Review: Zalman ZM-80 VGA Heatpipe Cooler

Cooling
Viewing page 2 of 2 pages. Previous 1 2

TESTING

In order to test the performance of the Zalman ZM80 I used the hardware components below:

  • Asus P4S8X v1003a motherboard
  • Pentium 4-2.53B CPU
  • 256M Corsair PC3500 RAM
  • Triplex GeForce4 Ti4200 64mb (3.3ns default) VGA
  • Windows 2000 SP3
  • Vantec 470W Stealth PSU

I struggled long and hard to determine which particular piece of software would REALLY heat up my Ti4200 card. After trying dozens of benchmarks, demos, and everything I could find I settled on continuously looping (without titles) the Point Sprite test in 3D Mark2001 SE.

I used an Extech dual channel K-type thermocouple to monitor and record both the ambient and GPU temps. The thermocouple probe is mounted to the small step between the GPU core and its mounting pad (where the red arrow is). The probe is touching the side (not the top) of the CPU core.

I ran five series of tests to determine the performance of the stock heat sink and fan assembly and the Zalman ZM80 matched with various fan positioning and completely fanless.

  1. The reference is the stock HSF.
  2. In the test without fans, the ONLY fans in the system were the fan mounted to the Pentium 4 heat sink and the fans in the power supply.
  3. In the test with normal case fans I used 2 80mm YS Tech rheostat adjustable fans set to produce LESS NOISE (by my subjective judgment) than the CPU heat sink fan. One fan was mounted as a rear exhaust and the other as a front intake.
  4. In the last series of tests I used a Zalman BR123 mounting bracket to suspend an 80mm Enermax thermistor adjustable fan above the video card. One test had the thermistor positioned between the rear heat sink bracket and the back of the video card. On the last test I positioned the thermistor in the ambient environment to keep the fan as slow as possible.

Temperatures were measured continuously and recorded every 30 seconds. The chart and table below shows the temperatures measured at various times into the test loops.

Time
0 (idle)
0.5
1
2
5
10
20
Stock HSF
33.6C
34.8
35.5
36.0
37.3
37.3
37.3
ZM80 no fan
46.6C
48.8
49.2
50.2
52.6
54.5
57.9
ZM80 w/ case fans
37.9C
39.1
39.8
40.9
42.6
42.9
43.4
ZM80 fan above
36.7C
38.1
38.9
39.8
39.9
39.6
39.9
ZM80 slow fan above
37.4C
38.5
39.1
40.0
41.0
41.6
41.6

Note: Ambient room temperature during testing was 22.8C.

OBSERVATIONS & CONCLUSIONS

In a system with no airflow, the temps reached are a little high for my personal taste. The card doesn't specifically "misbehave" (no crashes, no tearing, no hangs, etc.), but there are noticeably more white pixels in the background while running the Point Sprite test, than with the other 2 test conditions. This condition disappears when even a small amount of airflow from case or direct fans is introduced. It's not a terrible malfunction, but an anomaly I noticed nonetheless.

I wasn't able to find thermal ratings for Nvidia's GPUs anywhere so I can't offer an opinion if 58C is officially too high or not. Clearly, in a system with a marginal amount of airflow the performance of the ZM80 is quite respectable.

Using the ZM80 along with a fan and the Zalman BR123 fan bracket produced the best results. However, the space considerations for this type of installation are important.

As you can see in the picture, the fan bracket cannot be moved to the optimal position because of the height and width of the ZM80. I couldn't center the 80mm fan above the video card because the bracket could move no closer to the card. In addition, if you choose to use the BR123 bracket you will likely lose the ability to install a full height PCI card below it.

(Editor's Note: The use of Zalman's BR165 fan bracket, which is designed specifically for use over a VGA card, would probably make this fan installation more optimal. BR123 is designed to allow a second fan to be attached over the VGA card when the BR123 is used with a fan over the CPU/HS. The photo shows the BR123 directly over the VGA, which is not the intended mounting method. This probably explains Brett's difficulty in centering the fan over the card.)

While the installation instructions are easy to follow, the actual installation can be a little difficult when it comes to lining everything up properly. There are simply too many degrees of freedom of movement, and positioning all of the parts correctly can be troublesome. There are numerous examples of both what TO DO and what NOT TO DO in the manual, but until all of the parts are together it's not clear how things are supposed to be on your individual card. The included Philips screwdriver in the kit is a nice touch though.

