Western Digital cooking up 20,000-RPM Raptor?

What’s faster than Western Digital’s new 10,000-RPM VelociRaptor? How about a drive with platters that spin twice as fast? Surprisingly, Bit-tech has word that Western Digital is actually planning a VelociRaptor model with a 20K-RPM spindle speed.

WD itself hasn’t announced anything yet, but Bit-tech claims the drive is in the works and will be WD’s weapon against high-speed solid-state drives. The 20K-RPM VelociRaptor will supposedly resemble the current model, with the same 2.5" form factor and custom-built 3.5" enclosure that doubles as a heatsink. WD will use the enclosure for both cooling and noise dampening with the faster model, allegedly resulting in a "silent" drive. Considering the existing VelociRaptor did fairly well in our noise level tests, that’s not such an outlandish claim.

Bit-tech claims WD is still working on the noise-dampening housing, but it goes on to say details are "incredibly light at this stage" and that it hasn’t learned of a release time frame yet.

Comments closed
    • Bensam123
    • 13 years ago

    I was hoping a 15k Raptor would someday be released. You can find used SCSI-320 gear on ebay that is cheaper per gig then a Raptor by about half the price. Not to mention they spin faster, have higher transfer rates, are built like trucks, and they have all the frill benefits of SCSI standards. So this tips the scale for me as you don’t need a extra controller, they’ll be faster, and you wont need special cables

    Noise is a big issue. I bought 4 of these and even if you don’t hear any mechanical noise between the seeks, the ‘wooshing’ sound they create can drive you up the walls. I can only imagine what a 20k RPM drive in a smaller enclosure will do. Smaller things do tend to generate loud, annoying, ear piercing noises later on in their life.

    • green
    • 13 years ago

    this might end in tears (literally)

    1x cd drive = 500rpm
    4x cd drive = 2000rpm
    8x cd drive = 4000rpm

    48x cd drive = 24000rpm
    52x cd drive = 26000rpm

    i remember a sony drive running at only 40x when your bought a 52x
    you had to do a little trick to unlock 52x speeds “at your own risk”
    the problem being when cds spun too fast they shattered
    sometimes it would just screw over the drive completely
    sometimes it broke free of the tray and spat out shards
    (this wasn’t just sony drives mind you)

    a hdd is constructed much stronger obviously
    but if that thing does self destruct, it’ll end in tears

      • UberGerbil
      • 13 years ago

      Well, the other problem with CDs is that they’re quite flexible so they tend to wobble and flex, which makes it hard for the laser to track and read them, and they’re pressed en mass so there’s always a little variation in how true they spin. HD platters are much stiffer and engineered to far higher tolerances, but of course they’re spinning at higher rpms as well. With an HD any significant ripple would lead to a head crash, and any significant wobble would set up so much stress and vibration you’d probably burn out the motor (or, again, crash the head), and either of those would likely happen before you had a platter actually come apart. As a last resort, well, the external case for HDs are made of metal for a reason.

      • axeman
      • 13 years ago

      Your figures are way wrong. CDs are going to shatter way before 26k rpm. From wikipedia: “As of 2004, the fastest transfer rate commonly available is about 52x or 10,350 rpm and 7.62 megabytes per second, though this is only when reading information from the outer parts of a disc. Future speed increases based simply upon spinning the disc faster are particularly limited by the strength of polycarbonate plastic used in CD manufacturing, though improvements can still be obtained by the use of multiple laser pickups as demonstrated by the Kenwood TrueX 72x which uses seven laser beams and a rotation speed of approximately 10x.

      FYI at 1x speed, the rotational speed was only around 200rpm at the end of a disc, 500rpm was at the beginning of a CD

      The 52x claim has to deal with the max data rate, which is calculated by dividing by 150kb/sec which is even inaccurate measure of performance in itself, since 1x playback (audio) is actually 176,400 bytes a second. So even at at 10k rpm, the fastest reader is only doing an honest 44x at the very end of a disc, and perhaps 18x at the beginning.

