Zalman ZM80A-HP VGA Heatpipe Cooler Review

Admin - 2007-02-08 17:18:48 in VGA Cooling
Category: VGA Cooling
Reviewed by: Admin   
Reviewed on: January 7, 2003
Price: $40 USD


Back in November, Zalman released a new version of the ZM80-HP VGA cooler, the ZM80A-HP. Both models come from the CNPS (Computer Noise Prevention System), which uses shorter heat conductance paths, and surface areas 3-10 times larger than traditional coolers. That's right, no fan, and no noise. So how well does the ZM80A-HP perform? Let's take a look....


  Weight 400g or 0.88lbs
  Dissip. Material Aluminum
  Dissip. Area 1300 cm² or 1.39 ft²
  Heatpipe Gold Plated Copper Tube
  Compatibility Useable with VGA cards having holes around the chipset


Important Notes
Two important things to note about this heat sink. First, is the weight, at 400g, the video card equipped with the ZM80A-HP should be tightly attached to the PC case. And special care should be taken when moving or transporting the case. It'd suck to drive 18 hours to a LAN party, only to arrive and find your video card snapped in two.

The next thing to note is that the ZM80A-HP will not work with the Matrox Parhelia. Looks like the Matrox owners get the shaft again.

Closer Look

The ZM80A-HP comes with everything you need to install the heat sink, and then some. Included in the package was:


For most people, the hardest thing in installing a VGA heat sink is the first step. Removing the old heat sink. Luckily for me, the VisionTek GF4 card that I use has a heat sink that is clipped in, and removes very easily.

After you have your old heat sink removed, it's time to clean off the old thermal paste. Some people may be able to remove the paste with a cloth, while others like me, may have to use a razor blade.. I also used rubbing alcohol to help remove the group. Some people don't like to do this, so do whatever you feel comfortable with. Just be careful so you don't damage your GPU.

After you've removed the old heat sink and thermal compound, it's time to begin installing the ZM80A-HP. And that begins by reapplying some new thermal compound. :) Zalman provides thermal grease with the ZM80A-HP; however, I decided to use Artic Silver 3 instead.

Zalman has a note in the manual that states, "Most GPU's on the video card are concave the curvature of which is directed toward the center. Thus, more thermal grease is required at the center. Also, in the case of the ATI Radeon 9700 Pro, the protruding graphics core is surrounded by a protective frame, which is higher than the core itself thus generous amounts of thermal grease must be applied to the core." You could do this, or you could lap your GPU and make it flat. :)

If you are unsure if your GPU is flat or not, you can apply the thermal grease, place the heat sink on it, and then remove the heat sink. If the compound is even on the heat sink, the compound was applied properly.

After applying the compound, the front and rear base assemblies must be attached. Two sets of assemblies are included with the ZM80A-HP. One large set for cards based on the GF4 Ti where the mounting holes are far apart, and the other set for cards where the holes are closer together.

The base assemblies are labeled "front" or "back" to indicate which side of the card they are attached to, and are mounted with the arrow pointing toward the AGP slot of the card.

More thermal compound is applied now. This time, to the grove that runs the length of the mounting base assemblies, and to the base that will make contact to the large heat sinks.

Now the heat pipe is installed. It fits in the grove where we just applied the thermal compound, like so.

As you see, the heatpipe is angled slightly.

Just when you thought you we finished applying compound, we get to apply some more. This time, to the back of the heat sinks, which we will mount to the base assemblies.

The heat sinks mount to the base assemblies via four screws. And with that, the heat sink is installed.

You may have already guessed it. The ZM80A-HP does take up the PC slot next to your AGP port. Not a problem for most people, unless you are one of those weirdo's with 6 PCI devices. :P

How a Heatpipe Works
Before I actually move into the testing, I'm going to explain the heat pipe, so that those who don't know what one is will (hopefully) have a better understanding of it, and why it is used.

Zalman uses a heat pipe design, that consist of a vacuum tight copper tube, a wick (woven fabric, fiber, or other material, that looks like a braid), and a fluid (trying to find out exactly what Zalman uses, most people use a mix of water and cooling solvent). As the front heat sink gets warmer, the heat pipe under the front plate gets warmer as well. This turns the liquid into a high pressure vapor that moves toward the cooler side of the heat pipe, as the vapor reaches the cooler plate, it condenses and releases the heat to the rear plate. The wick then helps move the condensed liquid back to the other end of the heat pipe where it can repeat the process.

