Running dual video cards, and having weird reboots and freezes? Can you prime95 all day, but not play a game? Might be your chipset overheating!
Cadaveca takes us through the pros and cons of a number of aftermarket cooling solutions for chipsets, an area often dismissed without thought.

In the past months we have seen a plethora of motherboards and chipsets debut…between AMD’s AM2 and Intel’s Core 2 Duo, supporting parts are popping up on a seemingly monthly basis. With the added power and functionality these new platforms provide, it can prove very difficult to keep these chipsets cool, with most OEM’s going so far as to adopt heat pipe solutions on their “Enthusiast” lines. However, being “Enthusiast” products, these products tend to end up being used by just who they were intended for…enthusiasts.
Not satisfied with what’s been given, seeking even more performance, or just seeking silence, an enthusiast needs some options, and quite a few companies are willing to provide. Today we look at four of these options, three passive coolers, and one active cooler. Let’s take a look…
First up we have the ZALMAN ZM-NBF47, a blue aluminum passive cooler. ZALMAN has a history of providing quiet cooling, and proclaim themselves to be “leading the world of Quiet Computing Solutions”. Having many different products and a contract with a certain gamer, they are well known.

Inside the package are the necessary mounting clips for both looped mountings and push-pins for thru-holes, a small tube of thermal paste, a blue case badge sporting the new “Zm” logo, and some instructions. The “flower” shaped heat sink is built in the same style as the rest of ZALMAN’s flower-style products…a series of aluminum fins are sandwiched together with a couple of bolts, anodized, and given a flat bottom…this one weighs in at 57grams and measures 81(L) X 37(W) X 47(H)mm.
 The ZALMAN logo can be seen on each fin and they claim “The use of ZM-NBF47 with any of the following CPU coolers will provide a more stable and noiseless PC environment: CNPS7000 Series, CNPS7700 Series, CNPS8000 Series, and CNPS9500 Series.”…presumably due to the airflow that the mentioned coolers provide. The bottom is milled to a shiny surface that really betrays the fact that this cooler is assembled from about 26 parts.  At the angle provided in the previous picture it’s easy to see the finish and although the reflection isn’t crystal clear, it still appears highly polished with very faint mill marks marring the reflection. The cooler mounts attach to the bottom of the heat sink with a small screw and nut, and the package has different arms for loops and pins, which can be flipped over for added length. Due to its size, it cannot be mounted in every motherboard with every configuration, so please check ZALMAN’s site for motherboard compatibility. But what about the others? Thermaltake’s chipset cooler, the Extreme Spirit II, is a heat pipe based cooler in all copper, with an attached 40x40x10mm LED fan. Opening the package reveals the unit itself, nuts, paper washers and bolts to attach to the motherboard, thermal paste, a circular mounting pad, and instructions. Wrapped around the cooler is the fan’s power cord, well sleeved in black nylon. Thermaltake claims this unit can install on any motherboard that has mountings that use loops or push-pins, and although they are correct, they don’t mention anything about how they accomplish this feat might interfere with other installed components. More on that later. The cooler has a solid copper base attached to a small copper heat pipe, and is cooled by a set of copper fins which act as a radiator for the gas enclosed in the heat pipe. The fan is attached to the radiator with a shroud and two screws, and the mounting clip is screwed to the middle of the base and has adjustable metal arms attached to the clip with screws as well. Although the cooler is 48x40x70 mm, roughly the same size as the ZALMAN cooler, if placed on its side, it is a hefty 122grams more for a total weight of 169grams. The circular base of the unit comes with a plastic film over it to prevent scratches in transit…remember to remove it before installing! So as to easily compare the finish I used the case badge from the ZALMAN cooler as a reference item.
You can easily see the mill marks on the base by peeling back the cover, unlike ZALMAN’s cooler, but the reflection given is a much clearer image due to the direction of the lapping job, and the marks appear finer in comparison. The plug for the fan is of the three-pin variety, most commonly found on motherboards, but there is not a Molex adapter for those that like to not have fans lugged into their boards at all. In attempting to mount the unit to test, I ran into a few problems…because of the design of the mounting, although it may appear to be positioned in any fashion, I found issue with clearing video cards on motherboards that have chipsets between their graphic slots…most notably the “LanParty” series from DFI, even though Thermaltake’s website claims that the cooler fits all motherboards. I had planned on testing these coolers on multiple platforms, but this problem quickly limited me to the use of only one of the many boards I have lying around. Something to keep in mind, at the very least.
The third and fourth coolers hail from Thermalright, known for their XP-90 and XP-120 CPU heat sinks, each of which enjoyed being one of the best air-coolers, upon release. Their newest line of chipset cooler features two models, the HR-05 High Riser and the HR-05 SLI High Riser. To meet their name these coolers have a heat pipe attached to a 77x77mm radiator, and a nickel-plated copper base.

