Written By:
Date Posted: April 29, 2002
Global Win Coolers
How many of us haven't heard of Global Win? I don't think many if any have not heard of this company, in relation to cooling products that is. Any good computer store will usually contain at least some Global Win coolers for sale. Where and when was Global Win created? It began in 1992 in a small office space in Taiwan and has grown to over $10 million dollars in sales annually. Okay, that sounds idealistic but still we can't deny that Global Win is a product leader for heatsinks.
My first experience with Global Win coolers occurred when I needed to cool my overclocked K6-2 350 that was running at 500MHz at 2.6+volts. The only fans I could get in my city were generic fans that didn't cool the processor to under 50°C under load. So I ordered a FDP-32 and this big heatsink allowed me to cool this hot little processor.
So when I looked for a heatsink for my Duron processor I immediately looked and picked up a FOP-38. Since then I had bought a Dragon Orb 3 to cool my Thunderbird, but when asked if I wanted to review the CAK-II 38 I jumped at the chance to return to Global Win.
CAK-II 38
So what is the design behind the CAK-II? Well just like the CAK-38 before it, this heatsink is made out of one piece of copper, and as is standard fair for almost any 60mm overclocking heatsink it comes with a loud 7000RPM Delta fan. The major differences between the CAK-II and CAK are that the CAK-II is slightly shorter than the original, it also now has a professional looking shroud to hold the fan, in fact the whole heatsink looks to be designed much better with 32 fins compared to 23 from the original, and lastly but most importantly the clipping mechanism has been redesigned (more on that in installation). For a more detailed information on this cooler (directly from Global Win) click on the link.
The contents of the box are as follows:
- CAK-II38 Heatsink/Fan
- Global Win Case badge
- Installation manual (good idea to read this)
- 3-pin to 4-pin/RPM monitor (nice idea)
Installation
The installation of the CAK-II38 went very easily as the shroud/fan was already connected to the heatsink and the 3-pin to 4-pin/RPM monitor was also already connected. Apart from finding the one missing pin that holds the shroud on everything else was easy, remove the thermal pad, apply the AS3 and then put it on the CPU.
The clipping mechanism has really improved over the FOP series of heatsinks. With those heatsinks it was really necessary to use both a screwdriver and a pair of pliers to get the clip over the socket slugs. With the CAK-II this has been improved to just needing a flat head screwdriver to clip it to the socket slugs, as you can see from the picture in the previous pages' pictures. Of the heatsinks I have I'd rate it slightly behind the Dragon Orb 3 for installation due to the fact that while they are equally easy to install, the Dragon Orb 3 uses all six socket slugs while the CAK-II only uses two.
One very good thing about post installation of this heatsink is that it is fairly short, and unlike my DO3 it will allow me to place my motherboard tray back in my case (InWin Q500) without having to remove the fan.
Test Setup
All testing was done on the following three processor speeds:
CPU: |
AMD Athlon 1GHz @ 1.75v - 51.4 Watts |
AMD Athlon 1.2GHz @ 1.94v - 75.8 Watts |
Intel Celeron II 850 @ 1.75v - 23.5 Watts |
Motherboard: |
Abit KT7-RAID BIOS 64
|
Gigabyte 6OXET-C |
Memory: |
256MB PC-133 @ 2-2-2 |
Hard Drives: |
20GB Quantum LM, 40GB Maxtor
|
Sound Card: |
Sound Blaster X-Gamer 5.1 |
Operating System: |
Windows 2000 Professional Service Pack 2 |
Front Side Bus |
100MHz(200MHz DDR) |
100MHz FSB |
Video Cards: |
Hercules Prophet 4500 (Kyro II) |
Coolers: |
Thermal Take Dragon Orb 3 (7000RPM); Global Win FOP-38 (7000RPM); Global Win CAK-II 38 (7500RPM); Global Win CAK-II "32" (5500RPM) |
The Test systems/coolers looked like this:
The tests were run directly after bootup, and then the processors were allowed to cool for approximately 15 minutes. Sisoft Sandra 2002's burn in test was used with 23 repetitions of CPU Arithmetic and Multi-Media benchmarks being run. Two tests were done for each speed, as follows, 1GHz then 1.2GHz then 1GHz and lastly 1.2GHz. If the temperatures were off more than a degree between both tests a third test was run and the two closest temperatures were used.
The Athlon's used the socket thermometer to read the temperatures, while the Celeron used its more accurate internal thermal probe to read the temperature, both systems used MBM 5.1.0.2 to tell the temperatures from the CPU's. Room temperature was measured using a Radio Shack Indoor/Outdoor Thermometer with Dual Display and was never more than 0.2°C off stated room temperatures.
Sound levels were measured using a Radio Shack Digital Sound Level Meter from 1 meter away from the heatsink/fan. The only other fan running at that time was the Enermax power supply fan.
