Testing Methodology

For testing Micro-ATX and full ATX cases, we use the following standardized testbed in stock and overclocked configurations to get a feel for how well the case handles heat and noise.

ATX Test Configuration
CPU Intel Core i7-2700K
(95W TDP, tested at stock speed and overclocked to 4.3GHz @ 1.38V)
Motherboard Gigabyte GA-Z68MX-UD2H-B3
Graphics Card ASUS GeForce GTX 560 Ti DCII TOP
(tested at stock speed and overclocked to 1GHz/overvolted to 1.13V)
Memory 2x2GB Crucial Ballistix Smart Tracer DDR3-1600
Drives Kingston SSDNow V+ 100 64GB SSD
Samsung 5.25" BD-ROM/DVDRW Drive
Accessories Corsair Link
CPU Cooler Cooler Master Hyper 212 Evo with Cooler Master ThermalFusion 400
Power Supply SilverStone Strider Plus 750W 80 Plus Silver

Each case is tested in a stock configuration and an overclocked configuration that generates substantially more heat (and thus may produce more noise). The system is powered on and left idle for fifteen minutes, the thermal and acoustic results recorded, and then stressed by running seven threads in Prime95 (in-place large FFTs) on the CPU and OC Scanner (maximum load) on the GPU. At the end of fiteen minutes, thermal and acoustic results are recorded. This is done for the stock settings and for the overclock, and if the enclosure has a fan controller, these tests are repeated for each setting. Ambient temperature is also measured after the fifteen idle minutes but before the stress test and used to calculate the final reported results.

Thank You!

Before moving on, we'd like to thank the following vendors for providing us with the hardware used in our testbed.

Assembling the Antec Eleven Hundred Noise and Thermal Testing, Stock
POST A COMMENT

39 Comments

View All Comments

  • Stahn Aileron - Monday, April 09, 2012 - link

    Would it be too much to ask that you actually place the ambient tempurature during testing somewhere on the graphs themselves so we have a readily accessible reference point right there on the graphs? It would also work as a reminder to readers what the conditions are. Reply
  • Dustin Sklavos - Monday, April 09, 2012 - link

    The ambient temperature is going to vary from test system to test system, that's why I switched to listing the delta. Reply
  • Rick83 - Monday, April 09, 2012 - link

    You did check, that the delta is a constant over ambient temperature though, did you? It may not be for every case. Also, of course, fan speeds will be impacted by ambient/internal temperature.

    If you can, you should probably run two series of tests, one in the morning and one in the evening, and then either average that or use one measurement, but at least comment the other. At least for noise, we need ambient temperature, as otherwise that value is completely without base for comparisons.
    Reply
  • kandrtech - Monday, April 09, 2012 - link

    Those familiar with thermodynamics, and the equations utilized, would agree that the delta approach is the best. Variances of a few (or 10) degrees on ambient will not appreciably change the delta results. By appreciably, I mean you'll see differences out to one or more decimal places . . . . Reply
  • niva - Monday, April 09, 2012 - link

    Are you talking 10 deg C or F?

    Ideally your ambient temperature should be somewhere in the +/- 5 deg of 70 deg F. These are the normal temperatures most households are kept at. There may be a significant difference between the noise produced by components at 65 deg F, and someone's house which may be normally kept at 85 deg during the summer daytime because of lack of AC?

    And I'm talking about idle situation here...

    The point was valid, just include the temperature in your test data.
    Reply
  • bobbozzo - Monday, April 09, 2012 - link

    The article describing their new methodology for case testing seems to indicate that ambient temps are maintained between 71-74F.

    http://www.anandtech.com/show/5709/introducing-our...
    Reply
  • ShieTar - Tuesday, April 10, 2012 - link

    But those familiar with modern PC design are aware that fan-control systems generally try to achieve a constant CPU/GPU-temperature. Thus when you raise the ambient temperature to somewhat higher levels, CPU/GPU fans tend to speed up, giving you lower Delta-T values at increased noise.

    Thus it is important to still run these tests at comparable ambient temperatures, and if this is indeed checked at each test, it should be no problem to change the title of the temperature graph to read "Delta over ambient at 20+/-2°C". Or whichever is the range that is controlled and accepted by the tester.
    Reply
  • O8h7w - Saturday, April 14, 2012 - link

    I feel perfectly good about showing the temperature as delta above ambient instead of absolute temperature. But it seems many readers would like to see the ambient temperature at the time of testing reported as well, and I have to agree.

    The way of doing this that would make perfect sense in the graphs is to modify the labels to look like this:
    Antec 1100
    @ 23°C ambient
    Reply
  • Lucian2244 - Monday, April 09, 2012 - link

    Good review, i was wondering how it would look with a mATX in there.
    Is it just me or their cases get uglier and uglier ?
    Reply
  • Iketh - Monday, April 09, 2012 - link

    It's not you... this case is ugly... AND stupid...

    Enough with filterless side vents already! Why even have side vents? That's a damn 80's design, speaking figuratively...
    Reply

Log in

Don't have an account? Sign up now