Testing Methodology

For those that didn't catch our brief word on our updated testing methodology from our SilverStone FT03 review, I'll give you a quick primer. After our first two case reviews since 2009 went over with mixed reception, we went back to the drawing board and worked up a better, hopefully more streamlined approach to case testing. First, we've standardized the parts we're using for case testing. For Mini-ITX and Micro-ATX enclosures, we're using the testbed I described in the FT03 review. For full ATX enclosures (e.g. mid-towers like the BUC and larger full-tower cases), we've kicked things up a notch. Below is the kit we're testing with; you'll see some overlap in a few places from our Mini-ITX board where the same components could handle the same jobs:

Full ATX Test Configuration
CPU Intel Core i7-875K (95W TDP, tested at stock speed and overclocked to 3.8GHz @ 1.38V)
Motherboard ASUS P7P55D-E Pro
Graphics Card Zotac NVIDIA GeForce GTX 580 (244W TDP)
Memory 2x2GB Crucial Ballistix Smart Tracer DDR3-1600
Drives Samsung 5.25" BD-ROM/DVDRW Drive
Kingston SSDNow V+ 100 64GB SSD
Western Digital Caviar Black 1TB SATA 6Gbps
CPU Cooler Zalman CNPS9900 MAX with Cooler Master ThermalFusion 400
Power Supply SilverStone Strider Gold 750W 80 Plus Gold

For full ATX cases, we need to know not only how well they muffle sound but also how well they handle overclocked hardware. The Intel Core i7-875K we're using can run at an overclock nearly 900MHz faster than stock (and with a healthy dollop of voltage under it to boot). While going for maximum thermals would probably involve using an X58-based platform, LGA1155/1156 tend to be more popular and using that standard gives us the flexibility to potentially test air- and water-coolers down the line. It also bears mentioning that when we test under stock settings, the DDR3 runs at 1333MHz; it only goes up to 1600MHz when we're testing with the overclock in place.

My primary concern involves the BIOS-based fan controls. ASUS offers a decent amount of granularity in controlling the CPU fan, but I'd like to know what you think the ideal settings are. I currently have it set conservatively, to try and keep the processor below 60C, which is how I'd likely set it for my own system. But how would you set it? Would you just use ASUS's default "Silent" setting? Would you set a higher temperature threshold? Let us know.

Our actual testing procedures are unchanged from the Mini-ITX/Micro-ATX testbed, so here they are again:

Acoustic testing is standardized on a foot from the front of the case, using the Extech SL10 with an ambient noise floor of ~32dB. For reference, that's a silent apartment with nothing running, testing acoustics in the dead of night (usually between 1am and 3am). A lot of us sit about a foot away from our computers, so this should be a fairly accurate representation of the kind of noise the case generates, and it's close enough to get noise levels that should register above ambient.

Thermal testing is run with the computer having idled at the desktop for fifteen minutes, and again with the computer running both Furmark (where applicable) and Prime95 (less one thread when a GPU is being used) for fifteen minutes. I've found that leaving one thread open in Prime95 allows the processor to heat up enough while making sure Furmark isn't CPU-limited. We're using the thermal diodes included with the hardware to keep everything standardized, and ambient testing temperature is always between 71F and 74F. Processor temperatures reported are the average of the CPU cores.

And last but not least, it's important we thank the vendors who made our testbeds possible.

Thank You!

We have some thanks in order before we press on:

  • Thank you to Crucial for providing us with the Ballistix Smart Tracer memory we used to add memory thermals to our testing.
  • Thank you to Zalman for providing us with the CNPS9900 MAX heatsink and fan unit we used.
  • Thank you to Kingston for providing us with the SSDNow V+ 100 SSD.
  • Thank you to CyberPower for providing us with the Western Digital Caviar Black hard drive, Intel Core i7-875K processor, ASUS P7P55D-E Pro motherboard, and Samsung BD-ROM/DVD+/-RW drive.
  • And thank you to SilverStone for providing us with the optical drive and power supply.
Assembling the IN-WIN BUC Noise and Thermal Testing, Stock


View All Comments

  • Dustin Sklavos - Tuesday, May 10, 2011 - link

    My gaming drive alone is 585GB, and that's all legitimately purchased content.

    My scratch drive, where I put video I'm editing, has nearly a terabyte of video on it.

    My desktop has, on the whole, four hard drives (two in a RAID 1 and two in a RAID 0) and two SSDs (two in a RAID 0).

    So no, you don't have to steal content to fill that much space. You can get by just fine by making your own.
  • dagamer34 - Tuesday, May 10, 2011 - link

    You can never have enough hard drives for backups. Reply
  • FSWKU - Tuesday, May 10, 2011 - link

    Because anyone who needs a lot of space MUST be stealing content. That's the ONLY possible explanation. Lets see how much space I actually NEED:

    Windows Folder = 11.8gb

    Steam Folder = 17.3gb

    Music Folder = 31.3gb (all ripped from CDs that I own)

    Disc Images = 20.9gb (several flavors of Linux, Backtrack, and a few versions of Windows downloaded from my TechNet subscription to use in the VMs I develop on.)

