BBEG wrote:
Necroposting this one because it interests me a lot...
Was the driving factor behind the 955 chip the cost? Looking at the Sandy Bridge i5s, I wonder how much cooler they'll run at equivalent performance levels, thus allowing reduced fan speeds and lower noise.
Variety is more interesting -- for readers and for us; that was a main motivation. I doubt very much that noise could have been lowered with SB i5; 14 dBA idle and 17~18 dBA full load are extremely low SPL numbers. You would not hear this as a discrete sound source in a typical room. The overclocked settings saw no rise in noise at idle, but peaks of up to 25~26 dBA at full load, and this might be improved with SB.... but keep in mind that our test load is not the same as a gaming load -- it is substantially higher and, critically, steady state. Games are more dynamic, with power demand fluctuating up and down, which is why even OC'd, the max SPL during gaming rarely hit 20 dBA, which again is still very quiet.
Quote:
On that note, I notice that the cases chosen are optimized for cooling (air flow) and not for silence. To that end, how much of a difference would you expect the modern cases attempting to optimize both? Fractal Design Define series, Lian Li silent chassis, Silver Stone Fortress and Raven series, and Antec P-series cases... all arguably as good as or better than the Nine Hundred and Tempest Evo in terms of air flow, only they also make an effort to do it quietly. With or without some of the common sense case mods (cutting out the exhaust grills to replace with a fan guard for less air resistance, covering unused fans with noise absorbing foam lining a mounting plate, etc.), would this not have a fairly substantial positive impact on overall sound?
Raven & Fortress are not exactly "noise optimized", imo, though both can be used for very quiet systems. But to answer your questions, imo, I don't think those other cases would make this system any quieter, and even if they did have any effect on overall noise, I question whether that difference would be audible or appreciable -- ie, when the noise level is close to or below the room ambient, further improvements are difficult to hear. (Assuming the typical 1m or greater distance between user and PC.) I also think there might be some slight penalty in cooling with most of those cases except the Silverstones. Most of the cases you mentioned are a little pricier, too -- the Antec 900 is a <$100 case. The noise blocking effect of damping foam and thin sheet metal or plastic is modest at best, and any case that has good airflow has big holes which mitigate the sound damping. This is a primary reason for our general recommendation to reduce noise at the source (ie, choose quieter components to start with) rather than trying to suppress the noise after it has already been generated.
An aside: In most case, I actually prefer the sound of fans (and other noise makers) when they are unenclosed, rather than enclosed. Most components are easier to cool, and fan speeds can be totally minimized, making the overall sound extremely low in level and almost entirely broadband. I run a 4-core 3.2 GHz Phenom II HTPC on an open platform (a piece of plywood big enough for all the components) on an open shelf in the stand under my TV, and its single 500rpm 12cm fan is completely inaudible all the time.
Putting noise sources like fans and HDDs into a smallish box does several things:
* overall SPL is reduced a bit, maybe 3~6 dBA, but not evenly -- ie, sounds at some frequencies are more reduced.
* if sounds at some frequencies fall below audibility, the overall noise becomes less broadband, and the remaining audible noise becomes more tonal (ie, peaky).
* the case (enclosure) itself has air resonances, typically <200Hz, and these of combine with the noise of the components (esp. HDDs) to add a throaty bassy FM radio announcer effect.
All of the above makes an unenclosed quiet system quieter for me than that same system enclosed in a case. Of course, unenclosed systems are far less practical.