipes in its base. If a heatspreader is built into the GPU, then Xigmatek’s direct-touch heatpipe will still have advantages. It may be instructive to take a look at the physical size of your GPU.
As we lost a screw portion on one corner, a long screw with the same threading was used instead. Despite the crudeness of our methods, the base made good, tight contact with the GPU core.
Battle-Axe fully installed.
Unlike the Accelero S1, the Battle-Axe is quite easy to handle due to the presence of a metal housing. It’s also a lot narrower, extending only slightly past the edge of the card’s width.
The S2 has a slight curve.
Installed in our test system, the Battle-Axe occupied three extra slots, four including the video card PCB. The metal cover is slightly wider than the width of the fans, creating a sort of trough for the fan cables. Each fan has it’s own 3-pin cable.
Our test procedure is an in-system test, designed to determine whether the cooler is adequate for use in a low-noise system. By adequately cooled, we mean cooled well enough that no misbehavior related to thermal overload is exhibited. Thermal misbehavior in a graphics card can show up in a variety of ways, including:
Sudden system shutdown or reboot without warning.
Jaggies and other visual artifacts on the screen.
Motion slowing and/or screen freezing.
Any of these misbehaviors are annoying at best and dangerous at worst — dangerous to the health and lifespan of the graphics card, and sometimes to the system OS.
A custom-built internal variable fan speed controller to power the system fan
A custom-built external variable fan speed controller to power the VGA heatsink fan (if applicable)
Bruel & Kjaer (B&K) model 2203 Sound Level Meter, used to accurately measure SPL (sound pressure level) down to 20 dBA and below.
A summary of how our video card/cooler test platform is put together can be found here.
Our main test consists of ATITool’s artifact scanner running in conjunction with CPUBurn to stress both the graphics card and processor simultaneously. It is a realistic test that mimics the stress on the CPU and GPU produced by a modern video game, only more consistently. The software is left running until the GPU temperature stabilizes for at least 10 minutes at which point, both the CPU and GPU temperatures are recorded. In addition we also take measurements of the system’s overall noise level and power consumption using a B&K Sound Meter and a Seasonic Power Angel respectively. If the heatsink has a fan, the procedure is repeated at various fan speeds while the system fan is left at the lowest setting of 7V. If it is a passive cooler, the system fan instead is varied to study the effect of system airflow on the heatsink’s performance. If artifacts are detected in ATITool or other instability is noted, the heatsink is deemed inadequate to cool the video card in our test system.
Preliminary testing is also done at idle, and with only CPUBurn running for comparison. For idle results, the system is left stagnant for 10 minutes before ATITool is loaded and the first temperature it reports is used. We do this because on our test platform, after ATITool is loaded, it puts some kind of stress on the GPU, causing the temperature to climb immediately (even if it is left idle for hours beforehand) and the power consumption to increase by approximately 10W. We theorize that initially the card is in 2D mode, either underclocked or undervolted (or possibly both) and that ATITool automatically puts it in 3D mode, which would account for the rise in temperature and power draw. ATITool is left running in the background for the remainder of testing which is why the GPU temperature during CPUBurn will appear higher compared to idle. Consider this the difference between 2D idle and 3D idle.
The ambient conditions were 22°C and 18 dBA at the time of testing.
Results: Xigmatek Battle-Axe
System Fan Speed
CPUBurn + ATI Tool
Fans @ 12V: At 35 dBA it was uncomfortably loud, but we were expecting that since there are two fans, and not one. The majority of the noise was your run-of-the-mill growling turbulence.
Fans @ 9V: The fans exhibited a semi-high pitched buzzing. There was also a grinding noise coming from the motor.
Fans @ 7V: The acoustics were very smooth, with some slight ticking evident at 1m.
Fans @ 5V: The fans were essentially inaudible except for the ticking which could be heard within one foot of the case.
The GPU temperature remained basically constant throughout testing, no matter how fast the fans were spinning. This was a very surprising result, especially considering the difference in airflow and noise between 5V and 12V. This unusual behavior could possibly be the result of the lack of direct contact between the GPU core and the two heatpipes at the Battle-Axe’s extremities. The potential may be there for better temperatures. With numbers like these, we see no reason to have such loud fans. They could probably have gotten away with using just one. We considered attempting to jerry-rig the Battle-Axe with just one fan in the center, but during the course of our review, a second hex-screw broke, making it more trouble than it was worth to continue testing it.
Comparison: Battle-Axe vs. VF1000 & Accelero S1
Note: 1°C was added to the results for the VF1000 as the ambient temperature was 1°C higher at the time it was tested.
The Battle-Axe is significantly better than the Zalman VF1000, but falls short of our favorite, the Accelero S1 with a single Nexus 120mm fan. With two loud 92mm x 25mm fans, we were expecting it to blow the Accelero S1 out of the water, but the issue of no-contact for the outside heatpipes limits its performance.
The Battle-Axe is an exceptional VGA heatsink, capable of cooling all compatible video cards — if it can cool a X1950XTX this well, it can cool a Geforce 8800 series card easily. While the specifications do not list ATI HD 3800 series as compatible, it does not look incompatible. It may be the absence of a VRM heatsink that prevents it from being officially supported. We do recommend that some kind of fan control be used, and a fan swap or modification would not be a bad idea as the acoustics were not great.
While it performs superbly, from a usability standpoint, our experience with the product was primarily negative. It’s not often that mounting hardware gives us so much trouble in the lab, and even rarer for a screw mount to actually break… twice! It was simply a pain to install and uninstall, and we were afraid we would not be able to finish testing. The weak ramsink adhesive was also inexcusable.
As effective as its cooling is, the Battle-Axe’s performance probably could have been better on a larger GPU, because two of the four heatpipes did not make direct contact with our ATI X1950XTX core. Any performance advantage the two outside heatpipes provided here were minimal.
The nVidia 8800 series GPUs appear to have considerably larger cores than our ATI, and some of them incorporate heatpipes as well. The latest 9800GTX series appear to use essentially the same size die as the 8800GTX. All of these video cards would probably benefit more from the Battle-Axe’s four heatpipes than our ATI X1950XTX. Interestingly, more recent ATI GPUs like the 3850 and 3870 series are smaller, about the same size of the X1950XTX. so they would have the same disadvantage with the Xigmatek.
The underlining direct-touch technology seems even more impressive considering the size of our test GPU. Despite the challenges, the Battle-Axe performed admirably, beating out the more expensive Zalman VF1000. There is definitely more potential here than we tapped. That great potential cooling performance may be compelling enough for many potential buyers to overcome the longer list of cons in our summary below.
Xigmatek Battle-Axe VD964
Pros * Excellent performance with any amount of airflow * Fans easily replaced
Cons * Fans much louder than needed * No fan controller * Terribly fragile mounting hardware; virtually unacceptable. * Weak ramsink adhesive * Takes up three extra slots * Heatpipe spacing in base too wide for a small GPU