Magnetometers

1. Fluxgate magnetometer
The most common type of magnetomers is fluxgate.

Two coils are winded around a ferromagnetic core. An alternating current is fed to the primary coil. If an
external magnetic field is present the induced field is limited to one direction by the hysteresis of the core
material. The difference between the inducing and the induced fields gives a signal that can be observed.
The signal is observed with the pick-up coil.

On the left is a ring-core

sensor which has lower
noise that the bar type.
On the right is the sensor
made by the Danish
Meteorological Institute.
Here the sensing
elements are set in a
marble cube and
compensating coils are
winded around the
fluxgate sensors.
 Ukrainian LEMI-025 magnetometer that record correct one-second values.

Resolution along each component:

at 0.1-second file and Flash card data 0.001 nT
Temperature drift < 0.2 nT/°C
Noise level at 1 Hz < 10 pT rms
Components orthogonality error < 30 min of arc
Components orthogonality error after calibration < 2 min of arc
Sample rate 1 per second
Frequency band for 10 per second data output DC – 3.5 Hz
Digital output RS 232
Power supply 12 V
 2. SQUID magnetometer
The superconducting quantum interference device (SQUID) consists of two superconductors separated by
thin insulating layers to form two parallel Josephson junctions. The device may be configured as a
magnetometer to detect incredibly small magnetic fields -- small enough to measure the magnetic fields in
living organisms. Squids have been used to measure the magnetic fields in mouse brains to test whether
there might be enough magnetism to attribute their navigational ability to an internal compass.

SQUIDs are sensitive enough to measure fields as low as 5 a T (5×10−18 T) within a few days of averaged
measurements.[1] Their noise levels are as low as 3 fT·Hz-½.
 3. Induction coils

Induction sensors (also known as search coils), because of their measuring principle, are
dedicated to varying magnetic field measurement.

Induction sensor constituting with a winding (orange) surrounding a ferromagnetic core.

Frequency band of received signals 1 –70000 Hz

Shape of transfer function linear - flat

Transfer function corner frequency 20 Hz
Magnetic noise level, pT·H-1/2:
at 1 Hz ≤ 5
at 10 Hz ≤ 0.2
at 10 kHz ≤ 0.005
at 100 kHz ≤ 0.01
Transformation factor error:
at flat part of band pass without edges ≤ ±0.3 dB
at full band pass edges and corner frequencies ≤ 3 dB
Power supply voltage ±(6...12) V
Mass 1.7 kg

5 Proton precession magnetometer

A proton magnetometer measure the total magnetic field strength and is not very sensitive to direction.

In proton magnetometers a direct current flowing in a solenoid creates strong magnetic field around a
hydrogen-rich fluid (kerosine), causing some of the protons to align themselves with that field. The current
is then interrupted, and as protons realign themselves with the ambient magnetic field, they precess at a
frequency that is directly proportional to the magnetic field. This produces a weak rotating magnetic field
that is picked with inductor, amplified electronically, and fed to a digital counter whose output is typically
scaled and displayed as field strength.

Sensitivity: 0.022 nT / √Hz

Resolution: 0.01 nT (gamma)
Absolute Accuracy: 0.2 nT
Dynamic range: 20,000 - 120,000 nT
Long term stability: <0.05 nT/year
GSM-90F1: 1 sample /1 sec.
GSM-90F5: 5 samples / 1 sec.
Power: 12V 200mA max., 40mA average
RS232C parameters: programmable

IMAGE magnetometer network

Calibration system
At the Numijärvi Geophysical Observatory we have an accredited magnetometer calibration system. It is
used to calibrate three-component fluxgate magnetometers. I will describe the Nurmijärvi Magnetometer
Calibration Facility (abbreviation NuMCF) and present a calibration of one satellite magnetometer of
Lusospace, Lisbon Portugal. The NuMCF is part of the magnetic calibration and test laboratory (NuMCTL) of
the Nurmijärvi observatory of the Finnish Meteorological Institute, comprising of magnetometer and sight
compass calibrations and compass swing base measurements at airfields.

The three axes coil system is shown in Fig. 1 and consists of three sets of four square coils with side lengths
from 1.6 to 2.2 meters. The frame is made of aluminum. Two inner coils have 22 turns of 2 mm diameter
copper wire and two outer coils have 42 turns of 1 mm wire. The coils are placed on a 70 x 70 cm2 top of
concrete coming up to the floor level of the calibration room. A pillar, made of glass bricks and a marble
plate on top of it, is standing on the concrete basement. The pillar serves as a non-magnetic stable base for
the tested instruments.