Upon opening the box I was greeted by one of the most carefully packaged heat sinks ever…it made such an impact on me that I had to make mention of it here. Carefully nestled in the foam are the clips to attach the heat sink, a rather large tube of thermal paste, a couple of push pins, a Thermalright sticker, mounting instructions, and the heat sink itself, in a rather heavy plastic bag.
 The HR-05 weighs in at 145grams, a wee bit lighter than the Thermaltake cooler, and measures 77.9 x W19 (fin), W30 (base) x H105 (mm), without fan. The HR-05 SLI weighs 124grams and measures L78xW19 (Fin), W62.5 (Include body) xH105mm. Looking at the base, the entire image of the ZALMAN sticker is obscured. The fine mill marks left by the lapping process seems further “enhanced” by the nickel plating. Yes folks, that’s a copper base with a coating! Taking a look at the HR-05 SLI, we can easily see the differences between the two chipset coolers…most obvious is that the SLI model uses 2 heat pipes instead of one. Will this affect cooling? Let’s take a look at the unit first! The SLI model uses a slightly different base than the non-SLI model as you can see in the picture; however this one is nickel-plated copper as well. The pipes do not cover the middle of the heat sink, instead are slightly away from the mid-point. On its sister model, the heat pipe goes directly through the middle of the block…in one side and out the other. So, other than the use of a second heat pipe, and the “radiator” being slightly offset from the top, these units are identical. This is the stock cooler, our reference point. It comes from an ASUS A8N-SLI, the motherboard used for testing. Prior to doing any testing, the heat sink was removed from the board, and re-attached, using “Artic Silver5”, or AS5. By using the same thermal interface material, we can ensure that the factor that may affect any performance differences between these heat sinks is minimized. On to the testing! Test System: - ASUS A8N SLI NF4 Motherboard
- AMD Athlon 4400+ X2 w/ stock cooler
- Mushkin XP4000 2GB DDR
- Sapphire X1800XL
- Thermaltake 680W WR00049
- NZXT Lexa Case.
It should be noted that the case use does feature a 120mm fan blowing into the case from the side window. This fan is rated @ 38CFM @ 1800rpm. Although this will affect the performance of the passive coolers slightly, it also helps remove the hot air from the active cooler, but please keep this in mind when perusing the numbers. All temperatures are in Celsius, and were measured using SYSTool. Heatsink
| Idle Temp. | CPU 100% load Temp. | CPU & GPU 100% load Temp. | Gaming Temp. | Open Case GPU & CPU 100% load | Extreme Spirit II | 32 | 32 | 37 | 40 | 38 | ZM-NBF47 | 30 | 34 | 36 | 37 | 42 | HR-05 | 29 | 29 | 35 | 35 | 37 | HR-05 SLI | 29 | 29 | 34 | 35 | 37 | Stock | 42 | 46 | 55 | 58 | 56 |
So, in the end, the Thermalright heat sinks win. I readily assume that the numbers I got could be lowered with the addition of a fan on them, as can be seen by the temperature increase when the case door, and the attached fan, were removed. We can also see that the active coolers seem to benefit from the case door being off, most presumably because of the airflow from front to back inside the case is already warmed by the hard drives, as well as the fresh air from outside coming from the door fan, when the case is closed.
Each of these coolers did a better job than the stock heat sink, which is no surprise. It’s not hard to find reports online about these exact fans dying, and ASUS readily replacing them. Which brings another thing to mind…the active coolers are hotter than the passive coolers…and for good reason. The chipset on this motherboard is located between the two PCI-E 16x slots…making it so that if you install a card in the uppermost slot, the coolers end up drawing hot air off of the video card. This problem is also increased when you have two cards installed…and could be good reason for the stock heat sink failing in the first place. Unfortunately I do not have any pictures to show this, as while I collected data I had a slew of hardware failures…including a good majority of my USB devices, and my camera. I apologize for those looking to see how these heat sinks mount…hopefully once I get a replacement camera I’ll be able to add such pictures here.
In closing, considering the mountings, the temperatures, and all the other factors, clearly Thermalright has done a fantastic job or bringing to market some of the best chipset coolers I have ever had my eyes on. From getting the packages in the mail, and seeing how well they are packed, right to the real stuff, and how they performed, met…no…exceeded any expectation I may have had. Whenever someone asks me on the forum about which cooler to choose, you can bet that it will be Thermalright I recommend …and no folks, none of these products were provided for free…each was paid for out of my own pocket!
|