Specific, but basic, help for a magnetics newbie!

[agarrison]

I'm an astro engineer by training, but I'm working on a personal project dealing with magnetics and need some extremely basic assistance in knowing where to start looking for the engineering approach. Please forgive my fundamental language and description of the engineering involved.

I'm trying to create a function which will permit me to do the following:

I'll be using a 3-axis magnetic sensor and need to determine distance, velocity, etc. from the sensor to a "wire" of straight permanent magnetic material (assume 2 dimensional) as it approaches from an unknown angle to the axis of the wire. Closest approach (if it in fact crosses the source) will always be the same - about 1/2" from sensor to magnetic source.

Sensing the B field seems like the best approach, and I can derive a lot from measuring sensor readings (strength and direction) at two different times when approaching/departing the field crossing point. Velocity vector of the sensor will be known.

How do I find or create a function to determine the predicted field strength at a given sensor location? The wire could be many feet long, and I could cross over it in close proximity to one end or the other (ie: flux will be at a "steep" angle). I will probably know the strength of the magnet.

By the way, any suggestions for a material for the source "wire"? Permanent, not easily demagnetized, and as inexpensive as possible within reason.

Many thanks.

 

[prex]

Not sure to understand your setup: what you call a wire is a cylinder with the magnetic poles at the ends?
If the answer is yes then you'll get almost zero (unmeasurable) induction everywhere (also along the cylinder, close to it), except in the close proximity of one of the poles.
Also the choice of materials depends on field strengths you need to get, and finally on what you intend to measure and to what purpose.
You should be more specific on the points above.

prex

http://www.xcalcs.com

Online tools for structural design

 

[sreid]

Prex is correct. The field pattern is the same as end to end on a long cylindrical coil. Most of the magnetic circuit path is air so it's impedance is high and the magnetic field strength will be low almost everywhere. Some physics books show how to calculate this.

Since the circuit impedance is high, a large (length) magnet is not required. Buy a NeDyFe Rare earth magnet say one inch long and a half inch in diameter and attach 1/2 inch diameter steel bars to both ends.

 

[logbook]

Perhaps I am understanding what has been said differently to the other guys. Do you mean you have a long single electrically conducting wire through which you are passing a current? If so then the magnetic field would have cylindrical symmetry, falling off with the reciprocal of the radius from the centre of the wire. Putting a sleeve of "magnetic material" around the wire will have no effect on the field. The field strength in terms of H will be

H= current/(2*PI*R)

 

[agarrison]

prex, sreid and logbook - thanks for your replies. All were illuminating in their own ways. As I said, I know almost nothing about this stuff, and it shows.

Your replies prompt the following clarification and questions. The entire concept may be technically not feasible, and I'm very open to hearing this, if it's true.

1. The basic setup is for a magnetic material set into the floor about 1/8" below the surface. The floor is non-magnetic, non-metallic, etc. The "strip" of magnetic material can be as long as 6-8 feet. Numerous (up to 10-12) parallel strips would be about 1/2 inch apart, and the poles of the magnetic strips would vary to create a sort of "magnetic bar code" where the 3 axis sensor could keep track of the pole reversals (mag strip to mag strip). The angle that the sensor crosses the strips can vary from close to parallel to perpendicular.

2. (logbook) No electrical connections at all to the floor stripes. They would have to be permanent magnets.

3. (sreid) Interesting idea about the magnet at the center of two long steel bars at the ends. I'm going to think and study more about this possibility.

4. (prex) No cylinder is envisioned. The "wire" was originally conceived as a L-O-N-G (6-8 foot) thin solid bar, could be rectangular or circular in cross section.

Whatever the arrangement, when the sensor takes a reading, I need to determine when a set of strips is nearby (don't need to determine distance, just magnitude of the field in excess of the earth's effect of about .6 gauss), and when the sensor crosses the parallel stripes, I need to determine when the fields reverse, ie: when the sensor moves from on top of a N-S field to a S-N field.

With further apologies for the REALLY elemental descriptions, I sincerely appreciate the help. If anyone needs to know about Kepler equations or LaPlace transforms, or ion propulsion, it might salve my ego a bit!

Thanks,

 

[sreid]

With the application info provided I think what you need is bar magnets polarized with North and South poles on the long faces. These could be made in short piece (4-6 inches) and installed end to end with the N or S pole facing up. From a quarter inch (say) the field strength will be much more than terestrial magnatism. Ceramic magnets would probably be the most cost effective. Take a look at
Magnetic Component Engineering.

http://www.mceproducts.com/

 

[IRstuff]

Since the strips are only 1/2" apart, they must be read from much less than 1/2" away, otherwise the field from the adjoining strips would interfere with the measurement. This means that the magnets must actually be relatively weak.

TTFN

 

[agarrison]

In response to:

sreid: Good suggestion. This could be a very useful approach.

IRstuff: Thank you for the evaluation. The 1/2 inch between magnetic strips is not set - it just needs to be as close as possible. The sensor height will be no closer than 1/2" to the magnetic material. And I don't need absolute strength, I can sense "refersals" vs "non-reversals" of poles. Additive field strength will be sufficient for this deduction - kind of a binary bar code.

I'm finding this thread extremely helpful, and I appreciate all of your inputs.