36

Why do hard disk drives mostly have the same speed? They are mostly 5400, 7200, 10000 RPM. Why not 6000 or 8000? Is there any reason or benefit for having those speeds?

1

4 Answers 4

43

The spindle rotation speed of a hard disk depends on a spindle motor, naturally. Since there are only very few manufacturers of these motors, the available rotational range is limited. In fact, the current global shortage of hard disks is because of two flooded hard disk motor factories in Thailand.

The most common rotational speed was 3600 rpm in the 1980s. This is because oscillators and motors that are based around 60-hz are very common. Standard North American AC mains frequency is 60hz. 60hz translates to 3600 RPMs. Therefore 3600 rpm AC motors are widely available. Naturally, the designs that followed were based on the same rotational speeds/frequencies.

So the most common speeds ever used are 3600, 3600*1.5=5400, 3600*2=7200. Faster drives use 10,000 and 15,000 rotational speed, apparently they use different frequencies.

This StorageReview article discusses spin speed in depth.

11
  • I can verify that many early disk drives indeed did use (relatively) standard AC motors that would have naturally run at 3600 RPM. Nov 12, 2011 at 23:29
  • @haimg - Where did you read "direct-drive" in that article? I just see mention of "AC motors". BTW Storage Module Drives (SMD, such as those made by Control Data Corp.) used an AC, 1/2 horsepower, induction motor with belt drive. Don't know about 8" HDDs, but 8" floppy drives used AC motors and belts.
    – sawdust
    Nov 13, 2011 at 3:48
  • 7
    Whoa, AC motors? How would they do that when the power supply supplies DC (don't tell me it contained a DC/AC inverter?)? Nov 13, 2011 at 8:18
  • 5
    @BlueRaja: You really didn't want a DC power supply back then, for the drive motor. Check out en.wikipedia.org/wiki/IBM_magnetic_disk_drives - the earliest models ran at 1200 and 1800 RPM, using 2374 Watt.
    – MSalters
    Nov 13, 2011 at 12:20
  • 1
    @haimg - You're not alone; wikipedia also uses "storage module device". However, in the 1980s, I wrote the firmware and driver for a SMD controller, and I still have two CDC Hardware Reference Manuals for models BJ4A1/BJ4A2 (doc #83319200) and BJ7xx (doc #83308500) dated 1977. Title pages and text use the phrase "Storage Module Drive". Maybe CDC changed the meaning of "SMD" after 1977, but I doubt that. Considering that the proper name for the disk pack is "Storage Module" (and IBM used "data module"), SMD is a (transport) drive for a SM. Calling it a "device" is obfuscation.
    – sawdust
    Nov 14, 2011 at 7:43
9

Right now, you can get drives as fast as 15K. We're unlikely to ever see drives spin any faster, for two reasons:

  1. The shift to solid state for performance-critical uses
  2. Faster rotational velocities literally risk that the platter will rip itself apart.

When rotational velocity remains constant, linear velocity increases exponentially as you get further from the center. This is why all the 15K drives you see tend to come with 2.5" platters (the link above being a notable exception). They could make a faster drive if they reduce the platter size (and therefore capacity), but we're unlikely to ever see this because of item #1 above and because of motor issues outlined in other answers.

For the rest of it, I think the real answer has more to do with marketing. Sure, there are specific models of motor out there, but if a manufacture really wanted to market a 7500 or 8000 rpm drive, I expect they could work with their suppliers to make it happen. It's more believable to me that they find it easier to sell drives when there are just a few varieties that are well-understood by consumers.

3
  • 5
    FWIW, the relationship between linear velocity and distance from the center (i.e. radius) is linear, not exponential: circumference=2π*radius. Nevertheless, the velocity does increase pretty quickly, and the centripetal force increases with the square of the velocity.
    – Caleb
    Nov 13, 2011 at 5:17
  • FWIW, the problem at running at > 20 krpm is more about oscillations and stuff. I knew a guy who did a Ph.D. on magnetic (actively regulated) "bearings" for high-RPM drives. They ran easily at 20-25 krpm (if not more), but the design was too complex/expensive to be commercially viable.
    – Macke
    Nov 13, 2011 at 20:28
  • Also, a Formula One V12 engine from a few years back revved to 20-22 krpm. None of these ripped themselves apart due to the speed (Now it's regulated to 18rpm max)..
    – Macke
    Nov 13, 2011 at 20:30
8

If you look at the situation from HDD manufacturers' point of view, the reliability, the speed, etc. of drives should be tested separately. So the more variety of speeds you have, the more it increases your general expenses. Note that common HDD speeds are listed as 3600, 3811, 4000, 4400, 4500, 5200, 5400, 7200, 10000 and 15000 RPM (the first IBM hard drives actually spun at 1200 rpm).

1
  • Thanks for the link to brilliant site Dec 21, 2022 at 10:10
7

They don't!

That being said, the reason there isn't thousands of different ones is because it isn't likely that a 7200RPM one goes at exactly 7200RPM - these are the average speeds of the motors.

In addition, as much as there are only a few hard drive vendors in the world, there are even fewer specialist vendors who can produce motors of a quality and quantity needed for hard drive production. Because of this, why would they want to make 20-30 different motors instead of specialising/mass producing just a few.

However, there have been a few newer speeds creeping in such as 5600 and 5900 drives in recent years which are aimed to be "greener" than 7200RPM whilst maintaining good speeds.

As for why we don't have faster 100k drives.... All I can say is physics and thermodynamics! If it was possible, we would - but, generally, the faster the speed, the lower the capacity - because, you have to remember, it isn't just about the speed the disks spin at, the read head has to be able to keep up and read over the platter, in addition, the faster the spin, the higher the heat.

All this being said, I don't think we are going to see many more traditional hard drive technologies - give it 2 years and we are probably just going to have slower green disks for NASs/mass storage and SSDs for mainstream.

(And thankfully, the 4800RPM laptop drive is nearly completely dead!)

2
  • +1 for mentioning heat. At those speeds, the disk bearings and motor must get pretty hot. Nov 13, 2011 at 4:29
  • Actually, my understanding is that the rotational speed being tightly regulated is a very important requirement, because otherwise reading (let alone writing) data becomes much more difficult. This of course doesn't necessarily imply that the rotational speed has to take on some particular value (there would be no particular reason why you must have 7200 rpm, and 7189 rpm is unacceptable), but using standard speeds certainly do make things easier as you can use a small number of different model motors across your entire product line, thereby reducing cost.
    – user
    Dec 12, 2017 at 21:45

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .