Practical Guidelines for building a Magnetometer by Hobbyists

Part1 : Introduction to Magnetometer Technology

Practical Guidelines for building a Magnetometer by Hobbyists

Part1 : Introduction to Magnetometer Technology
W. Bayot
Version 1.2
22 June 2008
Practical Guidelines for building a Magnetometer by Hobbyists .................................................................. 1
Part1 : Introduction to Magnetometer Technology........................................................................................ 1
1.
Foreword ............................................................................................................................................... 2
2.
Introduction to the Magnetic World ....................................................................................................... 3
3.
What magnetometers are used for and by whom................................................................................. 4
4.
Types of magnetometers ...................................................................................................................... 7
5.
Basic physical principles used by a PPM ............................................................................................. 8
6.
System configurations of PPMs ........................................................................................................... 9
7.
Functional Blocks of a typical PPM..................................................................................................... 11
8.
Sensor Types ...................................................................................................................................... 13
8.1.
Single Solenoid Coil.................................................................................................................... 13
8.1.1. Fluid inside Coil Variant............................................................................................................. 13
8.1.2. Coil inside Fluid Variant............................................................................................................. 13
8.2.
Double Solenoid Coil .................................................................................................................. 13
8.3.
Multiple (> 2) Solenoid Coils ....................................................................................................... 14
8.4.
Toroid Coil.................................................................................................................................... 14
8.5.
Cylindro-Toroid Coil ..................................................................................................................... 15

10/11/08

Practical Guidelines for building a Magnetometer by Hobbyists

1. Foreword
The concepts and facts described in this paper are not new and they are probably quite obvious to
specialists in geomagnetism. However, they were quite mysterious for the standard hobbyist like myself.
This description tries to demystify some of these, making the difference between the critical working
conditions and the auxiliary aspects only subject to some optimization.
It should be taken by the reader as MY current understanding of the most important concepts required for
the building of a magnetometer. I am quite sure that I missed some points and that I even misunderstood
others. In advance, I apologize for any possible incoherence or misunderstanding and I am completely
open to any remarks and comments from knowledgeable people.
This was the result of reading a lot of papers on the Web (the best and the most accurate being the paper
of James Koehler available here) and numerous experiments that I have done in the last months but also
a lot of precious information from my temporary or permanent project team members, more specially:
My French friend Philippe Marie with whom I designed and implemented an all-digital PI
metal detector with discrimination properties.
Steve Brekke, my best Californian partner who designed our current PPM amplifier chain and
guided me on the routes of FFT calculation.
Joe Geller, my occasional partner near Albany,NY with whom I exchanged very interesting
messages comparing his design with ours.
John Oldham, a precious Canadian partner with a long experience and actual practice of
magnetometers in the field (the sea). He has given me so much good advice during my long
period of unsuccessful testing, when the damned precession signal was not showing up.
Carl Moreland, our well-known mag technical site and forum hosting partner who has
gathered so many pieces of data giving so many people a playground around which to share
their experiences (good and bad) on the technical aspects of all kinds of treasure hunting.
The version 1.2 of this document reports the result of the experience acquired during the last two years on
this subject.
I want to specially add my thanks to:
James Koehler, my best Canadian partner and now friend with whom I spent the last two
years improving our PPM project putting our two long past experiences in common for the
sake of this hobby project.
Paul Cordes, my second partner working with me on an alternative PPM instrument
implementation selecting different hardware and software solutions. See more about that in
Part 3 of the whole document.
In advance, I am grateful to those who will accept to proofread this paper and remove its most visible
conceptual and syntactical bugs.

