Noise

For many enthusiasts upgrading cooling the goal is maximum stable overclock, and they will live with the inconvenience of a louder system. For other users silence is the most important factor, and these users will forgo maximum overclocking if that increases system noise levels.

There are very few power supplies that do not have a fan. While Zalman and a few others do make expensive fanless power supplies, we have not seen a fanless unit larger than 500W, or one that would be used for seriously overclocking a system. With that in mind the noise level of the system with all fans turned off except the PSU was measured. The power supply used for the cooling test bed is the OCZ PowerStream 520, which is one of the quieter high performance power supplies.

We have also measured the Corsair 620W and Mushkin 650W power supplies which are reported to be quieter than the OCZ. Both the Corsair and Mushkin are indeed quieter at idle or start up speed. However, as soon as load testing begins and the PSU fan speed kicks up the measured noise level is almost exactly the same as the OCZ PowerStream 520W.

We are currently in the process of reevaluating our cooler test bed and planning some updates. The new configuration will include a "quieter" power supply with variable speed and noise levels. We will also update to a P35 chipset motherboard with all passive cooling to the chipset. Changes to the test bed will appear as part of a future roundup with the motherboard, PSU, and CPU all upgraded to more current configurations. We are also investigating a change to a quad-core processor as a further challenge to CPU cooling.

The noise level of the power supply is 38.3 dB from 24" (61cm) and 47 dB from 6" (152mm). The measured noise level of the test room is 36.4 dB, which is a relatively quiet room with a noise floor slightly below that of the OCZ PowerStream 520 PSU.

Noise levels are measured with the test coolers with fans at both low and high speeds. Both measurements were taken at 6" and 24" above the cooling fan on an open case side. Results are then compared to the other coolers/fans tested in this category. Measured noise levels in this chart should be considered worst case. Measurements are taken with an open side of a mid tower case 6" and 24" from the HSF. Real world would be a completely closed case with a further reduction in noise. Please look for both low and high speed results in the noise graphs.

Noise Level - 6

Noise Level - 24


Water cooling enjoys a well-deserved reputation for silence. We were therefore expecting noise performance at least as good as the better air coolers. The Xigmatek AIO comes close to this performance expectation, staying near or below the system noise floor at both low and high speeds. This is particularly amazing considering the Xigmatek is based on a smaller, higher RPM 80mm fan.

The Evercool Silver Knight, however, has to be considered a disappointment in the area of noise. The Silver Knight uses a larger 92mm fan and a physically larger pump, but it's still quite a bit noisier than the Xigmatek. The measured noise levels at 6" and 24" should tell you to stay away from Silver Knight if you are looking for quiet. Both the fan and water pump are noisier than we would expect with a water system. Noise levels at the tested high fan speeds are much higher than the measured noise with the best air coolers. When turned down to low speed noise levels of the Evercool are below the noise floor, but performance drops below what is already a pretty mediocre performance curve. We expect better from a design as innovative as the Silver Knight.

Cooling at Stock Speed Scaling of Cooling Performance
Comments Locked

26 Comments

View All Comments

  • mikelimtw - Thursday, July 12, 2007 - link

    First off, I would like to thank everyone who has taken the time to weigh in on the product, be it positive or negative. We take user comments seriously because we care about what our customers or potential customers think.

    I started reading this site when it first came online and have been a regular reader since. However, I also know that people like ourselves make up the top 10 percent of PC users who really know all the ins and outs of PC hardware. As such, it is also sometimes too easy for us to become jaded about products such as our AIO-S80DP cooler; especially when they don't meet our very high standards.

    I will admit that this product won't necessarily appeal to the hardcore user because it does not possess the highest performance. However, I would like to present another point of view for your consideration; that is, the point of view from most average PC users who do not possess the skills or knowledge that many of us take for granted.

    Xigmatek had a few considerations when creating this product:

    1) Ease of use
    2) Safety
    3) Performance
    4) Noise
    5) Weight

    Ease of Use:

    We realized that traditional water cooling isn't meant for everyone despite the performance of such systems. I've read more than enough horror stories about inexperienced users putting together a water cooling system only to accidentally fry their PCs due to leaks.