    • bdwilcox
    • 13 years ago

    The one figure I’ve never seen/compared is data stability estimates between SSDs and hard drive media. How long can we expect the data to stay intact before we get randomly flipped bits? Is hard drive media more stable than SSDs for archival/read-only puposes?

      • Krogoth
      • 13 years ago

      The data on the platter is more durable on a HDD. The problem is the damm motor giving out.

      SSD data has a more graceful death as you just lose ability to write/read on bad bits and it gradually gets worse with age.

        • UberGerbil
        • 13 years ago

        Well, yeah, but you don’t want to count on leaving a data-filled HD on a shelf for years and then being able to start it up and recover the data off it. HDs fail eventually when running but they can fail from stiction and other issues when they’re sitting for extended periods without power too.

    • NoOther
    • 13 years ago

    People there are many factors here. First, name all the other 10k SATA drive manufacturers. Aha, there is only 1. Remember the 15k drives are using a SCSI format and there are limitations to that as well. Many other companies don’t see a market in the 10k RPM range so they don’t care, and some of them are investing in the SSD market anyway. If WD who was the ONLY one to come out with a 10k SATA drive decides that a 20k will be cheap enough compared to the SSD, then that is what they will do. Also realize that SSD has limited capacity so far. There is nothing in the same range as the SATA drives right now.

    So it would make a lot of sense for WD to put out a fast higher capacity drive to catch the enthusiast market. Also this could very well capture more interest in the enterprise market as well.

    • Krogoth
    • 13 years ago

    Why?

    SSDs are going to render high-RPM HDDs obsolete in a few years.

      • mentaldrano
      • 13 years ago

      Sic transit gloria mundi

      • ish718
      • 13 years ago

      Not exactly, not with their prices so high, not even in a few years.
      I doubt the transition from HDD to SSD will be that smooth lol

      • NoOther
      • 13 years ago

      SSD drives have “lower” capacity. You cannot use them for everything, and you can only string soo many of them together, and then think about the cost of having to buy all those drives to get the same capacity as a SATA drive?

        • Krogoth
        • 13 years ago

        You fail to understand the raison d’etre for high RPM HDD.

        They exist solely because of superior access speed to the data not for capacity.

        SSDs are much faster at this then any mechanical based scheme.

        Thus there is really is no reason to get a high RPM HDD. The mainstream units are almost as fast overall while having far greater capacities then both SSDs and high RPM HDDs for less $$$$.

          • UberGerbil
          • 13 years ago

          Yeah, SSDs will kill off the high-rpm disks and mobile disks long before they have any serious impact on mainstream HDs. The window of opportunity for an HD like this is very limited, and it’s hard to see how WD could make back their R&D investment even if they also do an enterprise version. MLC tech, plus the major push by Intel and Micron on top of Samsung’s relentless advances (50%+ drop in prices year over year) mean that SSDs are going to be making inroads soon. For the people buying Raptors as OS-only drives, SSDs will be pushing into the conversation on a cost basis by the end of the year, since 64GB (or even 32GB) is plenty for that. As the mainstream notebook market starts to adopt the 128-256GB drives in the following years those will start looking good for OS+program drives on desktops. Media storage is the last bastion for HDs, and will be for a while, but you don’t need insane RPMs for that.

    • UberGerbil
    • 13 years ago

    ฯ€*2.5 * 20K = 125,663.71 inches/minute = 119 miles/hour. Which is 0.27 mach at sea level.

    Same math for a 3.5″ platter gives an edge velocity of 208 mph

    I believe a big issue with larger platters turning at higher speeds is finding cheap enough materials that will stay stiff enough to prevent small vibrations from causing head crashes. But I know there are bunch of other factors as well. Whatever the killer is, it’s not sonic booms.

      • jsncable
      • 13 years ago

      Is my math or your math wrong?

      I get ฯ€*2.5 * 20K = 157,000 inches/minute = 148.6742 miles/hour. Which is 0.19 mach at sea level.

