1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

What happens with HDD if power is lost?

Discussion in 'Storage' started by Swamp Monster, Jan 22, 2011.

  1. Swamp Monster

    Swamp Monster

    Joined:
    Oct 22, 2010
    Messages:
    560 (0.41/day)
    Thanks Received:
    83
    Location:
    Latvia
    I watched this video, and it says that:
    "heads floats on cushion of air, created of the spinning discs - an absolutely essential cushion that prevents heads from crushing the disk." - the same as in that document, but the document says that if discs stops spinning and head is still above data surface, it can touch. It is logical, because then no cushion exists.
  2. theonedub

    theonedub habe fidem

    Joined:
    Apr 30, 2009
    Messages:
    7,125 (3.75/day)
    Thanks Received:
    3,656
    This is like beating a dead horse, read my post above.

    The residual spinning of the platter after power loss is enough to maintain the air bearing long enough for the heads to park (there aren't any brakes so they spin and gradually slow down).
    Solaris17 and Swamp Monster say thanks.
    Crunching for Team TPU
  3. Frick

    Frick Fishfaced Nincompoop

    Joined:
    Feb 27, 2006
    Messages:
    10,384 (3.40/day)
    Thanks Received:
    2,084
    And Swamp monster, the PDF says exactly the same thing.
  4. ivicagmc

    ivicagmc

    Joined:
    Mar 17, 2010
    Messages:
    176 (0.11/day)
    Thanks Received:
    27
    Here is something interesting I found I wont to share... Didn't know that there is lubricant in HDD :)
    HDD operation is heavily influenced by temperature. Bearings, both in the spindle motor (either BB or FDB) and in the rotary actuator (BB) rely on lubricants. The lubricant's viscosity, like any lubricant, changes with temperature. Thicker when cold, thinner when hot. Colder temperatures are more problematic, since the increased viscosity significantly "stiffens" the bearing, resultant drag is high and motors may not be able to achieve their minimum operating speeds. But high temperatures can cause the lubricants to fail, with obvious consequences.

    In addition, mechanical dimensions of the entire HDD assembly will change as a function of temperature (the HDD components are primarily comprised of Aluminum and Stainless, which have different coefficients of thermal expansion. This usually isn't so much of a problem, since the mechanical tolerances allow for this, but it can result in reduced performance. And significant temperature excursions can cause glue bonds (bearing cartriges are usually glued together) to fail and/or result in brunelled bearing races.

    The recording heads are also affected by temperature. The critical parameter is something called "Fly Height", the distance above the surface of the disk at which the "Slider" rides on it's "Air Bearing". Too high (usually colder) can cause write and read errors. Too low (usually, hotter) can result in contact between the Slider and platter, permanently damaging the HDD.

    The disk surface itself is coated in a fine layer of lubricant. It's designed to provide some protection if the recording head (Slider) contacts the disk surface. Temperature effects this lubricant like any other. In particular, at very high temperatures the lube can "spin off" the disk, reducing the effectiveness of the protecting layer.

    The magnetics and electronics all behave differently as a function of temperature. Usually, the HDD is designed to compensate for the effects. But operating the HDD outside it's temperature limits can exceed the range of compensation, typically leading to read and write errors.

    As a general rule, higher temperatures are worse in terms of HDD longevity but the HDD should generally work OK up until the point when something breaks. At cold temperatures the HDD may not work at all, but it is unlikely that attempting to get it going will do it any harm, it just won't work.

    Mechanical shock and vibration are the enemies of HDD reliability.

    The recording head consists of a "Slider", which flies over the surface of the spinning disk (think of an Air Hockey puck and you'll get the idea). The Slider contains the read and write elements that allow data recording (think of a Cassette Tape recorder and you'll get the idea). The Slider is attached through a "Suspension" to the rotary actuator. The HDD positions the Slider at different radii across the Disk by rotating (think of a Phonograph Tone Arm and you'll get the idea). The "Air Bearing" created by the flying Slider is fairly robust, but shocks in the Z direction can force the Slider to contact the disk surface, and that's bad.