    The Xigmatek AIO-S80DP was designed to be easy to use and to bring the benefits of water cooling to the masses. The design integrates the best features of traditional air coolers and water cooling systems while providing reasonable performance. It mounts like a traditional air cooler but provides the near silent operation advantage of water coolers. Additionally, you can use our product for 2 years without the need for a refill - try that with a traditional water cooling system.


    Safety:

    Unlike traditional water cooling systems where you are required to cut tubing, connect tubing to water blocks, pressure test for leaks, etc. etc., the all-in-one design assures that there are no leaks when using this product. This built in safety factor is what allows us to expand the user base. We can now include people who like the advantages of water cooling, but don't have the technical savvy to pull off assembling such a system.

    Another added safety feature comes from the unique design and integration of the radiators to the resevoir. There is no tubing inside the radiators because they cut down the flow of water. We have implemented a unique free flow system that allows a massive 72 liters per hour of water circulation. The water continues to passive circulate in the cooler after the system (and therefore the pump) is shut down. This helps to continue to quickly cool your CPU thus extending its service life. This is also the reason why our 80mm design handily beats Evercool's 92mm unit in most performance metrics.

    Performance:

    Again, it was shown in the review that this product was not the top performer. However, as noted by another user, it is well within the top 10 performers out of the extensive list of coolers tested. It's true that we gave up some performance to the top air coolers - but not much. The difference was just over 100MHz of core clock speed. I consider this a reasonable tradeoff for providing the ease of use and safety factors associated with this design - keeping in mind that hardcore users may not be the main target audience for this product.

    Noise:

    This is a traditional strength of water cooling systems. While many of you may not mind the fan noise generated by top performing air coolers, I'd have to play Devil's Advocate and ask why water cooling systems exist in the first place. Water cooling exists not just for the performance but also for the silence. I think most PC users would prefer a quiet PC. Otherwise there's also no reason for ever larger air coolers using 120mm fans.

    While our AIO product is not silent it is reasonably quiet, despite the fact that is is based on an 80mm design. Again, this was a compromise for weight, size and pump noise. Overall, however, I think we did a good job on the noise side, and Wesley seems to agree.

    Weight:

    Nowadays, air coolers that are approaching 1Kg in weight are pretty commonplace, especially among the top air coolers which use a pure brute force approach to solving the thermal dilemma. By this I mean they use lots of large fins for surface area, lots of long copper heat-pipes and large fan units.

    However, many people either don't care or don't realize the mechanical stresses placed on motherboards with such heavy coolers hanging off of them. The 80mm design was implemented to minimize weight in an effort to not subject your valuable motherboard to undue mechanical stress. This has been a big consideration in our new air cooler designs as well.

    In summary, the key advantages that our flagship AIO water cooler brings to the table are:

    • Multi-platform support: P4-478, LGA775, AMD K8 754/939/940/AM2 mounting
    • Ease of installation: Unlike other water cooling kits, this mounts just like a regular air cooler without the hassle of cutting tubes, assembly, or risk of leaks
    • Great Performance: 150W-200W+ thermal dissipation, well suited to existing dual core and upcoming quad core CPUs
    • Extends CPU life: When the system is shut down, the unit continues to circulate coolant passively, continuing to quickly cool your CPU
    • High-reliability: Highly reliable ceramic pump and 12V DC motor
    • Easy to maintain: Coolant reservoir only needs to be refilled once every 2 years – no need to constantly check water levels of an external tank.

    Sincerely,
    Michael Lim
    Sales Director
    Xigmatek Co., Ltd.
  • Wesley Fink - Wednesday, July 11, 2007 - link

    Xigmatek has provided updated updated company and product information as well as MSRP and likely street prices for the Xigmatek AIO. p.3 has been updated to include this new information.
  • Justin Case - Tuesday, July 10, 2007 - link

    This has to be one of the dumbest concepts ever. The idea of water cooling is to reduce the number of fans by using a single large radiator and a very big, slow and silent fan (or two, at most).

    By reducing the size of the radiator, you end up needing a fan as large as a normal air cooler's, and everything else (water, pump, etc.), is just making the cooler _less_ efficient than a well-designed heatpipe or plain copper-fin cooler.

    In this house we obey the second law of thermodynamics. Unfortunately, in some houses, the laws of marketing carry more weight...