      Not picking,, just checking my math skills.

        • Spotpuff
        • 13 years ago

        You’re right.

        Sometimes pushing buttons on the calc.exe thing is hard ๐Ÿ™

          • Usacomp2k3
          • 13 years ago

          That’s why you use powercalc. ๐Ÿ˜‰

            • UberGerbil
            • 13 years ago

            Actually, I just did the equation doing google. But clearly I screwed something up somewhere. Oh well, close enough
            ยง[<http://www.google.com/search?&q=PI+*+2.5+*20000+inches+per+minute++%2F+speed+of+sound<]ยง

            • TheTechReporter
            • 13 years ago

            Close enough?
            Man, I thought you were a true geek.
            Obviously I was wrong, because real geeks don’t know the meaning of the words “close enough”. ๐Ÿ˜›

            Back on topic, the point here is that hard drives can’t just keep increasing RPM forever. SSD’s are already making up serious ground in terms of speed and capacity (already surpassing 1.8″ and smaller drives), and should even reach price parity given enough time. HDD’s may die a slow death, but it’s already the beginning of the end for them.

        • MethylONE
        • 13 years ago

        Considering that Mach is not a fixed number without including temperature and pressure you can’t really do that equation.

        (Yeah I know you said ‘sea level’ but that doesn’t mean much.)

          • UberGerbil
          • 13 years ago

          One ICAO standard atmosphere (29.92″Hg/101.3kPa and 15ยฐC) — and standard lapse rates for altitude ASL — was always assumed whenever we did these calcs in flight school or my aerodynamics classes. Within the range of real-world temperatures at sea level (and particularly the operating range for HDs) Mach doesn’t vary enough to make a significant difference for the single digit precision we’re quoting here. I suppose if you’re operating the HD during a hurricane the pressure drop might be enough to affect the Mach calculations, but then you’ve got bigger things to worry about.

          In any case, since we’re nowhere near the speed of sound anyway, normal variations in temp and pressure aren’t going to tip it over the edge.

      • moritzgedig
      • 13 years ago

      talk about diameter and revolutions:
      ยง[<http://www.techreport.com/ja.zz?id=324619<]ยง ยง[<http://www.techreport.com/ja.zz?id=324565<]ยง SSD will be faster because you can read from many places at the same time. what about a HDD with two arms? would they affect eachother to much?

    • albundy
    • 13 years ago

    anyone here actually work with a 15k scsi drive? all of them in my workplace are in separate enclosures in the chassis in an air conditioned room and they are very hot to the touch. i presume at 20k, you will have your fingers burnt good. that being said, would you want a 20k radiator right under your brand new blu-ray burner? also, how is wd going to deal with power consumption?

      • bdwilcox
      • 13 years ago

      They’re a lot of fun to hot swap. They act like crazy gyroscopes when you have to pull them while they’re still spinning.

        • Krogoth
        • 13 years ago

        Oh yes, the old HDDs were more infamous for that.

        I still have an old Quantium Bigfoot that will move around a little bit after it powers down.

          • albundy
          • 13 years ago

          heh, i remeber those brick drives. they were actually the size of a brick! those differential scsi drives would keep spinning like crazy.

    • bdwilcox
    • 13 years ago

    When one of these things lets go, it’ll be like a bomb going off in your server rack. Man, that’ll be fun.

    I don’t feel like doing the math, but won’t the edges of a 2.5″ platter spinning at 20,000RPM break the sound barrier? I believe this is what prevented 15,000 RPM 3.5″ platters.

      • Peldor
      • 13 years ago

      It’s “only” Mach 0.2

        • ew
        • 13 years ago

        That is accurate. Although the platters won’t span the entire width of the enclosure.

      • echo_seven
      • 13 years ago

      Okay, so no 100,000 rpm drives, then ๐Ÿ™‚

        • UberGerbil
        • 13 years ago

        Yeah. Bummer.