10/11/08

Practical Guidelines for building a Magnetometer by Hobbyists

2. Introduction to the Magnetic World

The Earth's magnetic field can be thought of as having three components:

down in the direction of gravity

to true North (not magnetic North)
and an East-West component at right angles to true North

The East - West component causes what is called declination, or the compass pointing to "magnetic
north" instead of true North. The true North component is in the horizontal plane, but the down
component is much stronger than the North component!. If a compass is not level then it will have a large
error because it is starting to read the down component of the Earth's field. Thats the reason why
conventional compasses should always be kept perfectly level.
The total value of the vector resulting from these three components depends on the geographical
location with its maximum values around the two Poles and minimum values along the Equator - look at
figure 4 on page 9 of the ampm-opt.pdf document.
The direction of this total vector is North-South with an inclination from the horizontal also depending
only on the location on Earth (50 to 70 North in Europe and Northern America but 0 on the Equator) look at figure 3 on page 8 of the ampm-opt.pdf document.
If you want to see the nominal value of the earth magnetic field in your region, go first on this link:
How to Get Your Local Field Value.pdf
Other sources of earth magnetic field modeling are the following:
http://geomag.usgs.gov/geomag/geomagAWT.html
http://www.ngdc.noaa.gov/seg/geomag/data.shtml
http://www.ngdc.noaa.gov/seg/geomag/jsp/IGRFGrid.jsp

Some underground or underwater objects or structures with permanent or thermo-remanent magnetism

are able to locally change these average value and direction. If you bury a permanent magnet and you
measure the earth magnetic field over and around it, you will find that the field greatly concentrates in its
immediate vicinity. Actually, any ferromagnetic object or matter will produce the same effects as well but
with much less magnitude than a magnet.
Magnetometers are all measuring the natural or human-made disturbances of the earth magnetic field.
These disturbances are called magnetic anomalies.
Important Note:
There is some confusion in the literature over physical units used in magnetism. SI units are now the
preferred units over the older CGS. In the older CGS system (still very much in use in USA), the
magnetic flux density was expressed in GAUSS (G) while in SI system, it is expressed in Tesla (T) or,
more practically, in nanoTesla (nT = 10-9T) with the following relation between the two:
1 Gauss = 0.0001 Telsa or 1 Gauss = 100,000 nT.
As an approximate reference, the nominal earth magnetic field is on average around 0.5 Gauss or 50,000
nT (while a small modern permanent magnet can easily give 3000 Gauss!!) with its maximum values
around the two Poles (70,000 nT) and minimum values along the Equator (25,000 nT).
Actually, the important measurement units in geophysics applications are Field Gradients expressed in
nT/meter. This is the variation of the absolute or vertical field value measured between two points distant
of one meter. Magnetic anomalies are detected by the value of the field gradient they produce when
surveying around and above them; higher the gradient, bigger the anomaly. It also means that small
anomalies require very sensitive instruments and also a large number of measurements close to each
other. There is absolutely no chance to detect an anomaly of 1nT with an instrument whose sensitivity or
resolution would be 10nT and an anomaly or object with a dimension of 10cmx10cm will probably be
missed with a 5mx5m search grid.

10/11/08

Practical Guidelines for building a Magnetometer by Hobbyists

3. What magnetometers are used for and by whom

The largest natural time-dependent magnetic anomalies are produced by magnetic storms
produced themselves by variations of the solar wind. A magnetic storm is a perturbation of the Earth's
magnetic field, caused by solar disturbances, usually lasting for a brief period (several days) and
characterized by large deviations from the usual value of at least one component of the field. They
can be predicted at FIXED meteorological stations by measuring the trends of the absolute values of
the earth magnetic field. The Dst is a geomagnetic index that monitors the worldwide magnetic storm
level. It is constructed by averaging the horizontal component of the geomagnetic field from midlatitude and equatorial magnetograms measured by meteorological stations from all over the world.
Negative Dst values indicate a magnetic storm is in progress, the more negative Dst is the more
intense the magnetic storm. One of the main practical applications of these measurements is the
prediction of near future radio-communication problems as magnetic storms disturb the ionosphere,
but some powerful magnetic storms have also disturbed the electrical power grid in some countries.
Outside of magnetic storms, there are other natural time-dependent slight and slow variations of the
magnetic field, they are called diurnal variations while smaller and quicker variations are called
micro variations. Thats the reason why the measurement of the space-dependent TOTAL
magnetic field (or one of its vertical or horizontal vector) is not sufficient to make actual field surveys
(See later on the chapter on PPM System Configurations)
This is a very good introduction of all these concepts:
http://geomag.usgs.gov/intro.php
This is a source of real magnetogram samples:
http://www.intermagnet.org/apps/dataplot_e.php?plot_type=b_plot
This is the plots you can get from there from any Intermagnet Observatory in the world. These are
from the DOURBES node in Belgium.