(Editor's note: The wide mounting range with many parts helps to explain why there are conflicting reports about the ZM-80 from both reviewers and users. It may be difficult to get exactly the same results twice even with the same parts because of the wide variance potential. This is a challenge I experience even with CPU HS that are so much simpler. The mechanical complexity of the ZM-80, presumably the result of trying to achieve usability with a wide variety of cards, has to be considered its weakest point.)

I was surprised that the back bracket and heat sink do not actually touch the back of the video card when mounted. They stand ~8mm off of the card and are mounted to bent metal arms that attach the bracket to the front bracket mounting screws. This arrangement is designed to insure that the back portion of the assembly is the cooler portion. This is fine for cooling the GPU, but it's not so good for trying to keep the rest of the card cool, especially in a tower style case. The back assembly acts as a sort of insulator for the back of the board and heat gets trapped between the card and the back assembly. The same issue also holds true for the components on the front of the card that sit under the heat sink. In a tower style case, the ZM80 is a GPU cooler, and less of a video card cooler.

There are quite marked differences in the the two supplied heat sink brackets. The one I had to use was quite tall and narrow, having the heat pipe sitting about 20mm above the GPU core. The other bracket is much shorter and wider with the heat pipe sitting around only 8mm above the core. This dimensional difference certainly has an effect on the cooling performance of the entire assembly since the length of the temperature gradient (difference between the temp of the bottom of the bracket and the top of the bracket) is quite a bit longer in one vs. the other. Although I couldn't test this I am sure that the cooling performance would be different.

Once installed, the weight of the entire card and heat sink assembly is pretty significant -- around three times the weight of the card alone. Moving your system with the video card installed would, in my opinion, present a potential for damage. In addition, the assembled size of the finished product will preclude use of the PCI slot directly adjacent to your AGP slot. If you are using this cooler with a PCI video card, you will likely lose the PCI slots on BOTH sides of the video card.

As the name suggests, SPCR readers are interested in SILENT PCs:

Is this an effective "silent cooler"? In the literal sense, I have to say no. I personally would not use this cooler on my Ti4200 card in a system with only passive cooling or really low airflow. The temperatures that I measured are just too high for my comfort.

Can this cooler be used to keep the GPU cool in a system with quiet, controlled but reasonable airflow? Certainly!

Can this cooler be used in a gaming intensive, overclocked system? A qualified yes... as above, I think using the ZM80 is effective where there is some airflow. (Editor's note: The 80mm fan over the card and HS with a bracket mount appears to be the best compromise of performance and low noise.)

PROS

- Very nice build quality
- Good manual
- No additional parts required -- everything is included, even spare parts!
- Acceptable performance in a system with some airflow
- Nice green handled Philips screwdriver included

CONS

- Alignment and proper installation difficult
- Excessive space lost due to significant size
- Potential damage during moving due to excessive weight
- Not really an effective fanless cooling solution for GF4200+ video cards

* * *

Our great thanks to online retailers Exotic PC and Silicon Acoustics for their generous contribution of review samples.

* * *

POSTSCRIPT - Jan 5/03 - by Mike Chin, Editor

After publication of the review, email exchanges with readers and discussions in the SPCR forums prompted the question:

How the ZM-80 compare with the original HS when both are cooled by a perpendicular 80mm fan over the card?

The original test data above shows what the ZM-80 does with the slow 80mm Enermax fan, but not how it compares with the original HS with the same cooling fan. So Brett ran this test with the original HSF back on, but with the original fan turned off, just to see whether the ZM-80 does provide a significant cooling advantage.

Time
0 (idle)
0.5
1
2
5
10
20
ZM80, slow fan above
37.4C
38.5
39.1
40.0
41.0
41.6
41.6
Stock HS, slow fan above
42.6
44.2
45.7
47.2
47.4
47.4
47.3

The answer is, yes, the ZM-80 does provide a significant cooling advantage. But it is surprising how decently even the small stock HS cools with a bit of airflow. The orginal HSF is amazingly efficient at cooling, louder and whinier though it may be in comparison to the slow 80mm fan.

Although it is not really an effective fanless cooling solution for today's hot VGA cards, with a slow-spinning perpendicular 80mm (or larger) fan, the ZM-80 is probably the only viable quiet cooling solution for hard core gamers.

* * *

Discuss this article in our Forums.



Previous 1 2

Cooling - Article Index
Help support this site, buy from one of our affiliate retailers!
Search: