The Asetek Low-Cost Liquid Cooler

First off, this is not a product review. This is a narrative of my experiences with the Asetek Low Cost Liquid Cooling CPU cooler. The main difference between this narrative and a product review is that no one could possibly replicate the exact scenarios that I am going to tell you about, just like you cannot replicate a lab’s scenario in everyday life. I am going to tell you about my experiences with it, and hopefully you will take away a lot more than charts and graphs would give you.

The first time I read about the Asetek LCLC was while skimming my first review of the HP Blackbird 002. For anyone under a rock for the past year or so, the Blackbird is the first child of the HP – Voodoo PC marriage. Complete with Voodoo DNA, this HP computer is actually a really neat package. It packs water cooled GPUs and CPUs into an “available for retail” box. By utilizing AseTek’s LCLC, we were told that 4.0 ghz overclocks on Core 2 Quads were possible, an amazing feat for any cooler, let alone one that was supposed to cost less than $100 in quantity.

Fast forward a few months to August of 2008. Chris, a consultant working for Asetek, shoots me an email saying “Hey, we’ve got this great thing you have to try.” I remembered back to the LCLC write-ups around the time of the Blackbird launch, and I was excited to try it out, so I said “yes, please.” I won’t lie, owning a computer business really does have its perks for a hardware nut like yours truly.

So three days later, a very nondescript box shows up from UPS. This thing is designed for OEMs, so no fancy box to show off “Insanely Good Thermals” or “Ridiculous Overclocks,” which I really appreciated. There’s only so much of that BS you can take in a day. Here’s a picture of what was inside:

Please pardon the very unprofessional photo of the parts on a white sheet. I do what I can...

<!--pagebreak-->

After ripping open the box and gleefully inspecting the contents, I hastily removed my Thermalright Ultra 120 Extreme and began to install the LCLC. For the record, here's a picture of the final installation:

Also for the record, here are the specs on the test box:

• Lian-Li A-05B with 3 SilenX IXP-74-14B fans (extra side blower)
• Core 2 Duo e8400 (3.7ghz, DDR2-1000, 1.45v)
• Asus P5N-D NVIDIA 750i Motherboard
• Corsair XMS2 DDR2-800 2x2GB
• XFX 8800GT video card with Zalman VF-900
• Creative Soundblaster X-Fi Fata1ity
• Hauppauge HVR1600 TV Tuner
• WD Raptor 150GB
• Seagate 7200.10 320GB
• Seagate 7200.10 320GB
• Lite-On DVD Burner and DVD-ROM
• Mitsumi Floppy/Card Reader

The only thing I could add to this machine is a second 8800GT, making it an excellent torture test for this cooler. The temperature inside this case must hit 40c when I'm really pushing it. I'm not really sure, but hey, not a review. Here's the other thing that makes this not a review but a story, this machine is my daily driver in the living room. That means that at any time a recording could kick on from Windows Media Center, who-knows-what kind of network activity is going on, and more so, I'm going to be listening to iTunes during the whole test.

Read on for the actual testing...

<!--pagebreak-->

I will start the actual "what happened" with a few notes about the cooler itself. First, the fan header is only a 3-pin which really doesn't matter. The pump is so quiet you'll never need to slow it down and it still reports the RPMs just fine with only 3 pins (which tends to read around 1300rpm for the record). Second, the radiator (aka Heat Exchanger) showed up just a little bruised. I seriously doubt it is anything that could effect the performance, but it should be mentioned any ways. Lastly, this kit showed up with only socket 775 installation brackets, although a retail package would have both AM2 and 775. Curiously, it did ship with two brackets, a push-pin model and a bolt-through model. After our first incident shipping a machine with a Scythe Infinity and having half of the push-pins snap, I'm very glad to see bolts. In fact, I have turned away some very good coolers because of push-pin-style installation brackets, they just don't hold up. All of this testing is with the bolt-through adaptor.

Any ways, in my haste to install this wonder of modern computing, I neglected to run any real temperature testing on the TRUE that this LCLC replaced, but the initial readings were pretty decent. I installed the radiator behind my existing 120mm SilenX fan, so that the fan was pushing the exhaust air over the exchanger. I fired up 64-bit Prime95 and started checking. I saw load temps at 60c after 5 minutes of Prime95 load, and just 62c after 20 minutes. I then let it cool down, and measured the temps after 5 minutes and 20 minutes. They were 41c and 40c respectively. I didn't really think that this was the best this setup could do, so I decided to rip it all out and try another configuration.