Natural space-dependent magnetic anomalies are caused by the juxtaposition of soils and
stones with highly contrasting magnetic properties produced by earlier volcanic activities. These
characteristic features are detected by geologists to find specific underground or underwater minerals
or metals (even gold since gold deposits are usually accompanied with "black sand"-magnetite
deposits) or, in petrology to locate gas or oil pockets kept within porous stones.
Artificial (man-made) magnetic anomalies are generated by:
o Buried or sunken objects containing a large proportion of ferromagnetic metal (iron or steel)
o Soil disturbances like filling holes with a soil of different magnetic characteristics than the
local ground or unfilled holes like tombs or wells.
o Burned features like fire pits, baked clay objects, ashes
o Permanent magnets buried specially to be easily located later on by geologists or
archaeologists.
10/11/08
4

Practical Guidelines for building a Magnetometer by Hobbyists

5
Currents circulating in the ground generate powerful and fast varying anomalies. These
currents are generated by powerful electrical activities like railways or power stations. These
anomalies can be felt kilometers away from any railway or power lines.

Thus, magnetometers are useful to:

o Public Works to locate pipelines or the steel bars of concrete foundations
o Geologists to locate coal, water, minerals, gold, oil and gas
o Environmentalists to find buried barrels of dangerous or polluted products
o Archaeological Research to locate :
old wall foundations,
refilled wells
tombs
old paths
Pits
Hearths
kilns & furnaces
Cooking sites
Midden deposits
Ditched enclosures
Trackway
o Military and Humanitarian deminers to locate and destroy UXO (bombs, mines, grenades)
o Sunken Ship Seekers to find cannons, anchors which help to locate the ship remains
themselves
o Treasure Hunters looking for iron-made containers (box or case) of precious material
o Relic Hunters looking for old man-made iron objects

In summary, magnetometers are able to detect the presence of Ferro-Magnetic material (Iron, Steel, Cobalt,
Nickel, Magnetite,) because they disturb the value and direction of the earth magnetic field.
However, magnetometers are COMPLETELY insensible to any other metals, they can not be used as metal
detectors even if they are sometimes used as complements in the research of targets.
Magnetometers also can not easily detect deeply buried (or sunken) small objects except with a highly
sophisticated instrument. Look at the figure 46, page 43 of the ampm-opt.pdf document, it shows that 1
pound of pure iron (not exactly a small object, equivalent to an ordinary hammer head) could be detected at 3
meters BUT would only generate a field gradient of 1nT. If you only own a mag with a sensitivity of 10 to 25nT
(typical of fluxgate and differential mags), then the same mass of iron could only be detected at 1 meter or
less IF and only IF your line of survey passes right above it. It is useful to remember that the field produced by
a given magnetic moment falls off as the inverse cube of the distance, not the inverse square.

