Department of Energy Argonne National Laboratory Office of Science NEWTON's Homepage NEWTON's Homepage
NEWTON, Ask A Scientist!
NEWTON Home Page NEWTON Teachers Visit Our Archives Ask A Question How To Ask A Question Question of the Week Our Expert Scientists Volunteer at NEWTON! Frequently Asked Questions Referencing NEWTON About NEWTON About Ask A Scientist Education At Argonne Hard Disc Magnets and Holders
Name: Ed
Status: other
Grade: other
Location: CA 
Country: N/A
Date: 10/16/2005

I am a computer tech for a local school district and over the years I have taken hard disks apart to see what makes them tick. From reading some articles on this site and others has been helpful but one question has not been asked. When one takes the magnets off the hard disk, normally two of them, top and bottom, only one side of the magnet (holders) is magnetized. The outside of the magnet holder is not magnetic like the inside (there a some exceptions to this). If I remove the magnet from the holder, it does not seem as strong a magnet as with the holder. How does this magnet holder do this magic?

Hi Ed-

When you look at a use of magnets, you have to also look at the passively magnetic iron they put around it. It is a lot like a "conductor" for magnetic fields, and it can conduct at least 4 times more magnetic flux per unit cross-section area than any of our permanent magnet materials can ever put out.

So it has uses:

a) steering the flux to a certain place and gap-shape

b) funneling, concentrating the flux from ~3000 Gauss at the magnet face to >15,000 Gauss at a small narrow gap. (In vacuum, 1 Tesla is 10,000 Gauss.)

c) conducting the flux from one magnet to the next with much lower opposition than air would present.

d) "keepers": when magnets push a field out into the space around them, the space pushes back.

Over time, the magnet material can let some of it's polarization slip out of it's grasp, and it gets weaker.

Iron short-circuits reduce this magnetic back-pressure, helping the magnet last much longer.

I have seen those disk-drive magnet structures, they and their removed arc-shaped slab magnets are very common in electronic hobbyist outlets. They are the field magnets for the head-positioning "motor" or "actuator" (actually a hefty galvanometer-movement) used by the control chips to move the read-head arm radially across the disk, to go from track to track and to hold the read-head accurately on a given track.

There are no mechanical or magnetic stopping-points in this arm-swinging actuator. just a continuously adjustable amount of left-right torque, operating over a long frictionless swing range. The control chips in the disk-drive change the coil-current moment-by-moment to whatever strength helps the read-head stay in the center of the desired data track.
|    <==  <=-  <=  <-   |
| vv ____^___^___^___^__|
| vv |  ----------------
| vv | |SSSSSSSSSSSSSSS|  "fringe
| vv |> --------------- \  field"     (P=pivot)    (read-head)
| vv |   ^   ^  ^   ^  ^               P
| vv |   ^   [c+c+c+.^......c-c-c-]------P------------R
| vv 
|   ^   ^  ^   ^  ^  (c=coil)     P 
| vv |< --------------- 
/                                   platter        (axle)
| vv |___________________
| vv     ^   ^  ^   ^  |
|    ==>  -=>  =>  ->   |     v,^,->,<- : direction of 
|_______________________|                magnetic flux 
    iron "pole-piece"                    lines; they make 
                                         a closed loop.

(Imagine I have impressionistically flattened out a couple of 90-degree angles with respect to the arm.) (The real iron does not need to look quite that thick.)

By being a low-resistance path from top to bottom, the iron pole-piece provides a fairly well-contained conduit for the magnetic flux between the outside faces of the two magnet slabs. You can see that the outside face on top is N and on bottom is S, so it is energetically favorable for lines to go from one to the other.

If there was no iron:

1) these "return-flux" lines would arc far out into the air around the edges of the magnets.

2) the effective air-gap opposing the magnets would be much longer, centimeters instead of millimeters, and the field crossing the actuator coil would be roughly 3 times smaller. The motor would be at least 3 times weaker for a given coil current, and your disk drive would be hotter, heavier, slower, and power-hungry when changing tracks.

It is conceivable they could polarize the magnets and coils with a different arrangement, which might look like this:

  iron "pole-piece 1"
     |    ->   =>  ->   |
       |NNNNNN   SSSSSS|
       |SSSSSS   NNNNNN|
        ----------------         (P=pivot)
         ^   ^  v   v              P
         ^   ^  v   v     (c=coil) P
       |NNNNNN   SSSSSS|
       |SSSSSS   NNNNNN|
      |  ^   ^  v   v  |
      |   <-  <=  <-    |     v,^,->,<- : direction of 
      |_________________|                 magnetic flux lines
    iron "pole-piece 2"

Then no top-to-bottom conduit is needed; the "return-flux" is only sideways in the top and bottom pole-plates. Even so, the iron pieces short some of the flux, make the field stronger, and mostly keep the magnetic field from spreading far and wide. Flux containment is valuable to prevent unintended side-effects (like erasing magnetic data on its own platter), and also helps the motor to always have the same strength even if somebody puts iron or a magnet just outside the box.

"When one takes the magnets off the hard disk, normally two of them, top and bottom, only one side of the magnet (holders) is magnetized."

This confuses me, it is not quite what I would expect. Perhaps I am not taking your meaning right.

Possible explanation: A further reduction of the magnetic path might be to magnetize the magnets in this pattern:
       |   <-  <=  <-  |
       |SSSSSS   NNNNNN|
          ^         v
       |NNNNNN   SSSSSS|
       |  ->  =>   ->  |

...although I think this takes lots of extra magnet material (thicker), where an iron plate would be thinner and cheaper and probably better.

The real benefit might come in if they use even more cycles:
        [ c+  c-  c+  c- ]------

Then the coil winding path starts to look almost zig-zag instead of circular loops, and less magnet thickness is needed, and the original magnetization is easier to accomplish because the length of the flux path in the permanent magnet material is shorter. It takes a very high-power pulse into an electro-magnet to magnetize those strong Neodymium magnets, so the magnet-makers appreciate an easier job and charge less. The ratio of air-gap to magnet length is lower, though, so maybe the field is a little weaker. it also would have much less travel before the torque would decrease or reverse. Electronics might then reverse the current, almost like running a stepping motor. And two overlapping coils would be needed, staggered a little with respect to each other. All told, I am not sure they ever do it that way.

You will have to feel out the N/S polarity zones on your magnets and see what part of all this fits.

Jim Swenson

Hi Ed,

I would think the magnet holder is made of soft iron, which has a very high permeability so the magnetic field much prefers going through the holder than air. If the holder is roughly a horseshoe shape, the magnetic field is guided to the gap between the ends of the horseshoe. This is then placed next to the disk so the field is strongest where it is most useful.

This naturally leads to a smaller field at the back of the horseshoe and a weaker field (because it is more spread out) when the magnet is removed from the holder.

Best, Dick Plano,
Professor of Physics emeritus, Rutgers University

Click here to return to the Engineering Archives

NEWTON is an electronic community for Science, Math, and Computer Science K-12 Educators, sponsored and operated by Argonne National Laboratory's Educational Programs, Andrew Skipor, Ph.D., Head of Educational Programs.

For assistance with NEWTON contact a System Operator (, or at Argonne's Educational Programs

Educational Programs
Building 360
9700 S. Cass Ave.
Argonne, Illinois
60439-4845, USA
Update: June 2012
Weclome To Newton

Argonne National Laboratory