12

I frequently use 3.5" hard drives in external "docks", swapping them in and out. The platters are heavy and have great angular momentum and I imagine that if the drive is moved - especially rotated out of the plane - while the hard disks are spinning that would be bad for the bearings, and reduce the life of the disk. So how long does it take for the disk to spin down so that it is safe to move? (I'm not so worried about the head mechanism and head crashes as I'm sure it retracts to a safe position extremely rapidly on power off.)

So my set of closely related questions (totally suitable for a single answer, though answering any one of the questions is of course acceptable and desirable) is:

  1. Is my imagination above, that twisting the disk drive out of plane while the platters are spinning can damage (or reduce the life of) the drive, correct?

  2. How long do I need to wait, after power off, before twisting the drive out of plane?

  3. How long should I wait, after power off, before moving the disk much at all?

  4. So anyway, how long does it take for those platters, rotating freely on bearings designed to be as friction-free as they can be made, possibly even in a helium medium, to spin down from 7200RPM to full stop? Or are there brakes (magnetic or otherwise) that slow them down?

  5. I have occasionally, unmindfully, moved a hard disk out of plane without waiting any length of time. I could swear I've felt a pretty strong gyroscopic resistance. Final question: Was that real or just my overexcited imagination?

(I added the "spinup" tag just for fun. There's no "spindown" tag - if there was it would probably have had a Q with the A to this question!)

9
  • It is entirely dependent on the specifications of the mechanical drive. Every type brand and storage size of drive would be different. – Ramhound Oct 1 '20 at 15:09
  • @Ramhound - those specifications aren't in the drives datasheets, are they? Are approximations available? Really, I'm just interested in guidelines ("best practices") so I don't damage drives (though my engineering-brain wants to know the actual values...) – davidbak Oct 1 '20 at 15:12
  • Given that the actual RPM numbers in the datasheets are not accurate according to recent technical articles, i am not sure you could trust the accuracy, of the numbers if they were listed. I don't have a link to the article I am talking about, the basic idea behind the article, is that the RPM values are closer to an estimation than a specification value (for many reasons most unknown to us common folks). – Ramhound Oct 1 '20 at 15:25
  • @Ramhound - WOW I did not know that! So in addition to computing capacity for marketing purposes different - and larger! - than every OS displays capacity, they're not telling us the true speed either! Next you'll be telling me their MTBF estimates are wildly off too! – davidbak Oct 1 '20 at 15:32
  • If you believe the article. I could go looking for the article, but I suspect it's on a website, I won't promote for personal reasons (but sadly has those rare articles nobody else has). In any event, the RPM being more of an estimate, has to do with increasing number of platters mechanical drives have today. – Ramhound Oct 1 '20 at 15:37
18
  1. As you state, gyroscopic forces on a spinning disk cause precession if the disk is tilted, so tilting a spinning platter would put some additional strain on the bearings. That said, it depends on how rapidly the disk is spinning and on how abruptly it's tilted. Gentle handling of even an operating is acceptable. Consider computers with HDD's operating aboard rolling, yawing ships... yet their lifetimes must be considered acceptable. (Of course, damage from salt spray has to be mitigated.)
  2. Since disk drive spindles are driven by permanent magnet motors, whether considered brushless or stepper, there is a cogging, braking effect, easily seen when operated at low speeds, as in this video starting at 6:10. The permanent magnets also induce eddy currents in the rotor, which causes the rotor to spin down in just a few seconds. On most modern HDD's, the platter is stopped within ten seconds or so of shutdown. Of course, older disk drives took considerably longer to spin down (or up). (See an early design model at 6:00.)

So, all things considered, within ten seconds of powering down, I'd be comfortable moving a 3.5" HDD.

11
  • 11
    Not just computers on ships; Laptops from yesteryear with 2.5" spinning hard drives before SSDs became a thing. And for a device that was moved more than a laptop: the iPod Classic; it used a tiny hard drive (up to 160 GB) to hold everything. – Cole Johnson Oct 1 '20 at 17:36
  • 1
    @ColeJohnson It still amazes me how reliable those HDD-powered iPods are. It's almost an unthinkable idea today. – 比尔盖子 Oct 1 '20 at 22:27
  • 11
    @ColeJohnson, 2.5" hard drives are designed to be moved. Looking at the spec sheets for a random 2.5" drive and a random 3.5" drive, the 2.5" drive has six times the operating shock resistance and four times the operating vibration resistance of the 3.5" drive. – Mark Oct 1 '20 at 23:12
  • 10
    Some drives also generate electricity from remaining momentum to finish outstanding operations. – Simon Richter Oct 2 '20 at 2:22
  • 2
    @ColeJohnson It's not just the iPod. Some time ago, CompactFlash memory cards were made in a variant that actually contained a tiny HDD inside instead of flash chips. These used to be common on the high end of the capacity range back then. My car GPS unit still runs on one of these. – TooTea Oct 2 '20 at 8:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.