Do it Yourself Hard Disk Repair

Add Your Comments

Only for the brave – Eric Shufro, Barry L. Failor, Jr.

Purpose:

To demonstrate a method for replacing the read / write head from a
Maxtor d540x-4k020h1 (20gb 5400 rpm, single platter) hard disk.

Disclaimers/Caveats:


1) I take absolutely no responsibility for any damage accrued to any piece of
hardware while attempting this procedure. It is 100% completely at your own risk. Opening the drive WILL VOID YOUR
WARRANTY!

The likelihood of success is STILL UNKNOWN since I have yet to receive a new controller.
However, if you read on, you may come to realize the feasibility of such a task. If anyone has a Maxtor d540x-4k020h1 and is willing to sell it, please email me – Thank You.

2) Opening a hard disk and exposing it to the air may cause permanent damage to the data stored within.

3) The alignment of the permanent magnets may pose an issue. They need
to be aligned exactly right. Take notice of the upper magnets position before
you unscrew it.

Problem Description:

A friend of mine who runs a decent size manufacturing company had kept various pieces of data
on this drive. Being only one year old, and part of his Dell computer, he never thought twice to back anything up.

Needless to say, the drive crashed without warning. Under close observation, I noticed that the drive didn’t
spin when powered up.

It seemed that either the controller was damaged, or the spindle motor had possibly burned out.

The data wasn’t critical and most of it just needed to be typed back in. The finances
weren’t available to send the drive out to a data recovery center, their work often costs between $1500 to $5000 dollars USD. He had decided to forget about it unless a less expensive alternative could be found.

I was definitely up to the challenge; especially knowing there was nothing to lose.

Troubleshooting:

I began by inspecting the controller for obvious damage (i.e. burnt parts etc). At this time, I was unable to see a problem.
However, I did remember that the drive did jitter when power was first applied, so I knew there was power on the spindle
motor, at least for part of a second.

Furthermore, if power was present on the spindle
motor, and it jittered momentarily
then stopped, then the motor was either stuck, or perhaps the applied voltage had dropped.

I briefly opened the drive just to ensure the spindle motor wasn’t jammed. It wasn’t.

Then I closed the drive
and began to troubleshoot an electrical problem. I removed the controller board from the drive and systematically masked off with masking tape the 20 pins or so between the controller and the internals of the hard drive.

After trying several pins,
I managed to discover a pin which in fact, when covered (and the controller back on the board) allowed the drive
to spin up properly. Unfortunately, the drive would not be detected by the BIOS, nor did there appear to be any head control. The drive would spin then slam the head side to side against the head stop inside the drive. This was
definitely not good.

It seemed this pin had some vital role in controlling the head movements. After re-opening the drive, I traced the faulty pin to its destination, a small chip on the side of the armature carrying the head.

BPA1
BPA2

There was no way I could repair this – UNLESS…


Eric Shufro

Solution:

Cannibalize parts from another hard drive to get the first one functional.

I purchased an identical drive on eBay. However, I first tried the easy way out. I moved the new controller on to the old drive and powered it up.

BIG MISTAKE. The new controller
immediately cooked itself. Now I was definitely sure that the chip mounted on the side of the head armature was directly responsible for dropping the voltage applied to the spindle motor. I imagine this pin shorts the controller which then stops applying voltage to the spindle motor. So I was exactly back where I began except that now I had two burned controllers.

I decided to do what had to be done and move the head from the good drive to the old drive. Moving the platters seemed impossible and obtaining the right tools would not be easy.

Required Tools:

  • Needle Nose Pliers.
  • Philips Head Screw Driver.
  • Small Flat Head Screw Driver
  • T9 Torx Head Screw Driver.

General Tips Before Starting:

  1. Attempt to work in the cleanest area possible. A garage is usually not a good choice.
  2. While working within the drive, work slowly and carefully. Be sure not to touch the platters with your fingers or tools! This will most likely damage the data on them.
  3. The goal is to remove the heads from the good drive and then later transplanting them into the broken drive. Again, we are moving the heads since the bad chip is affixed to the head armature which
    directly connects to the actual read / write head. Since there is little chance of removing the chip
    successfully, I am attempting to move the entire head / arm mechanism instead.

Eric Shufro

Procedure:

  • First, remove the screws holding the controller board to the back of the drive.
    Bottom
    The controller is the green thing
  • Pull the controller off.
    Controller Off
    Drive with controller off
  • Flip the drive over.
    Closed
    Time to unscrew
  • Remove the Philips head screws holding the drive cover in place. There are two T9 screws under the stickers, the hold the top of the spindle head armature bearings in place. Remove them. Remove any additional screws holding down the cover until the cover becomes loose.
  • Inside
    Don’t worry about the print, you’ll get a closeup later
  • Take the cover off of the drive and set it aside.
  • Remove the two screws holding the pin bracket that attaches through the casing to the back side of the controller. Do not try to remove this yet.
  • Remove the single screw holding down the top magnet above the head armature.
  • Inside
    Here’s your closeup
  • VERY CAREFULLY, use a small flat head screw driver to pry the magnet off of its seat.
    MagOff
    “Will this erase the charges on my credit card?”