    Video Folder = 50gb (all projects I've worked on or DVD's that I've backed up from my collection).

    Documents Folder = 20.4gb (Pictures, application files, training materials, savegames, misc drivers, voiceover projects, VHDs for the VMs I mentioned earlier)

    Thats 151.7gb, and I don't use my computer for nearly as much as most other people do. I keep most of my space free because I dabble in video editing. But this just means my 320gb (297gb once all is said and done) is just barely enough for what I do. But of course I MUST be stealing content if I feel like I need more space than the 80gb you limit yourself to...
  • Taft12 - Monday, May 09, 2011 - link

    <i>There's a lot you can do to mitigate it by using quiet fans and employing fan controls, but it just doesn't have a whole lot to really keep all that noise in and with a video card like the GeForce GTX 580 it might get a little louder than you'd like.</i>

    Fortunately, a PC case at this price point is unlikely to see a GTX580. A GTX 560 or 6870 is much more likely. This really sounds like the best option for a wide window of PC budgets and it sounds like you'd need to double the price you pay to get a case that is a definitive improvement.

    Great review Dustin!
  • fraginader - Monday, May 09, 2011 - link

    Would be nice to have a labelled diagram detailing various terms like 5.25" bay, SATA backplane etc. Newbies like me would appreciate it. Reply
  • micksh - Monday, May 09, 2011 - link

    One foot distance for noise measurement is actually too close. Towers are mostly standing on the floor so the distance to ears is rather 3 feet.

    This results noise from front fan contributing to overall system noise way more than it would contribute from 3 feet distance.

    One foot distance to front fan means two feet distance to video card. With doubling the distance noise pressure decreases 4 times.
    Let's imagine video card and front fan generate the same amount of noise. This method will result 6 decibel added to front fan noise comparing to video card noise.
    Whereas registering sound from 3 feet would make difference in distance negligible and amount of noise would be almost the same from both.

    Same thing happened in December in review of Silverstone GD04 case. Sound meter was too close to case fan and the case fan noise was artificially inflated in dBA measurement. And CPU fan noise was suppressed by difference in distance. This resulted incorrect conclusion that the case was loud. If you read Silentpcreview.com review of the same case they regarded GD04 case as quiet. But they measured it from 1 meter and they used very quiet components so their methodology should be more accurate.

    And I highly doubt that noise floor in silent apartment at night is 32dB. It's probably sound meter limit.
    Better sound meter is needed in order to get accurate readings.
  • Dustin Sklavos - Tuesday, May 10, 2011 - link

    Actually my tower is about two feet at most from my left ear. Some peoples' mileage may vary; I have a friend whose tower is on his desk. The fans in the BUC aren't the issue, and the system runs very quietly at idle from a subjective point of view. The BUC doesn't have much in the way of soundproofing, and that's why noise levels go through the roof when the GTX 580 gets stressed.

    32dB is actually pretty quiet. I'm sorry I don't live in a quieter area, but you'll also find the rated noise floors for most sound meters $300 and below is between 30 and 40dB.

    At the risk of being unprofessional, I'm getting tired of SPCR being brought in here as the bible by which all other reviews must be judged. The GD04, without proper fan control, is audible. SilverStone's engineers will even admit to that, and I think SilverStone tends to engineer their designs towards performance first. Then, when you employ fan control, you find that their cases are so well-engineered that they're able to handle considerable thermal stress without requiring substantial active cooling.

    As I told Tony at SilverStone when I met him at CES, I bought the GD04 with my own money, of my own volition, after having explicitly researched the case including your vaunted SPCR review. It was audible even across the room, where I had it stationed next to my television. It NEEDS fan control.
  • micksh - Tuesday, May 10, 2011 - link

    Sure, fan control is needed, I don't argue with that.
    If a single case is going to be utilized for two scenarios - high performance system and quiet system, it's better to use fan control for quiet scenario.

    I'm not questioning this review of BUC case. I'm just pointing that measuring noise so close may offset results in the future, for other reviews.

    You just need $2000 sound meter, then you will be all set.
  • Spivonious - Tuesday, May 10, 2011 - link

    Agreed. I'd like to see both 1ft and 3ft measurements. And 32dB noise floor in a nighttime apartment does seem high. I'd expect a level of 20-25dB, unless the meter is under the air conditioning vent, or the neighbors are having a party. Reply
  • JFish222 - Monday, May 09, 2011 - link

    I'm not familiar with specific of using sata backplanes but assume they use some form of voltage control and/or bridge chip.

    Can you go into further details about how this works (ie: if a bridge chip is used, supplier/model) and specify if it supports SATA rev3 (6Gb)?

    Thanks for another great article.

    - J

Log in

Don't have an account? Sign up now