        I had a friend who, in grad school, was working on a project involving some kind of ion deposition (I forget the details). The system they had set up involved levitating a mirror magnetically in a vacuum and then spinning it at some insanely high RPM like that. They had to armor the thing (and operate it from another room) to contain the shrapnel if the mirror ever contacted the walls of its containment vessel (or just came apart from internal stresses).

      • Shining Arcanine
      • 13 years ago

      Why not vacuum seal them? If there is no air, there is no sonic boom. ๐Ÿ™‚

        • UberGerbil
        • 13 years ago

        That creates other problems — assuming you can build something that retains a vacuum over its lifetime… if not, boom — primarily that the hard disks work by “flying” the head on a cushion of air over the platter. No air, no cushion, head crash. So now you’re off on your own designing something different from the rest of the industry. Oh, and lots of common lubricants don’t work well in a vacuum, so add that to your list of problems. How much did you want this to cost?

          • pins
          • 13 years ago

          I thought that in newer drives the head was levitated not by the air cushion, but by the electromagnetic force exerted on it by the bits on the platter.
          Can’t find a reference

      • Lord.Blue
      • 13 years ago

      There have been 18k rpm scsi drives made, but I am unsure if they are 3.5″ or 2.5″.

    • UberGerbil
    • 13 years ago

    You can’t fight Moore’s Law.

    Well, you can, but you’ll lose, eventually.

      • Krogoth
      • 13 years ago

      Again, it is not Moore’s Law. It is an observation.

      It already has been killed with the start of the Prescott. It is merely slowing down due to physical limitations of semiconductors and cost of doubling transistor count is outrunning economic viability to do it.

        • UberGerbil
        • 13 years ago

        Uh, yeah, I know. But that’s what it’s popularly called. Goodwin’s, Clark’s, and Sturgeon’s aren’t technically “laws” either, but nobody seems to freak out when they’re called that.

        When I’m feeling like a nitpicking pedant (or feel like I’m going to be pointlessly corrected by one), I call it “Moore’s Imperative”: an economic force that companies ignore or fight at their peril.

    • Peldor
    • 13 years ago

    It’s a nice product if they can do it in a timely fashion, but it’s not really going to slow the onslaught of ever cheaper SSDs.

    Seek times will still be milliseconds vs microseconds for SSD. Sequential transfer rates will be awesome if they could keep the same density platters. I don’t know if that’s possible at 20k rpm.

    And I think maybe it’s not really silent. It’s just pitched too high for human ears. ๐Ÿ™‚

    • Cova
    • 13 years ago

    If it was economically viable to make 20K RPM drives, even if only at a price that no sensible consumer would ever buy, someone (seagate probably) would already be making them for the enterprise storage market (which currently spends far more on 15K RPM fiber-channel drives).

    So – I’m calling BS on this. Someone who already makes a 15K RPM drive will be the first ones to release a 20K RPM one – and it’ll be for the enterprise market (SAS or FC interface) before it moves to the consumer market.

      • UberGerbil
      • 13 years ago

      Yeah, I agree, and that was my first thought: we’re going to get a consumer 20K drive when 15K is as high as it gets in the server market? The Raptor (like a lot of other things) was just enterprise tech trickling down to the enthusiast market, and had the virtue of having the R&D mostly already paid for. Even if they’ve been working on a 20K drive in the background for some time, I can’t see how they can introduce it at a price point even enthusiasts will find reasonable without the volumes they’d get from also selling a version of it in the enterprise market. Unless they think there’s a lot of demand for those $4K 128GB SSDs /[

      • eitje
      • 13 years ago

      you know, someone has to try an idea before everyone else can do it. ๐Ÿ™‚

    • bjm
    • 13 years ago

    Whoa, man.. Now that is some interesting news in storage development! The whole SSD situation is just too high priced for me. Raptors, on the other hand, are within reach. They just weren’t interesting anymore wrt performance from the 7200RPM drivers. If they can pull this off without a rediculous SSD-like price tag, I’m all for it!