    WARNING! These magnets are very strong. Do not allow them to get any closer to the platters than they are while positioned in their mounting places. THIS WILL DESTROY any data on the platters!

  • Set the magnet down away from the drive and prepare your needle noise pliers. . . .
    {mospagebreak}

    Eric Shufro

    The next step is to remove the head locking mechanism.

    HSAL
    HSAL2
  • Carefully squeeze this piece of plastic gently, but use a moderate amount of force to lift it off its mounting pin. It takes some effort, but it will come right up. Be careful! Both this plastic device as well as the one in the next step are both very
    fragile and can break easily if squeezed too hard!
  • Next, remove the head stop mechanism, located as shown in the picture just above. Again, squeeze gently and lift it straight off of the pin.
  • Gently slide the head off the platter. It will sit nicely in the air without the platter underneath it. DO NOT touch the heads with your fingers or any other material. This may damage them.
  • HSAL2
  • Now, prevent the head from moving by holding your finger above the bearing on top.
  • Finger
    Giving it the finger
  • While doing this, turn the drive over. You will see a
    Torx screw just below the head (below the controller you removed). Remove this screw carefully. This is the last screw holding the heads into the drive chassis.
  • Lift the head straight out of the chassis. When it is clear, gently pull on the connection block you removed the screws from earlier. It will lift straight out of the chassis with the heads.
  • Out
    Getting (hard drive) head
    Gutted
    This is what’s left
  • Take a deep breath, you are half way there.
  • {mospagebreak}

    Eric Shufro

    Now, do the exact same procedure to the drive with the bad heads.

    Once both sets of heads have been removed, simply put
    the good set of heads in the old drive chassis. The installation process is very similar to the removal process.

    When you insert new heads into the old chassis, be sure to screw the bottom of the head armature in while the heads are not over the disk.

    Then gently lift on the arm carrying each head (in this case there was only one) to slide it back over the disk. At this point you should test the arm by swinging it back and forth very slowly and carefully. It should not drag on the disk whatsoever.

    Reinsert the drive stop and head lock, and then the permanent magnet.

    Re-close the case, and screw the controller back on. Now you should be finished! Plug the drive in and you should be set to go.

    Final Caveats

    This is a last-ditch means to get data out, and you shouldn’t count on this working. If you absolutely, positively MUST get that data out or you’ll lose thousands and thousands of dollars, look at a data recovery center.

    If the data isn’t worth THAT much, and you’ve exhausted all lesser options, you might want to try this, but there’s a lot of things that can go wrong, and there’s no guarantee or even probablility everything will be OK if you do this.

    It’s an interesting project, though. {mospagebreak}

    Eric Shufro

    Having worked for a major hard drive manufacturer repairing
    damaged/destroyed hard drives for a living, I have one small note about the
    procedure taken to repair the drive:

    The read/write heads are more sensitive than can possibly imagined. There is a term that should be introduced here: Head Slap.

    Sounds like the heads smacking together violently, but its much less than
    that. If the top and bottom heads for a single double sided platter ever
    touch, they are damaged to some degree. Now here’s the fun part: You may
    not actually damage the reading/writing element, as they are the size of a
    spec of dust in the middle of each head. However, the heads are made of a
    silicon material, much like IC chips.

    The heads have an amazing role in the
    drive. As the platters spin, the disk surface generates a small “film” of
    air that spins in the same direction as the platter. As the air passes
    under the head, the design of each head is such that it forces a very small
    amount of air under the head and gently lifts the head off the surface of
    the disk, ever so slightly so that it doesn’t ride on the surface. (BTW, the inner most ring of the disk surface of all platters is called the landing pad.)

    In order to keep the heads from flying too high off the disk surface, a
    small amount of downward tension is built into the armature to provide a
    spring effect. When you slide the armature across the disk surface, you
    introduce small, probably microscopic, scratches into the surface of the
    platter. You also wear away a very small amount of the silicon of the head.

    Now when you make it to the edge of the platter, the two heads are no longer
    separated by the platter and slap together ever so gently. In the
    microscopic world, the slap is like smacking two skyscrapers together.
    Maybe they don’t splinter into a million pieces, but infinitesimal chips and
    cracks appear in the surface of the head. Not necessarily the read/write
    element, but definitely in the silicon around it.

    When you put the device
    back together, you introduce stress and over time – maybe a week, maybe a year – the surfaces scratch, wear down and the head doesn’t fly anymore. New term: Head Crash – when the heads simply grind the surface of the platter to dust while committing hara-kiri and turning the inside of the drive case into a black powdery mess.

    I’ve seen it – pretty cool – but not good for
    storing or retrieving data for that matter, as the data is now separated into little particles and splattered all around the inside of the drive case.