The Tests
Lets first look at how these coolers fare with a standard 51 watt load of the Athlon 1GHz, and note that all temperatures are sorted by C/W or as the equation used here is:
((Temperature under load - Room Temperature)/Wattage of the Processor):
Here we see the CAK-II beat the Dragon Orb but only by 1/2 of a degree, which is within the error range of Motherboard Monitor. The 'castrated' CAK-II performs the worst, but this is understandable as it is moving about 2/3 of the air that the other fans are yet it is silent to my ear. All coolers however do not let the temperature reach 40°C on any of the coolers.
What about overclocked results, from an Athlon 1.2GHz CPU that produces the same amount of heat as an XP 2000+:
We can see here that all these coolers could cool an XP 2000+ adequately. However we see the CAK-II 38 pull away from the rest of the coolers giving itself a 1.5°C lead over its closest competitors. The CAK-II "32" loses ground in this test as it gets 4°C hotter than any of the other coolers would at the same ambient temperature.
What therefore can we conclude about the performance of the CAK-II with fans? We see that it is highly dependant on moving allot of air quickly to help it cool itself, its C/W rating follows almost the exact same path as the airflow increase, with both the airflow increasing by 30% and the C/W rating decreasing by the same amount.
Fan Tests
Fan Sound Levels
So how loud is the Delta 7000RPM fan? The following chart shows some measurements from 1 meter away:
|
FOP-38 (Delta 7330 RPM) |
Dragon Orb 3 (7031 RPM) |
CAK-II 38 (Delta 7500 RPM) |
CAK-II "32" (5532 RPM) |
Sound Level (dB):
|
58 |
59 |
59 |
51 |
The sound results show this as all three high performance fans, two Deltas and an Evercool, are very close with no noticeable difference in loudness, though the Dragon Orb's fan had a slightly higher pitch to it. The slower CAK-II "32" fan was also the quietest, with the difference being like a car passing about 100 feet away at standard speeds (Delta Fans), compared with a you having a normal conversation with someone standing a couple of feet away (CAK-II "32" Fan). What would happen though if you had no fan on the CPU, and only had two case fans (60mm blowing air in, 80mm blowing air out) inside a full case?
Celeron 850 - No Fans
What happens if we remove the major source of noise, the fan. Well lets see how this works using a Celeron 850:
Now isn't that interesting, the Dragon Orb 3 destroys both other heatsinks, and the CAK-II is completely useless in this application, giving a temperature of 65°C from a 23.5 watt CPU. Why did this happen? Well we see the only differences between the CAK-II and the other two heatsinks is that the CAK-II is all copper while the FOP-38 is aluminum, and the DO3 is a combination of copper and aluminum. Why does aluminum perform better? We see if we look at our old physics textbooks that copper absorbs heat quicker than aluminum but it does not wan to conduct that heat away quickly. Aluminum however lets go of the heat it gathers much more quickly and has a better thermal conductivity.
Final Words
So what can we conclude about the CAK-II 38 from Global Win? As we can see by the results of the tests of the 1GHz Athlon and the 1.2GHz Athlon results, the CAK-II 38 gets better as the heat is increased. While it is in a virtual tie with both the Dragon Orb 3 and the FOP-38 at 1GHz, the CAK-II takes the lead by over a degree as the processor heats up. The results of the fanless Celeron 850 are somewhat surprising, as the copper CAK-II can't touch the performance of a aluminum heatsink or copper/aluminum heatsink in this test.
But how loud is the CAK-II 38 fan? Well it is quite loud as anyone who has previously used or heard the black label Delta fans. Using the quieter Thermaltake fan makes the heatsink almost silent but as would be expected it definitely impacts performance but is still able to keep the processor producing 70+ watts of heat below the magical 50C limit that many people like to see.
Would I recommend this heatsink? For the price that you can get it from you could pick up a Dragon Orb 3, but as we've seen the CAK-II 38 performs better as the heat increases and is just as loud as the Dragon Orb. The size of this heatsink is impressive as its almost 2.5cm shorter than the DO3, yet performs better. From a purely aesthetic view this heatsink looks very professional and looks like it was designed to do the best it can. The downsides? Its loud and I mean loud, once you've switched from a 7000RPM 60mm fan to a quieter fan you'll never want to go back. Other than that, the pins holding the shroud in place are pathetic (see the picture below), and don't seem to be able to last as I've had to push them in on more than one occasion as I feared they were going to fly off.
I would like to thank for making this review possible by sending the CAK-II 38 to me, so go check out their products, the water cooling systems look especially nice.
Good Points
- Excellent performance with a fan
- Clean, professional design
- Small cooler
- Good price
- 3 pin to 4 pin / RPM monitor
Bad Points
- Loud
- Pins holding shroud could be stronger
- Thermal Pad to be taken off the cooler
- Bad performance without a fan
|