    Desktop and Server drives use Contact Start Stop, the heads actually land on the disk surface when the disk is stopped. Mobile drives use something called "Ramp Load", the heads are lifted off the disk when the disk is stopped. For this reason, it is not a good idea to use a Desktop HDD in a mobile application. Desktop HDDs are prone to a kind of damage called "Head Slap" when stopped, and so are less robust when it comes to tolerating the types of mechanical shocks likely to be seen in a mobile application.

    Data is located on the disk surface in a series of concentric tracks. Each track is divided in to segments called "Sectors" (think of Pizza slices and you'll get the idea). To locate data on the disk, the HDD rotates the Actuator to the target track, and then it waits for the desired Sector to rotate under the Slider and past the read/write elements.

    The HDD has to control the Actuator very accurately, track widths are measured in microinches. Linear motion (in the X, Y, or Z direction) has very little effect on the HDD's ability to "Track Follow", the Actuator is balanced, but rotary motion (rotation around the Spindle Motor's Z axis) is very difficult for the HDD to handle. The HDD will monitor it's track position and abort read and write operations (and subsequently retry) if it's straying too far, so usually the worst effect of linear or rotary shock is reduced performance. However, very large shocks can result in mis-reads or mis-writes (the Slider strays on to an adjacent track while reading or writing).

    Vibration can excite resonances in the HDD's mechanical structure. Resonance has the effect of magnifying the vibration and it can result in the same kinds of errors and lost performance that large shocks will cause.
  5. Swamp Monster

    Swamp Monster

    Joined:
    Oct 22, 2010
    Messages:
    560 (0.41/day)
    Thanks Received:
    83
    Location:
    Latvia
    I really didn't read your post above, I guess I was writing at the time. I now understand how it moves heads away, and how it gets power to do sow, thank you for the information.
  6. theonedub

    theonedub habe fidem

    Joined:
    Apr 30, 2009
    Messages:
    7,125 (3.75/day)
    Thanks Received:
    3,656
    Oh I see, well glad the info helped. I just saw that episode of Decontsructed on TV the other night :D
    Crunching for Team TPU
  7. Swamp Monster

    Swamp Monster

    Joined:
    Oct 22, 2010
    Messages:
    560 (0.41/day)
    Thanks Received:
    83
    Location:
    Latvia
    I remember that some time ago (when IDE was popular) it was common knowledge that if HDD is moved from one house to another in winter, than we put it in a towel, so it don't get cold. If I remember right, sealing material would get hard and crack, allowing dust to get in.

    Anyway it is not a good idea to turn it on when it is cold, due to condensate danger:)
  8. Solaris17

    Solaris17 Creator Solaris Utility DVD

    Joined:
    Aug 16, 2005
    Messages:
    17,065 (5.25/day)
    Thanks Received:
    3,501
    Location:
    Florida
    changed but controller communication was not what i was talking about. tech like parking the control arm etc etc is good, but whoever wrote that that wont make a diffirence during power failure is right. What im talking about is physical NOT a argument passed to the controller. At a cerain point the heads will stop because of physical limitation. Almost like a support bar or a net under a trapeize wire. the head before hitting the platter will be stopped by the lack of ability to go further. or the arm being stopped by an indent in the plastic. I can draw pics if you want but I dont think they will. Manufacturers think of this stuff. They dont just assume electricity flows forever.
    Swamp Monster says thanks.
  9. Swamp Monster

    Swamp Monster

    Joined:
    Oct 22, 2010
    Messages:
    560 (0.41/day)
    Thanks Received:
    83
    Location:
    Latvia
    Thanks for expanded explanation! I didn't fully understand what you mean't in your first post, because you wrote that "it stops" and I wasn't sure if you mean platters or heads and what direction, because I tought that it isn't enough for head to stop if movement is in horizontal position. From this post I understood all, it's very good. So I guess longer posts is beter for understanding:toast:
    Solaris17 says thanks.

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)

Share This Page