Test two had the LCLC setup as I had seen it in both the Blackbird and a recent Anandtech review of Cyberpower's big dog. I installed the fan on the back of the radiator, and installed that fan on the case. I used Arctic Silver 5 instead of the stuff they ship it with between the copper and the CPU. I also really wanted to see this thing rock, so I stretched some electrical tape over any gaps between the radiator and the fan shroud. This made sure that all the air possible was going over the fins. Again, we were exhausting through this radiator, which is how I imagined it should be setup. I started up Prime95 and saw 60c after 5 minutes and 62c after 20 minutes, exactly the same as before. Cooling down showed the same, 41c and 40c after 5 and 20 minutes. I scratched my head, went to sleep, and called Chris in the morning with my results.

Read on for day two...

<!--pagebreak-->

After talking with an engineer at Asetek, I decided to run the same tests with the Thermalright cooler to see what was going wrong. I removed the LCLC and reinstalled the TRUE, complete with SilenX fan attached. I started up the Prime95 and got to 59c after 5 minutes and 60c after 20 minutes. The cool down was also the same, 41c and 40c after 5 and 20 minutes. Yes this is 2c lower at load than the testing from day one with the LCLC, but don't read into that too much. Remember, this is not a scientific endeavor and there could've easily been ambient temperature differences that cause that. You should walk away with the idea that the TRUE and the LCLC were nearly identical performers in this installation, which goes to show you that no matter how efficient a cooler is, it can't get temperatures lower than the air that is used to remove the heat.

I really wanted to see this thing rock though, so I devised a way to get the air colder around the radiator: turn the rear fan to intake and switch the side fan to exhaust. While I knew this would cause the northbridge and GPU to heat up, it would give me a peak at what we can do with this thing. Since the motherboard temperature would be pretty important in this test, I started off at stock settings to get a baseline for the MB temp. It was around 40c under full load with the TRUE, with the CPU at 52c. I then overclocked back up to 3.7ghz and loaded the machine again. This time the MB temps hit 42c, and again the CPU was around 60c. I didn't see much of a change in the MB temp, despite nearly a 10c change in the CPU temp. Lastly, I installed the LCLC in it's new cold-air mode, and loaded the cores. The motherboard hit a staggering 46c. This is a 4 degree change, double the 2 degree change between overclocked and not. The GPU temperature was also up about 6 or 7 degrees. Here's the kicker though: CPU under full load was a chilly 54c, just a hair above stock settings with the TRUE.

After seeing this huge drop in temps, I decided to see if that magical 4.0ghz mark was now within reach. Unfortunately, I went all the way to 1.55v in the BIOS, 1.48v or so measured by PC Probe, but to no avail. This processor and this board must just hate each other, as almost every other e8400 I've seen has hit 4.0ghz with no trouble at all. So at the end of the day, I had a much cooler-running processor, somewhat hotter motherboard and hotter GPU.

Read on for the conclusion...

<!--pagebreak-->

Interestingly enough, this experience reminded me of a conversation I had with a client two years ago about the difference between water-cooling and air-cooling a computer. He was a big fan of Mario's Overdrive PC down in Atlanta, and had spoken with Mario several times about the specs of his machine before he ordered it. He was moving out of a Voodoo PC Omen, and expected never to own an air-cooled PC again, however Mario explained to him the real advantages and disadvantages of each setup. While water may bring some pretty neat bling to your case, the fact is that air-cooled machines regulate the temperature of the whole machine better than water-cooled ones. In addition, even the best water cooling loops are still unable to cool the water below ambient temperature. That means that in order to get the best results, you have to blow the heat back into the case, making your components even hotter. And as for how quiet a machine is? Well, in order to maintain your motherboard temperature you'll more than likely install a small fan on your northbridge to keep it cool, which is louder than three good 120mm fans any day. Throw in the hazards of leaks and pump failures, and all of a sudden the added expense of water doesn't look so cool.

I won't say that I totally disagree with using this product, because it really does shine when it is given the right environment. If you have a very large case, or want to have the radiator outside of the machine, something like this can easily bring down your CPU temp without breaking the bank. My buddies over at FrozenCPU.com used to have a rebadged version of this called the NorthQ Siberian Tiger, but it has dried up recently. Still, finding one of these in retail drape should not be hard, and you should be able to get one for around $100. Really, that's not that expensive compared to a Thermalright Ultra-120 Extreme with a nice fan like the SilenX ones that we use. I still don't think you'll see this product showing up in a Nordic PC machine any time soon. Sorry Chris, it's a great little water cooler, but I'm just not sold on water yet.