      • kuraegomon
      • 13 years ago

      Have you seen the VelociRaptor? That thing is definitely relevant … against 7.2K drives. Still though, I’m thinking that 2 years from now will be an absolutely fantastic time to go for some 2- to -4-way RAID-0 SSD action. Mwuuaahahaha.

    • Jigar
    • 13 years ago

    If this news are true.. I think i will have to crank up my sound systems volume.. ๐Ÿ˜€

      • Jon
      • 13 years ago

      You won’t have to…read the article again. It’s going to be relatively silent, which means it’s going to be quieter than the 74GB Raptor’s, which are fairly quiet anyway.

        • Jigar
        • 13 years ago

        Don’t know, for some reason, my logic says it might turn out a lot noisier than it’s predecessors…

        • ludi
        • 13 years ago

        “alledgedly”, which follows shortly after the qualifier “WD itself hasn’t announced anything yet” — which is an awfully big pile of words to chew on for something that will spin in the upper end of the range of dental drills and Dremel tools.

    • Fragnificent
    • 13 years ago

    Unless they can do much better on capacity, I am done with the raptors..they are fast but don’t hold enough data for me. 7.2K 640GB ftw!

      • Taddeusz
      • 13 years ago

      Yea, I would have to agree. I have one of the 640GB drives on my main system running quite nicely. ๐Ÿ˜€

        • DASQ
        • 13 years ago

        On the contrary, I still want them to release them in single-platter low capacity designs for lower heat output.

        My Raptor is a pure OS drive, nothing else.

          • willyolio
          • 13 years ago

          yeah, if you’re enough of an enthusiast to buy a raptor, why would you have anything less than 2 hard drives?

            • DASQ
            • 13 years ago

            I’ve got 6.

            Out of SATA ports on my P5K-E ๐Ÿ™

            • Forge
            • 13 years ago

            Buy a better mobo. 7/8 ports on my EP35-DS4 in use. Also the PATA one.

            Also, poster #1 misses the point. Some of us don’t want to have to partition six ways or use part of our OS/programs disk for storage. For OS/programs, 640GB is about 400-500GB overkill.

            My best friend goes a step farther, putting OS on one small disk, programs on another, storage on all the rest. I think that’s going too far, but with the size of some of today’s games it’s making more and more sense.

            • Fragnificent
            • 13 years ago

            With all due respect I am not missing my own point. Which is, why should I pay a premium for a fast, tiny drive, when for LESS money I can have a huge, just oh so very slightly slower drive? I’d have to say that buying a Raptor for e-peen or omfg i Have t3h 10k rpMz!! doesn’t seem like a fincancially wise decision. Also, SSD’s are getting cheaper every month. Pretty soon, you can kiss moving drives goodbye. I won’t shed a tear, although I will miss the flutter of the read write head. ๐Ÿ™‚

            • Taddeusz
            • 13 years ago

            No matter how much an enthusiast you are reality has to play a part somewhere. The fastest 7200RPM SATA drives are in most cases faster than a 10000RPM Raptor. The only place the Raptor has the edge is in rotational latency due to the higher speed.

            Is the price premium paid for a Raptor really worth the slight edge over cheaper larger drives? IMHO, not at all.

            And I do have more than one drive in my main computer but only because the second drive serves as a file share for my converted DVD’s for SageTV.

            • Mithent
            • 13 years ago

            Agreed, but we’re in the minority. The VelociRaptor has changed things somewhat, but if you show Raptor real-world benchmarks to people who have them and show them how there’s no noticeable difference with modern 7200rpm drives, they’ll tell you that, never mind the benchmarks, it just “feels” better. Same with RAID0, and Windows Server 2008 as a workstation. People who know they have some kind of “advanced” setup always claim it’s far better in rather undefinable ways that benchmarks don’t pick up. I’d like to see double-blind tests, personally.

            • indeego
            • 13 years ago

            far better = we paid a lot of money so it better feel fasterg{

Pin It on Pinterest

Share This