On Mon, Dec 22, 2008 at 08:58:37AM -0500, Justin Piszcz wrote: > > Besides the power supply/surge/etc with a lot of drives-- I have read: > > http://www.wrightthisway.com/Articles/cat_reviews.html > "As I mentioned earlier, the enclosure will definitely handle drives using > Spread Spectrum Clocking (SSC), an increasingly common feature that helps > reduce electromagnetic interference, especially between drives in close > proximity with each others, such as you might have in a RAID setup, so > that is a definite plus here." The SSC is implemented by intentionally adding JITTER on a clock signal. Problems surface when such jittery clock is reading/writing disk surface.. SSC does "help" on EMC compliance by spreading the leaked clock signal to wider frequency range -- effectively lowering the power at any given single frequency, but at same time increasing total leaked power... SSC works fine with some things, with others it will need asynchronous elastic buffering (known as FIFO) when communicating with precission synchronous clocking domain. Harddrives, USB, SATA, etc. serial links I would not dare to run with SSC clocks -- and properly done circuitboard does not radiate electrical noise. > With 16-24 disks-- if they are close together, it sounds like a good idea > to use Spread Spectrum Clocking? > Does anyone here with a large-ish raid array use this? I would consider that BAD IDEA -- intentionally lowering system reliability by adding jitter on clocks is ... wrong way to fix the EMC issues. Hiding the leaked oscillator signal with SSC out of a device does not have any effect on reducing device susceptibility on externally originating RFI. I never use SSC, but then my designs can afford that 5 dollars more expensive circuit-board.. > Any enterprise-insiders care to comment? > Justin. Matti -- To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