  • bigpow - Tuesday, July 10, 2007 - link

    I'm keeping my Scythe Ninja B -stock-
    Good nuf for my modest O/C & ears
  • takumsawsherman - Tuesday, July 10, 2007 - link

    Now that we finished expressing how unimpressive these are, when are you gonna get a review sample of the IFX-14? I eagerly await your results.
  • ajramos - Wednesday, August 22, 2007 - link

    I too am eagerly awaiting a review of the Thermalright IFX-14 cooler. Working on a new QX6850 set up and need to pick up a cooler.

    Thanks for the hard work on the reviews...

    ajramos
  • erikejw - Tuesday, July 10, 2007 - link

    When I look at the performance of a 3.83 GHz overclock it is rated 6th of the 20-25 collers with 56c. All the others are running on maximum fanspeed all the time.

    Since the Xigmatek keeps a lower speed when it is not necessarry to work at maximum efficiency it will have higher rating in idle and at lower speeds.

    I rather have a silent system than a system that is superefficient in idle mode.
    To have a fair performance it would be great if you added temps when the Xigmatek fan is maxed in all scenarios.

    Anyway Xigmatek is not a stellar performer but the idea is interesting.

    Another note when I'm at it. I like to build as silent systems as possible and since your system base dB is high it is impossible to know which coolers is most silent since
    50% of them are moresilent than your system.

    It should be easy to add a category where you just plug the fan in a 12V supply out of a case and make the same 6" and 24" away measurements.

    This cpu cooler series is awesome, keep up the good work guys.

  • Frumious1 - Tuesday, July 10, 2007 - link

    Okay, I have to get this off my chest. Sorry. Please bear with me, because I do find the cooler tests somewhat useful, but they could be so much more!

    First, the text is about what - 80% filler BS that gets repeated every time? Why not just do an article saying, "this is how we test coolers" and refer back to it for new readers? Cut out all the crap, which would allow space for....

    First, screw the X6800 overclocking! I know, you love having that $1000 CPU (and will probably upgrade to the QX6800 for the new testbed?), but let's be honest: how many AnandTech readers have such high-end CPUs? I know I don't!

    Actually, scratch that last comment somewhat. What I really want is MORE. There is a market for X6800 OC testing. There is also a market for E4300, E6600, and Q6600 (and hell, QX6800) overclocking! I bet for every X6800 OCer you have reading AnandTech, there are probable 100 people with something in the E4300 to E6600 range (maybe even 1000).

    This is really important, because the OC results you're getting with your X6800 may not really apply to an E4300. Who buys a $120 CPU and pairs it with a $75 HSF, though? Well, me for one, or at most a $200 CPU. But temperatures seem to have a bigger impact on lower end CPUs, and where you go from 3.73GHz retail HSF to a maximum of 3.94GHz on air cooling with the Ultra-120 eXtreme, I would bet you could see a 400-600 MHz difference on some other CPUs.

    Now, while we're at it, others bring up AM2 occassionally. Seriously, get some AM2 tests in there. I don't care if it's a minority platform; get at least one decent AM2 CPU into your cooling tests and we could all have one central location to get quality results. The Barcelona chips (er... Phenom I suppose) at the very least better be included when they become available.

    I know, it's a lot more work for you. Let's be real, though: how long does it REALLY take to do all the tests for a single HSF? If we're going to get all these cooler reviews one at a time (or in rare cases two at a time), can't we at least get some more details on the real performance? Still, to keep it manageable, how about two CPUs, one high-end and one midrange? An E6600 would be a good choice now, although I'm very curious to see Q6600 as well (especially post-price-cuts). Then at least one AM2 CPU for comparison (which at this point would have to be midrange, given the most expensive isn't much more than $200).

    Maybe you don't even need to do this for every HSF article, but at least one good comparison of say five or so HSFs with more than just the X6800 would be GREATLY appreciated. Let us see exactly how the results you're getting on the extreme end of the price spectrum might translate down to us mere mortals!

    Cheers!
  • Wesley Fink - Tuesday, July 10, 2007 - link

    Every time we do not repeat test procedures in the review we spend half the time in comments answering questions about how we test. We tried it, and it didn't work, because most readers will NOT go back to check how we test. Take a look back at the first Ultra-120 eXtreme review where we just reported results and referred to the original review for test methods. Almost all the comments are about how we tested which was clearly spelled out in the referenced review.

