Magnetic calibration

What Is a Magnetometer and How It Determines Heading

A magnetometer measures the strength and direction of the magnetic field around it. Inertial sensors use this measurement to determine heading, the same way a compass does: by comparing the measured field to the Earth's own magnetic field.

Earth's magnetic field isn't perfectly aligned with its geographic axis, and it isn't the same everywhere: its direction and strength change with location and drift slowly over time. To account for this, SBG Systems' inertial sensors rely on the on-board World Magnetic Model (WMM), a model of Earth's magnetic field, to know what the local field should look like at a given position and date. This is also why magnetic north and geographic (true) north aren't the same, the angular difference between the two, called magnetic declination, changes depending on where you are on Earth.

Why Magnetometers Are Sensitive to Their Environment

A magnetometer cannot tell the difference between the Earth's magnetic field and any other magnetic field nearby. Once the sensor is mounted on a vehicle or structure, nearby materials, metal parts, screws, wiring, motors, generate or distort the local field, and the sensor measures the sum without knowing which part comes from where.

This can introduce heading errors reaching several tens of degrees if left uncorrected.

Magnetic Interferences

Hard Iron Interference

Hard iron interference is caused by magnets, magnetized objects (screws, bolts), or high-current wiring. It adds a constant offset to the measured field, whatever the sensor's orientation is.

Soft Iron Interference

Soft iron interference is caused by ferromagnetic materials placed near the sensor, iron and some steels for example. Unlike hard iron sources, these materials don't generate their own magnetic field. Instead, they react to the existing field and distort it in a way that depends on direction.

Because the distortion changes with orientation, soft iron can rotate the measured heading by tens of degrees. This makes it harder to deal with than hard iron.

Fixed vs Moving Sources of Interference

Not all sources of interference behave the same way over time.

Fixed sources: materials that stay in a constant position relative to the sensor, such as a vehicle's own structure or a permanently mounted magnet.
Moving or variable sources: parts that move or whose magnetic field changes over time, such as rotor blades, motors, or wiring carrying variable current.

The first category creates a stable, predictable disturbance. The second creates one that keeps changing, which is far more difficult to deal with regardless of the sensor used.

As a general rule, keeping the sensor as far as possible (over 3 meters) from variable sources, such as motors or wiring carrying changing current, is one of the simplest and most effective ways to reduce this kind of interference.

Why Calibration Geometry Matters: 3D vs 2D

To compensate for hard and soft iron distortion, a calibration needs to sample the magnetic field across as many orientations as possible.

A 3D calibration, where the sensor is rotated freely through all axes (roll, pitch, and yaw), covers the full sphere of possible orientations. This lets the calibration characterize the complete hard iron offset and the full soft iron distortion in every direction, giving the most accurate correction.

A 2D calibration only rotates the sensor around a single axis, typically yaw, while it stays level. This only samples a narrow band around the equator of that sphere. Any interference that doesn't act evenly within that plane, for example an offset that varies with roll or pitch, is not measured and cannot be corrected.

Best results are always given when a wide range of orientations is covered.

Limited degrees of freedom

If the sensor cannot be rotated freely in all orientations, for example a fixed ground vehicle, or a sensor bolted in place with limited access, the calibration will still be accurate, but only for the orientations that were actually covered during the procedure.

Example: A large tracking antenna pedestal, such as a SATCOM or radar antenna, often has its azimuth rotation mechanically limited to around 300°, not a full 360°. A magnetometer mounted on it can only be calibrated within that same range, leaving a blind spot where interference remains uncorrected.