10/11/08

Practical Guidelines for building a Magnetometer by Hobbyists

10/11/08

Practical Guidelines for building a Magnetometer by Hobbyists

4. Types of magnetometers
There are now a number of types of portable magnetometers based on different physical principles:
Instruments based on Magneto resistive and Hall effect sensors are not used to
measure the earth magnetic field even if they are sensitive to the value of magnetic
fields. They are used in the industry to detect ferromagnetic objects at short distances
or to measure higher magnetic fields. We shall not go in more details about them
here.
The fluxgate sensor has a special magnetic core (high-Mu material) being saturated
within the magnetic field to measure a single axis vector of this field. The basic
principle is to compare the drive-coil current needed to saturate the core in one
direction as opposed to the opposite direction. The difference is due to the external
field. Full saturation is not necessary; any non-linearity will do. As the core
approaches saturation, the signal picked up in the sense coil will show the nonlinearity. Flux-gate magnetometers are widely used in geophysical research to
monitor the earths magnetic field at observatories located throughout the world. At
each observatory, three separate flux-gate magnetometers are used; one oriented in
each of three orthogonal directions. For amateurs, there are single flux-gate
magnetometers commercially available; the most common of these are manufactured
in England and which have an output in the form of a frequency-modulated waveform;
the output frequency is proportional to the magnitude of the component of the earths
magnetic field which is parallel to the sensor. Amateur magnetometers built using
these elements usually are built in the Gradiometer Configuration to measure
Vertical Field Gradients. In actual practice, they can detect field gradients of no better
than about 100nT.
The standard Proton Precession magnetometer is to be discussed in more details
in the following chapters. In short, the protons of a proton-rich fluid are magnetized to
align in the same direction and stop their normal precession. When they are released,
they resume their precession in phase and induce a voltage at a frequency
depending upon the ambient magnetic field. This frequency is measured and the
corresponding field value is calculated. This type of mag can give sensitivities better
than 1 nT with a relatively low sampling rate of 1 measurement every 1 to 3 seconds.
An Overhauser PPM uses a proton-rich fluid mixed with a very small quantity of a
special chemical component containing free radicals. A constant magnetic field of RF
frequency (around 60MHz) is applied on the fluid. This produces what is called the
Overhauser effect providing the same polarization effect from a polarizing field much
smaller than in standard PPM and in much less time. The net effect is that
Overhauser mags take 4 times less power consumption than normal PPMs and can
give up to 100 times better signals and faster sampling rates (more than two
measurements per second).
Cesium is the most widely available optically pumped alkali vapor magnetometer. It
can give sensitivities better than 0.1 nT with a very fast sampling rate but this
technology remains completely out of reach from standard hobbyists for its
complexity and the cost of required special material and professional adjustment
instruments.

In summary, the types of mags that can practically be built by hobbyists are essentially
the Fluxgate and the PPM.

10/11/08

Practical Guidelines for building a Magnetometer by Hobbyists

5. Basic physical principles used by a PPM

Normally, proton particles contained in fluids turn around their main axis but
this axis also oscillates at a slower precession speed (a spinning top shows this
very clearly) whose value uniquely depends on the value of the ambient
magnetic field. In a normal status, the precession movement is not
synchronized between protons and thus has a canceling effect.
A sensor filled with a proton-rich fluid is placed in a coil at right angles to the
ambient magnetic field and a strong current energizes the coil, creating a
strong magnetic field - the so-called polarization field, which orients all the
protons of the fluid in the same direction and stops their precession. When the polarization current is
sharply ended, all the protons resume their normal precession but now, IN PHASE and thus, the
accumulated action of all these very small magnets induces a VERY SMALL (at the Volt RMS level)
Audio Frequency Sine Wave signal in a signal pick-up coil (which could be the same as the
polarization coil). This sine wave signal has a frequency depending ONLY on the TOTAL value of the
local magnetic field vector and lasts a few seconds (the relaxation time) during which the precession
of the protons still stays in phase. This frequency is called the Larmor frequency. This voltage is
amplified and its frequency precisely measured to determine the ambient field strength.
The exact relationship between ambient magnetic field (B in nT) and Larmor frequency (f) is the
following:
B = f * 100,000 * 2 * PI / 26751.2 or B = f * 23.4875
f = B * 0,04256
For example, for an average ambient field of 50,000nT, f will then be 2.128 KHz
From the above relation, it is quite obvious that a difference of 1 Hz in frequency corresponds to a bit
more than 23 nT. If a system should be built with a resolution of 1nT, then the Larmor frequency
should be evaluated with at least an accuracy of 1 part in 50,000 or approximately, a precision of 0.01
Hz to have any chance to be successful. Moreover, depending on the proton-rich fluid that has been
used in the sensor, the duration of this signal with sufficient RMS amplitude will only be from 1 to 5
seconds as it exponentially decays. Thus, the frequency evaluation must be made as soon as
possible after the polarization cut-off to get its maximum amplitude (and maximum SNR). However,
this RMS voltage will not exceed a value in the order of 1V. Its value will mostly depend on the fluid
being used, on the polarization current and on the inductance of the signal pick-up coil.
There is a long list of proton-rich fluids that can be used in PPM sensors but, in practice, some liquids
are usually rejected because they are potentially dangerous to manipulate. The most usual liquids
being used by hobbyists are:
Simple distilled or de-mineralized water (better boiled to remove dissolved oxygen longer
relaxation time) giving a comfortable relaxation time of 3 seconds but subject to freezing.
Methyl, Ethyl Alcohol, Kerosene (lamp or stove oil), Diesel Fuel or charcoal lighter fluid
with a shorter polarization and relaxation periods of 0.5 second but not subject to freezing
Automotive windshield washer fluids
Benzene with a longer relaxation time of 4 to 5 seconds but rather difficult to find by nonprofessionals.
You can see here the exponential decay of the magnitude of a typical PPM signal generated after its
polarization current has been cut off. Listen to it here.
This is its equivalent frequency spectrum (FFT) with its high and very narrow peak showing that it is
a pure sine wave with some background noise. Its Signal-To-Noise Ratio (SNR) is around 35dB at
2KHz, this is not too good but, if we remove a good amount of this noise from the raw signal (with
analogue or digital filtering) , we get this FFT result. As you see, the Signal-To-Noise Ratio (SNR)
jumped up to 90dB. Now, listen to it here and hear the difference.
The magnitude of this signal was rather low but with a higher gain amplifier, this is what the digital
filtering should give as result.

