While I am by no means an expert in electronics or Hall sensor use with different magnets configurations, the basic placement of the Hall over the four pole neodymium magnet is simply zeroed over the center of the flat face width of the magnet. The circuit used with the Hall can easily establish the center, along with the meter circuit or an adopted meter. The movement of the Hall is simply parallel to the magnet face and not toward or away from the magnet face. The distance between the Hall and magnet is rather one which has to be determined by the user for their individual degree of sensitivity/gain they need or wish. Of course for the reader of this; the magnetic field junction cannot be seen, but between each half of the sides of the magnet their are actually two magnets joined together and it is here where the circuit "zero" is centered. The "zero" line runs vertically up/down the magnet.
For Robert and myself, we use a simple "magnetic viewer", (which he gave me acouple to use), which illustrate the field. This simple device is simply a flexible thin bag which contains a oil and very fine particles of iron filings, which when brought upon a magnet illustrates the junctions, edges of magnets quite well. Unfortunately, at the moment the viewer doesn't seem to be available for purchase, or at least, a known commercial source is not known.
The configuration we use with the four pole neodymium
magnet is for the most part; quite a departure from the normal magnet/s
configurations for use in other applications.
See:
Hall-effect IC applications guide (AN27701) by Allegro/Sanken Microsystems
The web site also covers their products and carrys
a large selection of notes, which can be much more helpful in hall use
understanding than this site can carry.
Allegro/Sanken
Microsystems
We have used only two slightly different configurations
for our seismometers. The simple
"one" magnet configuration of above, and the second,
which is simply using another magnet with a spacer inbetween, which effectively
boosts the sensitivity of the hall. The simple one magnet approach
works very well with normal "garden gate" type horizontals. For the
2 magnet approach; it works very well for the S-G type seismometer.
Their seems to be a precaution though with the 2 magnet approach, and that
is that there is a real need to have the Hall sensor closer to one of the
magnets, than to the other magnet, i.e., don't center the hall exactly
inbetween the two magnets. The reason for this, is that the two magnets
fields do have an interaction magnetic junction exactly inbetween the magnets,
and in this area, the Hall sensor will not show as much sensor diversion.
To clarify further, if you can imagine an invisible "+" with the pair of
magnets set on edge with a spacer inbetween; the parallel junction to the
length of the magnets is kind of a dead zone; while normally the line running
vertically is the real "zero" we would want the sensor to be most exposed
too.