    Second, the format is the same, but the repeat material is basically the Test Configuration page. We do have a paragraph or two that is basically test procedures repeat in other pages, but we always update every page with the data from the current review. Comments and analysis are also updated on every page, in addition to charts and graphs, to reflect the coolers or coolers being tested.

    we have talked much more about AM2 fit in the recent reviews, and that should be obvious. Since overclocking is still very poor on current AMD processors (compared to C2D or C2Q) testing cooling capabilities is not particularly challenging or revealing of cooler performance with AMD. That is why many top AM2 coolers like the Zalmans did very well on AM2 test platforms, and not so hot on Core 2 Duo.

    Last, we use the X6800 not because it costs $1000, but because it has unlocked CPU ratios. This allows much more flexibility in testing. The new test bed will likely use a 1066 Quad Core like the Q6600 on a P35 motherboard, and we will use the 1066-1333 options to make up for the lack of adjustable ratios. With the new pricing the Q6600 will be $266 at 9x1066 or 2.4GHz. A simple setting to 1333 gives 3.0Ghz at default voltage in our experience. That will match the fastest new 1333 CPU. OC will then push the FSB up from a 1333 setting.
  • Frumious1 - Tuesday, July 10, 2007 - link

    You put way too much emphasis on the unlocked ratios. An E6600 with decent DDR2 RAM can easily hit maximum CPU overclock without running into problems using a 1:1 ratio. I at least never worry about unlocked ratios, unless the mobo for some reason doesn't support FSB adjustments (which wouldn't be the case on any true OCing mobo like the EVGA you use). Your mobo reviews always have FSB OC results, often pushing beyond 500 MHz at reduced multipliers. I doubt an E6600 is going to come anywhere near 500 FSB with the stock 9X multiplier.

    But then, I'm only running an OCed E4300. That seems to top out at 3.0 to 3.3 GHz, but that's with a Thermalright SI-128; wish I would have purchased the Ultra-120 instead, now! I don't have the parts or the ability to run all the same tests you do. Which is why I'm asking about this stuff - you know, would that better HSF have improved my OC from 3.0GHz to something more?

    Second, even if OCing isn't great on AM2, there's still a big question of temperatures and performance relative to the stock cooler. I doubt 6000+ will reach much beyond the rated 3.0GHz, but I the better HSFs have to cool such chips a lot more than the stock. Noise levels would also be good to know for the stock AMD coolers.

    What I'm getting at is that you did an AM2 vs. Core 2 memory comparison a while ago showing that memory timings were typically about the same between the platforms (what ever happened to RAM reviews?), so that established a baseline where you could claim that whatever timings and voltages you could get with Core 2 Duo would usually translate over to AM2 without difficulty. We have never seen any indication of how AM2 performs. I realize to someone like you AM2 may be dead right now, but some of us still have such systems - or the occasional fanatic that refuses to buy Core 2 - and it would be nice to know how your results with an X6800 on temps compare to AM2.

    Let's say one article, looking at the Thermalright Ultra-120 eXtreme and a stock AMD 6000+. If you want to put in a bit more effort, throw in the Tuniq 120 and a Scythe Ninja Rev. B. It's not a perfect solution in terms of getting us information we'd like to see, but at least we would have a better reference point. At stock load, the temperature delta between the Ultra eXtreme and the Intel Retail is a whopping 24C, and at the max stock HSF OC it's 28C.

    What's the delta between retail AMD and the TR 120X? What's the noise level of a stock AMD vs. the TR 120X at load with a 6000+? For that matter, what's the power draw of a 6000+ vs. an X6800 - somewhat unrelated but still useful information. I don't know - at least not without looking at some competing websites - and I bet you don't either. I'd say that's a pretty useful piece of information, though!

    Oh, while I'm going on, would it be too much to ask for some power draw figures for your test results? I'm not talking about on every review, but I'm very curious about how much the X6800 uses at 2.93 GHz, 3.75, 3.83, 3.90, etc. I'd assume you're ramping up voltages as the overclocks increase, so it would be interesting to see what's happening in power requirements. Could be very good information for OCers looking at a new power supply.

Log in

Don't have an account? Sign up now