back to article Feeling a bit gassy? Toshiba floats 16TB helium whopper

Toshiba has promised January sample shipments for its MG08 helium-filled disk drive, which inflates current disk recording technology to a record 16TB capacity. Rivals Western Digital and Seagate max out at 14TB for normal format drives with the former having an Ultrastar DC HC620 shingled media drive with 15TB capacity. This …

  1. Crisp Silver badge

    How do they stop the helium from leaking out?

    From what I understand it's very hard to keep tiny gas atoms like helium in one place. Apparently one of the problems with transporting hydrogen is that those tiny little atoms try and migrate through things. Like the walls of the pipe.

    1. navidier

      Re: How do they stop the helium from leaking out?

      Physics, mainly. Actually, it's not that hard to contain He but if the pressure is around atmospheric there is not much driving force to lead to escape anyhow. If there is any leakage path for He, it's likely to be much harder for N2/O2 to leak back the other way.

      There is a problem the other way. It's apparently not unknown for ancient "vacuum" tubes to have accumulated a significant amount of He because it can diffuse in through the glass envelope and electrode seals much more easily than N2/O2.

      1. Crisp Silver badge

        Re: How do they stop the helium from leaking out?

        That's pretty interesting stuff.

        Have an upvote!

      2. Anonymous Coward
        Anonymous Coward

        Re: How do they stop the helium from leaking out?

        >Physics, mainly.

        Well...

        >Actually, it's not that hard to contain He but if the pressure is around atmospheric there is not much driving force to lead to escape anyhow. If there is any leakage path for He, it's likely to be much harder for N2/O2 to leak back the other way.

        This I cannot agree with. It appears you confuse atmospheric pressure with partial pressure. Helium diffuses though a lot of materials, so special care and special materials and gaskets have to be used to keep it sealed in.

  2. Anonymous Coward
    Anonymous Coward

    Will it sound different?

    More squeaky?

    Enquiring minds want to know :).

    On a slightly more technical note, I'm wondering if helium changes the flows inside the disk and alters the dynamics of turbulence of platters and moving head assemblies - just out of interest.

    1. Thoguht Silver badge

      Re: Will it sound different?

      That's part of the reason why helium works well because its lower density means that there is less drag at the platter surface and so less turbulence and the heads can sit closer to the platters giving improved areal density.

    2. Anonymous Coward
      Anonymous Coward

      Re: Will it sound different?

      So begins the era of hard drive helium huffing.

  3. Roland 2

    WTF is 550TB/year workload rating ? limited endurance or something else

    I keep seeing endurance figures that seem ridiculous for a HDD.

    For this model, the Toshiba spec sheet's exact wording is:

    "550 Total TB Transferred per Year Workload Rating"

    If this is an endurance limit as for an SSD, this would be less than 0.1 full drive read or write per day!

    Or is this a prerequisite for the 2.5 Mh TTF ?

    If it is, it means : "this drive may last, provided you don't use it"

    Can some bright mind explain?

    1. Groaning Ninny

      Re: WTF is 550TB/year workload rating ? limited endurance or something else

      I was wondering the same thing. I suppose if this is aimed at archives it makes some sort of sense. Even my possible home use-case (ripping/transcoding/storing/playing DVD movies) would probably be below that 550TB per year, but this really isn't what I'd want to put in my datacentre for general storage.

    2. Come to the Dark Side

      Re: WTF is 550TB/year workload rating ? limited endurance or something else

      You are indeed correct. Workload Rate Limit is the upper tolerance range for throughput before it will affect risk of failure for the drive. 550TB appears to be the threshold for most "heavy" use designated drives across the manufacturers.

      Basically the manufacturers state that if the overall workload is kept below this threshold then the reliability of the drive will be as advertised. Exceeding the WRL rating reduces the reliability conferred.

      1. Kobblestown

        Re: WTF is 550TB/year workload rating ? limited endurance or something else

        "Basically the manufacturers state that if the overall workload is kept below this threshold then the reliability of the drive will be as advertised. Exceeding the WRL rating reduces the reliability conferred."

        It's probably just an excuse to not honor the warranty.

        1. Come to the Dark Side

          Re: WTF is 550TB/year workload rating ? limited endurance or something else

          Would not surprise me.