Here you can see a cut away of the heat pipe, which exposes the wick. Also, in the picture below, you can see a larger portion of the wick. This is probably one of the more important pieces of the heatpipe.


Now that the heat sink is installed, and I've somewhat explained how the heat pipe works, it's time to fire up the computer and see how the ZM80A-HP works.

The ThermalTake GeForce 4 Copper Cooler is probably one of, if not the most commonly used VGA coolers available, and because of that, I decided to use it to compare the ZM80A-HP. I said earlier that Zalman is known for using surface areas 3-10 times larger than traditional coolers, and as you can see in the pictures, the two Zalman heat sink plates easily dwarf the TT cooler.

The test system that I used is as follows:

To make testing as fair as possible, I used Artic Silver 3 when installing both heat sinks. I also ran the computer for 80 hours to allow the compound time thicken, as recommended by Artic Silver, Inc. Temperature measurements were taken by a CompuNurse with a probe placed next to the GPU. Thermal compound was cleaned off and reapplied between the use of the TT GF4 cooler, and the ZM80A-HP.

Idle & Load Temp - Not OC'd
To measure the idle temperature, the computer was shut off for 20 minutes, then powered up, and sat at idle for 5 minutes. To measure the load temperature, the computer ran FutureMark's 3DMark2001 for 15 minutes. During this time, the highest temperature that reached was recorded.

Idle & Load Temp - OC'd 310/700
The OC'd idle temp was measured in the same way as the non-OC'd idle temp. As with the other measurements, the OC'd load temp was measured in the same way as the non-OC'd load temp.

As you can see in the results, the ZM80A-HP didn't keep the GPU as cool as the Thermaltake cooler. I expected this would happen, after all, the Thermaltake cooler uses a fan to cool off the heat sink. :) Though, I suppose this is where the Noise vs. Performance debate comes into play. The Thermaltake GF4 cooler has a noise level of 29dB(A), where the ZM80A-HP produces no noise at all.

Now that we've seen how much the temperature varies, how does that affect the video card's performance? We all know that IC (Integrated Circuits) perform better if they are kept cooler... just how much does a 5-7C difference make?

FutureMark's 3DMark2001SE Build #330 - Not OC'd
For "benchmark" testing under 3DMark2001, I closed all unnecessary system and user processes inside of Windows XP. Then ran 3DMark using the default testing options.

FutureMark's 3DMark2001SE Build #330 - OC'd 310/700
Same method was used to test the card when it was OC'd as I used when it wasn't OC'd. Same services were closed.

So far, the TT cooler leads the ZM80A-HP by 5-6C and 10-20 points in 3DMark. 10-20 Points really isn't something most users worry about, except maybe the hardcore, extreme, l33t, gamers that want to push it to the max.... Heck, I can't tell a difference when playing BattleField:1942 or UT2K3. Though, knowing that the card stays a few degrees cooler, would make me feel better. :)

Speaking of UT2K3, let's see what the max FPS looked like, shall we? To get the result for Max FPS, I used the Unreal Tournament 2003 Benchmarking tool and ran the Flyby-Phobos2 test running at 1600x1200.

Unreal Tournament 2003 - Flyby-Phobos2 Benchmark - Not OC'd

Unreal Tournament 2003 - Flyby-Phobos2 Benchmark - OC'd 310/700

As with the 3DMark test, the UT2K3 test shows the ZM80A-HP with a slightly less score. Is it something that you would notice in a fire fight? I didn't.


Well, there you have it. If you are looking for quiet, the ZM80A-HP cannot be beat. However, it does sacrifice a slight bit of performance by not having a fan to keep it cooled. Zalman states in the manual, "You can expect satisfactory cooling performance with the product alone. However, you can enhance the cooling capacity by using the product in conjunction with Zalman's CNPS6000-xxxx or CNPS6500B-xxxx, which is accompanied by the FB123 (fan bracket)."

I believe the ZM80A-HP would do extremely well with a fan added to it, as suggested by Zalman. A low noise fan would probably cool the ZM80A-HP well below the TT GF4 cooler, and something like the TT SMART Case Fan II or Vantec Tornado would work mush better. Though, this is just an assumption at the moment. I'll see if I can come up with a FB123, and test the ZM80A-HP with a few different fans. I'd really like to know what this thing can do with a 80mm Tornado. :) Look for an update to come in a few weeks.