10/11/08

Practical Guidelines for building a Magnetometer by Hobbyists

6. System configurations of PPMs

A single sensor PPM measures the total ambient earth magnetic field value and displays and/or logs
it in units of nanoTesla. This simple configuration is being used in fixed geomagnetic stations spread
all over the world as a network.
A PPM to be used for field surveys must be built with portable power supply and it must be as light as
possible to be easily carried. As already explained, the detection of magnetic anomalies is based on
the measurement of field gradients and not of total field values. Thus, using a single sensor PPM
would theoretically and simply mean to subtract total field values taken consecutively at survey grid
distance to get the required field gradient values. This simple method could indeed be used for
surveying small area, provided that the consecutive measurements are made very close to each other
in time. Why is that? Remember that the total earth magnetic field value slowly changes during the
day (diurnal variations). (Note that anyway, during periods of magnetic storm, it is better to stay at
home.)
If the required precision of the results is not too high (e.g. during a first overview survey of a large
area), these variations would not influence the interpretation too much; otherwise, they will disturb the
gradient values over the whole length of the survey line.
One possible solution to this problem is to take two measurements at each crossing point of the
survey grid, one high above the ground (e.g. 3 meters high) and the next, immediately after, closer to
ground (e.g. 1 meter high). The difference of distance of 2 meters will be sufficient to produce a field
gradient if a magnetic anomaly is produced under this point. It is clear that this method is only valid for
ground surveys and cannot be applied for underwater search.
Important Note: Never make measurements with the sensor(s) lower than 0.5 meter from the ground. The best results would
be that your values would be too much influenced by very small ferro-magnetic particles at the surface and the worst would be
that your signal would completely disappear.

