4 ms·

If my memory serves me right, 7200 rpm result in higher spinning but reduced seeking/head noise (which I personally dislike more). edit: I'm probably wrong

by rbinv 6y ago

bayindirh 6y ago
7200 RPM drives indeed have faster access (in terms of seek latency) and higher sustained throughput but, heat and power is not simple variables as one might think. Especially RED drives are marketed for NAS systems (from home to SMB size). When a drives are installed into a closure, things like power and thermal budget comes into play. NAS boxes are designed with strict limits and if they cannot cool themselves (and the drives) off or cannot supply the required power to drives; data corruption, performance/life span degradation may occur. In extreme cases (with large storage arrays) calculated overall power budget can skew and may cause blackouts in some parts of the data center. There's a presumed ~2W difference in power consumption between 5.4K and 7.2K drives. We can safely install 60 of them in a 4U enclosure today. If we add 4 enclosures to a rack, that creates a ~480W difference, which is a one HPC server or 2-3 low-load, 1U or 2U servers. This is indeed significant. While some of these drives are in MyBook enclosures, REDs may end up in very different places and may cause problems. In that case, head or spindle noise is not even the smallest concern. If we're talking about noise, continuous (spindle) noise is much more degrading than head/seek noise. For me, even benign noises like USB fans have much more noticable effects after the day than mechanical keyboard noise for example.
- ballenf 6y ago
  > There's a presumed ~2W difference in power consumption between 5.4K and 7.2K drives. Is that purely extra friction due to higher speed? Or is it spin up from sleep needing more power?
  - bayindirh 6y ago
    If we consider WD80EFAX [0], which is one of the suspect drives, that figure is for idle power consumption where disk is spinning but heads are not doing anything. So, that difference is for sustaining higher spindle speeds. The peak power requirement is 1.79A (~21.5W) which is probably consumed when starting from full-stop state (stand-by, sleep or power-off where motor is turned off). [0]: https://media.flixcar.com/f360cdn/Western_Digital-3805661149-eng_spec_data_sheet_2879-800002.pdf https://media.flixcar.com/f360cdn/Western_Digital-3805661149... [0]: https://products.wdc.com/library/SpecSheet/ENG/2879-800002.pdf https://products.wdc.com/library/SpecSheet/ENG/2879-800002.p...
noisem4ker 6y ago
Why would rotation speed inversely affect head noise?
- rbinv 6y ago
  Yeah, I came to the same conclusion after thinking about it some more. Looks like a faulty memory.
- sandworm101 6y ago
  Ever drive a manual transmission? At very low RPM the engine will seem to shake more. It isn't so much that there is less vibration at higher speeds but that it occurs at a higher frequency. That means the amount the part physically moves from one side of the oscillation to another is less, just as a higher-revving engine doesn't bounce around as much as it would at lower RPM. The total amount of noise/vibration may be similar but the change in frequency can still have other effects.

If my memory serves me right, 7200 rpm result in higher spinning but reduced seeking/head noise (which I personally dislike more). edit: I'm probably wrong

by rbinv 6y ago

If my memory serves me right, 7200 rpm result in higher spinning but reduced seeking/head noise (which I personally dislike more). edit: I'm probably wrong

bayindirh 6y ago

7200 RPM drives indeed have faster access (in terms of seek latency) and higher sustained throughput but, heat and power is not simple variables as one might think. Especially RED drives are marketed for NAS systems (from home to SMB size). When a drives are installed into a closure, things like power and thermal budget comes into play. NAS boxes are designed with strict limits and if they cannot cool themselves (and the drives) off or cannot supply the required power to drives; data corruption, performance/life span degradation may occur. In extreme cases (with large storage arrays) calculated overall power budget can skew and may cause blackouts in some parts of the data center. There's a presumed ~2W difference in power consumption between 5.4K and 7.2K drives. We can safely install 60 of them in a 4U enclosure today. If we add 4 enclosures to a rack, that creates a ~480W difference, which is a one HPC server or 2-3 low-load, 1U or 2U servers. This is indeed significant. While some of these drives are in MyBook enclosures, REDs may end up in very different places and may cause problems. In that case, head or spindle noise is not even the smallest concern. If we're talking about noise, continuous (spindle) noise is much more degrading than head/seek noise. For me, even benign noises like USB fans have much more noticable effects after the day than mechanical keyboard noise for example.

ballenf 6y ago
> There's a presumed ~2W difference in power consumption between 5.4K and 7.2K drives. Is that purely extra friction due to higher speed? Or is it spin up from sleep needing more power?
- bayindirh 6y ago
  If we consider WD80EFAX [0], which is one of the suspect drives, that figure is for idle power consumption where disk is spinning but heads are not doing anything. So, that difference is for sustaining higher spindle speeds. The peak power requirement is 1.79A (~21.5W) which is probably consumed when starting from full-stop state (stand-by, sleep or power-off where motor is turned off). [0]: https://media.flixcar.com/f360cdn/Western_Digital-3805661149-eng_spec_data_sheet_2879-800002.pdf https://media.flixcar.com/f360cdn/Western_Digital-3805661149... [0]: https://products.wdc.com/library/SpecSheet/ENG/2879-800002.pdf https://products.wdc.com/library/SpecSheet/ENG/2879-800002.p...

noisem4ker 6y ago

Why would rotation speed inversely affect head noise?

rbinv 6y ago
Yeah, I came to the same conclusion after thinking about it some more. Looks like a faulty memory.
sandworm101 6y ago
Ever drive a manual transmission? At very low RPM the engine will seem to shake more. It isn't so much that there is less vibration at higher speeds but that it occurs at a higher frequency. That means the amount the part physically moves from one side of the oscillation to another is less, just as a higher-revving engine doesn't bounce around as much as it would at lower RPM. The total amount of noise/vibration may be similar but the change in frequency can still have other effects.