          Back-of-fag-packet calculation puts it an average throughput of ~140mbit per sec, 24x7 to keep below the WRL.

    3. katrinab Silver badge

      Re: WTF is 550TB/year workload rating ? limited endurance or something else

      How long would it take to fill up one of these drives? I'm guessing about 3 days. So if you have the drive reading and writing 24/7 at maximum capacity, you would have about 3 times that workload; and you should probably consider SSD rather than spinning rust for your use case.

      1. dfsmith

        Re: WTF is 550TB/year workload rating ? limited endurance or something else

        Consumer SSD endurance is less than 500 overwrites these days. If your workload is like you describe, I'd suggest adding another HDD. E.g., Samsung 850 PRO is 150TB written total; or 30 days under your scenario. Of course, you'd need about 30 SSDs to match the capacity, which takes you out to about 2.5 years before you have no space left.

        1. katrinab Silver badge

          Re: WTF is 550TB/year workload rating ? limited endurance or something else

          Obviously you wouldn't use consumer grade kit, but if you did, then at one read or write every 3 days, then they would last at least 4 years - if they were all writes.

  4. GWS
    Happy

    Perhaps more importantly,will stored audio files sound sqeeky when played back?

    1. Jedit
      Devil

      "will stored audio files sound sqeeky"

      That depends on how old your gramophone needle is.

      1. Anonymous Coward
        Anonymous Coward

        Re: "will stored audio files sound sqeeky"

        they were referring to breathing in helium

  5. Black Betty

    Why not increase the platter size and/or the height of the enclosure?

    What is so bloody special about the 3.5 inch 1/4 height form factor that manufacturers can't revert to 5.25 platters and/or taller enclosures?

    1. AIBailey

      Re: Why not increase the platter size and/or the height of the enclosure?

      Bigger diameter platters mean more distance for the head to travel across the disk surface, which in turn translates to longer seek times. You've also got the problem of vibration and noise being potentially greater for a larger platter size, so physically bigger drives have tended to spin more slowly, meaning a lower data transfer rate. You could try and mitigate the transfer speed problems by packing more sectors per revolution, but again your seek times go down as you wait for the necessary information to be underneath the head.

      The last time I recall a company trying the strategy of larger, cheaper drives was Quantum, with their Bigfoot range in the '90s, however I don't recall them being particularly successful.

      1. Purple-Stater

        Re: Why not increase the platter size and/or the height of the enclosure?

        I completely understand why manufacturers avoid the larger platter, but am surprised that they don't use the full-height format again, to simply double the number of platters. If weight/resistance is an issue on the motor, it shouldn't be that big of a deal to use two motors (one for each half of the platters). Has any HD manufacturer ever tried using more than one set of read/write heads, in order to have more data throughput with lower spin speeds?

        1. Black Betty

          Re: Why not increase the platter size and/or the height of the enclosure?

          Actually you can more than double the number of platters, because stacking two drives into a single enclosure eliminates the top of one case and bottom of the other.

          I'm not sure that it's ever actually been implemented in a real world product, but dual (multiple?) sets of heads has certainly been talked about. One option here would be a single set of write heads combined with two or three sets of lighter weight read heads.

      2. Black Betty

        Re: Why not increase the platter size and/or the height of the enclosure?

        I'm aware of the latency issues, but big drives are generally used for long term archival storage, which as a rule does not require rapid access times. For that purpose petabyte class solid state arrays exist.

        My personal experience with Bigfoot drives is that they failed due to timing issues when the system bus speed was increased beyond 33 MHz.

  6. Dave Hilling

    So does this mean hard drives will get cheaper ..lol

    Doubt it. I mean they aren't super expensive but even getting several 8TB or 10TB is prohibitive to me. Usually I can get multiple times more 4TB drives for example with more capacity than the 8s for the same price. I am running out of room in my case using smaller drives so it would be nice for the larger drives to come down in price a bit since I would imagine the mechanicals etc of multiple drives should be more than a single 8TB with mechanicals or is it just me?

    EDIT: Looking at amazon it looks like prices may be starting to edge down for 8-10TB so maybe soon ill grab a couple. I am personally not a huge fan of shucking drives from externals but it really is the most cost reasonable route.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2019