A Differential PPM uses two sensors mounted on the two extremities of a long staff and connected
in series to a single amplifier chain. The signals of the two sensors add together, are amplified and
feed headphones. If the area being tested has a constant magnetic field, the precession frequency
induced in both sensors is the same and the operator hears a gradually decreasing in amplitude but
steady tone which disappears within the relaxation time. In the presence of an anomaly, however, the
magnetic field at each sensor is not the same and signals of different frequency are produced. As
these signals are added together, they interfere to produce a wavering tone, increasing and
decreasing in amplitude until it finally dies out. This is an amplitude beat whose frequency depends on
the field gradient between the two sensors; the more beats are heard per unit of time, the stronger is
the magnetic anomaly causing them. This configuration is the simplest and lowest cost you can find
but, because the measurements are non-quantitative and because the human ears are used as tone
amplitude measuring devices, it is also quite crude in sensitivity. You could not expect this
configuration to detect any anomalies creating gradients much lower than 25nT. Why is that?
Because the imperfect human ears are good at detecting tiny changes of frequencies but are very bad
at detecting amplitude changes (Thats the main reason why the vertical scale of Audio Spectrum
Diagrams is logarithmic = dB).
This is an example of signal going out of such a configuration. It is the signal sample posted by Phil
Barnes on his web site. It gives a beat of around 3 Hz corresponding to a gradient of about 70nT. phil
barnes.wav. As you can see on this PCM stream picture, the magnitude of the signal shows some
clear beats for about 1 second.
This FFT curve of the same second of signal duration shows a large and relatively wide peak at
around 2KHz.
The best configuration used by professionals to survey large area with great precision is made of two
complete and identical magnetometers:
o One is put on a local recording base station, i.e., diurnal station monitor, near but outside
the surveyed area. This mag records, with precise time stamps, the total earth field values,
including its diurnal variations, during all the duration of the survey.
o The second one is the mobile station making the real measurement on the crossings of the
survey grid and recording them with time stamps as well.
At the end of the survey, the two data stream are merged based on the time stamps and the field
values captured at the fixed station are subtracted from the raw measurement values captured at the
same time, thus canceling the bad effect of the time-dependent variations of field. The net result is a
10/11/08
9

Practical Guidelines for building a Magnetometer by Hobbyists

10
3-D map of field values of the surveyed ground relative to the field value of the fixed station. Note
that these values are NOT gradients.
To show the difference, take for example a rectangular anomaly crossed by a survey line.
In this configuration, when crossing over one limit of the anomaly, the values will change of
level up or down depending on the kind of anomaly and will stay at this level until the survey
operator crosses the other limit of the anomaly.
In a gradient measurement configuration, you will detect a positive peak of gradient when
passing over and another negative peak when leaving the surface. The gradient data
displayed as a 3-D map needs more experience to be interpreted than the first. (See more
details in the ampm-opt document)
A slightly more sophisticated configuration uses a simplex radio data link to communicate between
the two mags in real time. This avoids the final post-processing and allows verifying the quality level
(unexpected level of noise or possible operating errors) of the measurements while they are made.
1. A special gradient measurement configuration is called a Gradiometer. It is also made of two
identical magnetometers whose sensors are attached to each extremity of a long staff usually held
in a vertical position by a survey operator. In this configuration, the polarization and signal pick-up
cycles are synchronized on the two mags but the resulting signals coming from the two sensors are
kept in separate amplifier chains and frequency evaluation processes. The two total field values are
then subtracted to get a net and clean gradient value representing, without any time-dependency, the
difference of earth magnetic field as sensed by the two sensors, one closer to ground and the other
farther by 1 to 3 meters.
In summary, all system configurations (except the differential mag which only requires a single amplifier chain)
require two complete and identical magnetometer systems to be built. However, one of those is enough to
start the testing and experimentation of the amplifier, the polarization control and various types of sensors.

10/11/08

10

Practical Guidelines for building a Magnetometer by Hobbyists

11

7. Functional Blocks of a typical PPM

Analog Power
Supply

Sensor
Coil(s)