Theoretically, the Alldred and Scollar (1967) coil system of this size can produce uniform fields with errors
less than 0.001 % in a volume with diameter of about 30 cm at the center of the system. A calculation
based on real dimensions of the Nurmijärvi coil system, give 18 cm for the corresponding diameter for Y-
component, 25 cm for X and 30 cm for Z.
Fig. 1: Calibration coil system of the Nurmijärvi Geophysical Observatory.

The magnetic directions of the coil system were determined with a DI-fluxgate, a non-magnetic theodolite
having a fluxgate sensor fastened to its telescope (Kring-Lauridsen, 1985).
Fig. 2. Angle errors between the orthogonal magnetic fields in the center of the coil system and 15 cm to
north, east, south and west (above) and 18 cm above and below the center (below).

The coil constants were measured by using a proton magnetometer in the centre of the coils. The Earth’s
field of two of the three components was first compensated to a value close to zero. Then large positive
and negative fields were generated in the third component and the field in the centre of the coils was
measured. With this method the coil constants can be measured with accuracy of ±0.002 %. This
measurement is done once in every year and the values (Nov. 2006, 20 °C) were in [nT/mA]: X: 42.401, Y:
48.275 and Z: 36.970. The coil constants have temperature dependencies of -0.0025 %/°C for X component
and -0.0020 %/°C for Y and Z components.

Fig. 3. Funtional diagram of the magnetometer calibration system.

Often the exact knowledge of the orientations of the magnetometer sensor's magnetic axes with respect to
the sensor's mechanical axes is needed. With this information the sensor can later be installed e.g. in a
satellite so that the true magnetic directions of the magnetometer sensor are known with respect to
reference directions of the satellite. At NuMCF a theodolite can be installed on a pillar outdoors 60 m south
from the coil system and along the magnetic South axis of the X-coils. By using a light beam from the
theodolite (see Fig. 4) and a mirror fixed to the magnetometer sensor (in the centre of the coil system), the
orientation of two mechanical axes of the sensor can be measured with accuracy better than ±0.03 °. The
third axis is measured with a spirit level.

The temperature variation is generated in a thermally insulated box (Fig. 5) that has a square hole in the
bottom for a marble cube. The heating elements are fixed to the marble and the tested magnetometer
sensor is standing on the cube. The purpose of this arrangement is to avoid tilting of the underlain pillar
due to heating or cooling. The box can be heated up to +60 °C and cooled by using dry ice down to about –
30 °C.

The estimation of the uncertainty of the measurements is based on the EA-4/02 publication of the EAL
Committee 2 (1999) of the European co-operation for accreditation. The combined standard uncertainties
for the current measurement are:

u(IX)= ±0.0036 mA
 u(IY)= ±0.0034 mA
u(IZ)= ±0.0049 mA
And the standard uncertainties of the coil constants:

u(SX) = ±0.00092 Nt/mA

u(SY) = ±0.00099 Nt/mA
u(SZ) = ±0.00109 Nt/mA
The uncertainty budget for the X-component is as follows:
Quantity Estimate Standard Probability Sensitivity Contribution to
X-comp. uncertainty distribution coefficient stand. uncert.
xi u(xi) ci ui(y)
I 80mA 0.0036mA U-shape 0.0063nT/(mVmA) 0.000023nT/mV
S 42.40nT/mA 0.0009nT/mA normal 0.012mA/mV 0.000011nT/mV
U 6784mV 0.06mV normal -0.00007nT/mV2 -0.000004nT/mV
B 3392nT 0.16nT normal 0.00015 /mV 0.000024nT/mV
T 0.5nT/mV ±0.000035nT/mV
±0.007 %

After making an uncertainty budget for all the components we get the standard uncertainties of the
sensitivity coefficients:

u(TX) = ±0.007%
u(TY) = ±0.007%
u(TZ) = ±0.008%

The EA-4/02 publication for calibration laboratories states the use of an expanded uncertainty, obtained by
multiplying the standard uncertainty u by the coverage factor k = 2. The assigned expanded uncertainty
corresponds to a coverage probability of approximately 95 %. Here the multiplication with k = 2 gives the
expanded uncertainty of ±0.02 % for the transformation coefficients of all the components.

u(T) = ±0.02 %

For the angles between the components of a three-component magnetometer the corresponding
uncertainty is ±0.02 degrees.
Data processing
At the Observation Services of the Finnish Meteorological Institute all 10-second data from the IMAGE-
stations is checked manually. Timing errors and external disturbances are removed.

There after the data is transferred to the open database where anyone can fetch it.
https://ilmatieteenlaitos.fi/avoin-data