    (Another BTW: By scraping the magnetic film off the platter and randomly
    distributing it around the inside of the drive case, you effectively render
    the disk unreadable, which no amount of money will enable a drive recovery
    company to bring back. Food for thought if you need to truly delete files
    from your hard drive.)

    So how do you do it without destroying the drive? Build yourself a comb
    system that will spread the heads apart and lift them off the platter
    surface before you move them at all. We use a series of metal fingers that
    are tapered to get under the armature and lift the heads.

    Then swing the
    armature and comb assembly out and remove, but if you take the comb out –
    Head Slap! If we ever allow heads to touch, they are in the trash can a
    second later. The point is, you can’t see the damage done any it may not
    be catastrophic, but it will lead to early drive failure.

    Just thought I’d let you know. You could probably put together something to
    accomplish the head lifting armature swinging maneuver, but be careful.
    Make sure it securely holds the heads apart while rotating them out from the
    platters and keeps them apart while you make the switch.

    And never let any thing touch or slide along the surface of the platter!

    I’ve included a very simple quick sketch of the comb I’m talking about:

    Diagram

    Also, the whole cleanliness thing is crucial.

    Not that you can’t do it at
    all, just be very sure that no dust or small particles enter the case.
    Imagine if one of your eyelashes hit the platter surface. If it’s there
    when you spin up the drive, it might get wedged under a head and accomplish
    pretty much the same thing as a chipped head. Remember that black powdery
    mess?

    The following (TM Microscopes) link shows what happens to the physical disk when a drive is shocked (not electric, physical) – a different kind of head slap occurs.

    In this case, the head lifts off the media and smacks back down with enough force to deform the magnetic layer of the disk. Sometimes you can see them as anywhere from 1 to 4 dots in a rectangular pattern on the surface of the disk. The four images on this web page show a visible image next to the magnetic signature of the same area.

    Their point in showing this is that the little lines of magnetic force are actually altered by the collision. When you alter those lines, you corrupt data. So even though it may not be major, you can easily lose data by simply dropping a HDD. And it doesn’t take much to do that.

    When I was working at the manufacturer, we played around with a mock-up of an empty HDD outfitted with accelerometers. Those are the little guys that tell you how much force is being exerted on a object. (Same thing they use in crash test dummies.)

    We’d play with them and see how high we could drop them without exceeding a given value. (Memory fails me and I can’t remember what it was.) Typically, it only took a drop of 3/4-1 inch onto a hard surface to see damage.

    Suspension systems are much better now, but still don’t absorb a lot of the shock. And once you dent the surface of a platter, it’s like putting a divot on a putting green. Part of the material is depressed, while the rim raises up above the surface.

    Spin up the disk and the head will smack it every time it passes over that spot. Eventually you get a head crash. Another thing to be careful with.

    Barry L. Failor, Jr. Studies Engineer – POWER Engineers, Inc.

    Leave a Reply

    Your email address will not be published. Required fields are marked *

    You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

    Discussion
    1. Your best bet is to create a thread of your own to get more exposure than a 3 year old thread, but to give a quick answer, have you tried the drive in a PC? Perhaps the enclosure is what is out of whack. So if you attach it to a different power and data source, perhaps it will work... Give that a try.
      Really appreciate this article. I'm checking out an external hard drive (Seagate 3.0) that's stopped working. When I apply power, the drive spins, the armature moves back and forth several times, then it shuts down...any thoughts?
      can u plz tell me what is the problem with my hard disk. here is the video like of the opened hardisk.

      http://www.youtube.com/watch?v=P7cDZkAtTlw
      You definitely can't release them by hand - just exposing the platters to dust or letting dust into the housing that holds the platters can be a big problem. Touching the platters will leave fingerprints, oils, or residue, which could be problematic when the little read/write head passes over that section at high speed. Once the platters do spin up, they spin very fast - a spec of dust at high velocity banging around your platters and your microscopic read/write head could spell catastrophic failure.

      From the sounds of what you described, I would consider your hard drive dead. It sounds like the logic board is fried, and after replacing the board it still isn't recognized in BIOS - if replacing the logic board worked, the board should display in BIOS even if the mechanical side of things still has a problem. Since it isn't recognized in BIOS, that means there is another problem you haven't identified yet.

      Also, it sounds like there is a mechanical failure with the platter mechanism seizing up, which indicates a failure with the motor or bearings of the drive. That is another sort of problem that can't be fixed at home.

      If you really need the data, your best bet at this point is sealing up the drive and pursuing options with data recovery companies who have clean rooms to recover data. That can be expensive, so if its not really important, your best move may be to write it off as a loss.

      Sorry I don't have much good news, but I hope this helps understand the situation.
      I am owning a Samsung 7200 rpm, 320GB hard disk and its not showing up in the bios after the power failure. I identified the damage in its board, so changed the board with another similar board and then identified that the platters inside the hard drive have pasted or locked together. The lock doesn't seems to get released even though the hard disk starts on after changing the board and again the hard drive is not showing up in the bios. Please advice me on how i am gonna release the platters, is it OK if i release it manually with my hand or is there any other alternative methods to recover data?