AntiRinging
Circuit

Polarization
Control

Sensor
Tuning

Low noise
Pre-Amplifier

Band-Pass
Filter

Amplifier
Analog
Output

Gel Cel
Polarization Battery

Test PC

Digital Power
Supply

Zero-Crossing
Counter/
Timer

Comparator

Analog
Output

Local
Recording
&
Display

Digital
Output

PC Link

Basic Proton Magnetometer Modules

2. The Polarization Control module switches the polarization battery to feed the sensor(s) with a large
DC current, then switches off this current as fast as possible. An adjustable automatic timing
mechanism can trigger the switching durations or the program of a micro-controller or a PC can
control them.
3. The Anti-Ringing circuit absorbs the ringing produced by the large inductance of the sensor coil(s)
before the polarization control connects the sensor to the pre-amplifier for the signal pick-up phase.
4. The Sensor Tuning Capacitors are put in circuit after the polarization cycle and during the signal
pick-up phase. They are carefully adjusted to tune the sensor coil at the nominal precession
frequency of the local region and thus, act as a supplementary signal amplifier as well as a filter of
unwanted frequencies.
5. The Low-Noise Pre-Amplifier Stage is a very critical part of the system. It has to increase the V
level signal magnitude up to the mV level without introducing any signal distortion nor any internal
noise and preferably, without amplifying too much the external electro-magnetic noise naturally
picked-up by the coil of the sensor and the various cables and wires.
6. The center frequency of the Pass-Band Filter is adjusted to the nominal precession frequency of the
local region and its bandwidth set to a few hundreds of Hz around it.
7. The Amplifier stage increases the sine wave signal magnitude to the minimum level required by the
following signal processing stages.
8. The analogue Output gives a sine wave signal sufficient to be studied in details by PC-based Audio
Editing and Spectrum analyzing tools through any integrated audio card.
9. As a first simple solution, the sine wave signal is converted into a clean square wave and its signal
level adapted to TTL level by a Comparator stage.
A micro-controller counts a pre-defined number of zero-crossings of the square wave signal and
very precisely measures the exact time it takes. This allows calculating its precise periodicity and
thus, its frequency. The frequency is then converted in magnetic field value (expressed in nT units) by
a simple multiplication by a constant value.
10. The signal processing method described in 9. is rather simple to implement but it is very sensitive to
noise, more specially to spikes. A single missing or extra period changes the result by 1/n with n
being the total periods used to make the measurement. For example, if one uses 256 periods, 4% is
lost for reach missing or extra period. As a more sophisticated solution giving much more protection
10/11/08

11

Practical Guidelines for building a Magnetometer by Hobbyists

12
against noise, we have experimented a so-called Phase Shift algorithm. The principles of this
method will be described later but the tests made up to now show that this method is immune against
a low SNR.
11. The current magnetic field value is shown in real-time on a local display and appended to a vector of
values in local memory.
12. A serial line allows the downloading of the vector of measurements to a PC for further study,
analysis and plotting.

10/11/08

12

Practical Guidelines for building a Magnetometer by Hobbyists

13

8. Sensor Types
8.1. Single Solenoid Coil
This is the simplest sensor type and the easiest to build by Hobbyists even if it is also the one picking
up most of the electromagnetic noise around except if it is shielded inside a large box.
A cylindrical form is selected with a inside diameter fitting a small non-magnetic recipient like a glass
or polythene bottle containing a proton-rich fluid.
A relatively thick magnet wire is wound on this form. The wire diameter and the number of turns are
not very critical, usually around 1500 to 2000 turns of wire AWG 16 to 22. The DC resistance should
preferably be kept under 10 ohms to get a DC polarization current of more than 1 Amp from a 12V
gel cell battery. If a higher voltage battery is being used, it is possible to wind the coil with thinner wire.
The Inductance of the coil should be around 30 to 50 mH. The magnetic field generated in the coil
(and on the protons of the fluid) by the polarization should be a minimum of 100 Gauss.
A Solenoid coil sensor calculator has been written by Jim Koehler and posted by Carl Moreland. It
calculates the approximate inductance, resistance and polarization current given the dimensions of
the coil, the voltage of the battery and the diameter of the wire. It also evaluates the length of wire to
plan. The other calculated parameters of the sensor will be discussed in more details later.
A Solenoid sensor should preferably be positioned perpendicular to the local vector of the earth
magnetic field. For regions where the inclination of the local field is greater than 45, the main axis of
the sensor should be horizontal while if <45, it should be held vertical. If the field has a (quasi-)
vertical direction, then the horizontal direction of such a sensor is not important but, in equatorial
regions, where the earth field is much more horizontal, the sensor should point to the North to get a
maximum magnitude of signal.
Note that if a Solenoid coil is positioned in the general direction of the local earth field, the precession
signal will be null.

8.1.1. Fluid inside Coil Variant

This is the traditional method of polarization, a bottle is inserted inside the coils and its
contained fluid is subjected to the powerful field generated by the coil.

8.1.2. Coil inside Fluid Variant

This is a more professional version where the coil is wound first around an hollow form and the
whole winding is blocked and electrically isolated with epoxy. The final coil is then introduced
inside a bigger container filled with the proton-rich fluid.
The advantages of such a configuration are:
a better cooling by the fluid itself of the heat generated by the polarization current
the supplement of protons activated by the polarization field (not only inside the coil but
also around it).
Each of the following sensor configurations can also be made in this variant.

