Fluxgate

 Article 36322 (112 more) in sci.physics:

From: pad@galaxy.nsc.com (Paul Denny x8349)

Subject: fluxgate compasses

Keywords: how it works

Date: 4 Sep 92 16:51:43 GMT

Organization: National Semiconductor, Santa Clara

Lines: 70

I recently asked here how fluxgate compasses work and received the

following explanation by email from Franklin Antonio. I felt that

it was an elegant explanation that would be of interest to sci.physics

and as such I have reproduced his explanation. All credit for this

posting are due to him and any errors remain mine - flame me not him.

enjoy! (I did)

 

From: Franklin Antonio <ANTONIO@qualcomm.com>

In a recent sci.physics posting, you ask how fluxgate compasses work.

The trick that requires explanation is how you can get a coil to produce

a signal which is proportional to the strength of one component of a

STATIC magnetic field.  (ie the Earth's magnetic field)  Obviously, if you

can pull that off, you can do it twice, with two coils oriented 90 degrees

apart, then take the 4-quadrant arctangent of the two signal levels, and get 

the direction of the static field.  (or if you prefer doing it in 3d, you

can do the same with three coils.)

To make a signal which is a measurement of a static magnetic field, the

fluxgate compass gates the flux of the earth's magnetic field.  

In the absence of local magnetic materials, the earth's magnetic field

looks locally uniform.  That is to say the lines of force are curved

but gently so, so that locally you can think of them as uniform and

straight, like the lines on a piece of writing paper.  Now, consider

a piece of magnetic material of high permeability.  Perhaps a piece of

iron or ferrite.  Put this into the previously uniform field, and now

the lines of force bend, because the lines would rather go thru the

high permeability material than thru free space.  If we had a way to

gate the permeability of that hunk of iron, that is to vary it in a

regular patern high/low/high/low/... then we would make the lines of

force of the earth's magnetic field move in and out of the iron.  Now

we would have a varying (rather than static) magnetic field which

is proportional to the strength of the static (Earth's) magnetic field

which we were trying to measure.  Of course, now that it's varying, we

can measure it easily.

How to make the permeability of that magnetic material gate on and off?

Since permeability is in general a nonlinear function of total field,

you can do this with bias.  Consider a toroidal magnetic core (iron or

ferrite) with two coils wound on it.  One coil is for bias, the other

for sensing.  Put a large gated signal into the bias winding.  Here

large means large enough so that the nonlinear properties of the magnetic 

material come into play.  When the bias signal is off, the core has a 

high permeability.  When the bias signal is on, the core has a lower

permeability.  Now observe the sense winding.  On it, we will see two

signals.  One is the bias signal obviously, and we need to filter that

out.  The bias frequency is chosen to be high enough that we can easily

lowpass filter the sense output to eliminate the bias.  The second 

component in the sense winding will be caused by some of the earth's

magnetic field being alternately pulled into the core (by the permeability

of the core material), and displaced (when the permeability drops).

This signal is then rectified, and measured.

This is usually done using two (or three) cores, oriented at right angles,

so that the resulting signals are the components of the field in the

corresponding directions.

In a practical implementation, the bias might be something like a 100 kHz

sine wave, which is gated on/off at a rate of a few per second.  Bias level

would be many ampere-turns, so to put the core well into saturation when

bias is on.

Accuracy is limited by your ability to wind a very uniform (symmetrical)

sense coil, and other such non-ideal characteristics of the implementation.

There is a NASA patent.  I had a copy once, but no longer have it handy.

Check your favorite patent database, and search on "flux-gate".

End of article 36322 (of 36434)--what next? [npq]