8.2. Double Solenoid Coil

The single Solenoid Coil type of sensor is picking all the electro-magnetic noise around. The very high
gain of the amplifier chain of a PPM and its voltage limiting diodes connected to its input makes it a
relatively good radio receptor with an antenna made of the sensor coil. A double coil sensor can be
built to reduce this noise and produce a canceling effect on signals coming from the outside of
the coil. There are two possible coil arrangements:
1. Two identical single coils sensors are built as described above and kept very close to each other
(glued side by side) in order to sense the same magnetic field. They are connected in series to
add their precession signal voltages and, at the same time, to cancel the effect of noise voltages
from outside the coil arrangement.

10/11/08

13

Practical Guidelines for building a Magnetometer by Hobbyists

14

2. Two separate coils are wound on the same axis around the same bottle. There must be a wide
gap between the two coils because the two polarization fields cancel each other there. They are
connected in series and in opposition (wound in the opposite direction) to get the same effect
Coil1

Coil2

Bottle
Sides
Connecting
Wires

as in 1. Above.

8.3. Multiple (> 2) Solenoid Coils

This configuration is used by a number of commercial systems. It is made of 3, 4 or 5 coils glued
together at regular angles to make a complete circumference. A three-coil system is made of coils
glued together and shifted by 120 while a five-coil system is made of coils shifted by 72.
The purpose of this rotation of the coils is to eliminate the signal dip due to local magnet fields and the
angle of the sensor in relation to it. As one coil is null, the others take over. The coils must be wound
in the same direction and connected in series.
Three or more separate solenoid coils as pictured also create a self-shielding effect call (phase
canceling) or Hum Bucking.

8.4. Toroid Coil

A toroid coil has many advantages over the solenoid like its natural insensibility to external noise
and positioning direction. However, it is more hard manual work to wind and it potentially contain
less volume of proton-rich fluid.
The same coil sensor calculator can be used to calculate its functional parameters.
Jim Koehler describes the details of construction of such a sensor from the largest ring on the FisherPrice Rock-a-Stack toy in his paper starting at page 36.
A slightly larger model can also be built from eight 45 PVC tubing bends glued together. This method
has the advantage of making much easier the winding of the sensor. Look at the building phases one,
two and three. Between the phases two and three, the two halves of the toroid are glued together, the
core of the sensor is filled with the fluid and the two still unwound gaps of the toroid are wound.
Corrective note about these two last pictures: The aluminum foil wrapped as pictured around the
toroid should NOT be used as shielding. A toroid seems to be naturally protected against external
disturbances and thus, does not require any shielding.
A toroid sensor should also be positioned perpendicular to the local vector of the earth magnetic
field. For regions where the inclination of the local field is greater than 45, the main axis of the sensor
10/11/08
14

Practical Guidelines for building a Magnetometer by Hobbyists

15
should be vertical while if <45, it should be held horizontal. Its horizontal direction of such a sensor
is not important.
Note that if a toroid coil is positioned in the general direction of the local earth field, the precession
signal will still be half the maximum it could be when it is positioned in a perpendicular position.

8.5. Cylindro-Toroid Coil

A cylindro-toroid sensor has the same advantages and limitations as the simple toroid one but it has
an added advantage; it is able to contain a larger volume of fluid. This is an example of the stillunwound core of such a sensor.
In summary, the simplest sensor to build is the single Solenoid coil and it should probably be the first one to
try even if it picks up external electromagnetic noise. When the amplifier chain has been proven to work with
this sensor, the other types of sensor can then be built and tested to optimize the whole system. The natural
sequence is:
1. Single solenoid with internal bottle
2. Double solenoid with internal bottles
3. Triple solenoid with internal bottles.
4. Toroid or cylindro-toroid coil
5. Triple solenoid inside a container

10/11/08

15

Practical Guidelines for building a Magnetometer by Hobbyists

16

End of Document

10/11/08

16