Compass in Smartphones — Essence, Working Principle and Application

What Is a Digital Compass in a Smartphone?

Digital compass in a smartphone is a sensor based on a magnetometer that measures the Earth’s magnetic field strength and determines the azimuth (angle relative to magnetic north). Unlike a mechanical compass, a digital compass uses microelectromechanical systems (MEMS) to detect the magnetic field. The magnetometer measures the field along three axes, allowing correct direction determination at any smartphone tilt.

How the Sensor Works

A modern MEMS magnetometer is a microchip measuring 2×2×1 mm, integrated into the smartphone body together with an accelerometer and gyroscope (as part of a 9-axis IMU). The sensor measures not only the direction but also the intensity of the Earth’s magnetic field (25–65 µT depending on latitude). The presence of metal objects and parasitic magnetic fields from speakers, vibration motor, and camera create distortions, so software correction of readings is necessary for accurate operation.

The first digital compasses in mobile phones appeared in 2009 with the release of the iPhone 3GS — the first smartphone with a built-in magnetometer. Today, magnetometers are installed in the vast majority of smartphones, including budget models. According to Yole Group, the MEMS magnetometer market for mobile devices exceeds 1.5 billion units per year, confirming the demand for this technology.

Working Principle of a Magnetometer

The core of a MEMS magnetometer is a sensing element that changes its electrical characteristics under the influence of an external magnetic field. Modern smartphones use the Hall effect or the magnetoresistive effect (AMR — Anisotropic Magneto-Resistance). The principle is based on the fact that a current-carrying material changes its resistance when exposed to a magnetic field, and the magnetic field vector is calculated from the voltage change.

Hall Effect and AMR

Hall effect sensors detect voltage changes at the edges of a semiconductor plate caused by the Lorentz force acting on moving charges. Magnetoresistive sensors (AMR) use a permalloy (NiFe) film whose resistance depends on the direction of the magnetic field. AMR sensors provide smaller size and higher sensitivity compared to Hall sensors, which is why they dominate in modern smartphones.

Magnetometer data is combined with accelerometer readings to correct tilt and determine the north direction in the horizontal plane. The sensor fusion algorithm uses the Madgwick filter or Kalman filter to combine data from the magnetometer, gyroscope, and accelerometer. This provides a stable and accurate azimuth even during fast rotations and device shaking.

Types of Magnetometers in Mobile Devices

Three main types of MEMS magnetometers are used in the mobile industry: Hall effect, anisotropic magnetoresistive (AMR), and giant magnetoresistive (GMR). AMR magnetometers are the most widely used in smartphones due to their optimal balance of sensitivity, power consumption, and cost. Leading manufacturers include AKM (Asahi Kasei Microdevices), STMicroelectronics, and Bosch Sensortec.

The sensitivity of an AMR magnetometer in a typical smartphone is 0.3–1 µT with a dynamic range of ±1000 µT (±8 Gauss), which is sufficient for accurately measuring the Earth’s field (30–65 µT) and compensating for parasitic fields. The sensor sampling rate can reach 100–200 Hz, but most applications use 10–50 Hz to save energy. Automatic compensation for temperature and housing stress deformations is implemented at the sensor driver level.

Digital Compass Calibration

Compass calibration is the process of compensating for parasitic magnetic fields created by smartphone components: speakers, vibration motor, camera, and battery. Without calibration, magnetometer readings can deviate from the true direction by 10–30 degrees. The calibration system determines the offset matrix (hard-iron bias) along three axes and applies it to correct all subsequent measurements.

Calibration Algorithms and Methods

The standard compass calibration method is drawing a “figure 8”: the user rotates the smartphone along a trajectory resembling the digit 8. During movement, the sensor records the magnetic field in all possible orientations, and the algorithm accumulates points and approximates a sphere with a center offset from zero. The sphere radius corresponds to the Earth’s magnetic field magnitude, and the center represents the parasitic offset vector. After approximation, the system calculates correction coefficients for each axis.

Modern calibration algorithms from Qualcomm and MediaTek perform continuous background calibration without user intervention. They analyze magnetic field changes during device rotations and automatically update correction parameters. If the sensor detects a new parasitic field (e.g., when attaching a magnetic case), the system initiates recalibration within 1–2 seconds of continuous device movement.

Compass Usage in Mobile Applications

Digital compass is used in a wide range of mobile applications. The primary use is in navigation and mapping services (Google Maps, Yandex.Maps, 2GIS), which determine the device’s orientation on the map to correctly display the user’s direction of travel. In augmented reality (AR) applications, the compass ensures correct placement of virtual objects relative to cardinal directions.

In astronomy applications (Star Walk, SkySafari), the compass combined with an accelerometer determines the direction the smartphone is pointing and shows the corresponding area of the night sky. Fishing and tourism apps use the compass to mark points of interest on the map with azimuth reference. GPS fitness tracking uses compass readings to determine running or cycling direction, improving route display accuracy.

Using the Compass via Android API

To access the digital compass on Android, the TYPE_MAGNETIC_FIELD sensor from the Sensor API is used. Developers register a listener for magnetometer reading changes and receive an array of magnetic field strength values along three axes in microteslas. To determine the azimuth, magnetometer data is combined with accelerometer data through a rotation matrix using sensor fusion.

import android.hardware.Sensor
import android.hardware.SensorEvent
import android.hardware.SensorEventListener
import android.hardware.SensorManager

class CompassManager(
    private val sensorManager: SensorManager
) : SensorEventListener {

    private val magnetometer = FloatArray(3)
    private val accelerometer = FloatArray(3)
    private val rotationMatrix = FloatArray(9)
    private val orientation = FloatArray(3)

    override fun onSensorChanged(event: SensorEvent) {
        if (event.sensor.type == Sensor.TYPE_MAGNETIC_FIELD) {
            event.values.copyInto(magnetometer)
        } else if (event.sensor.type == Sensor.TYPE_ACCELEROMETER) {
            event.values.copyInto(accelerometer)
        }
        
        SensorManager.getRotationMatrix(
            rotationMatrix, null,
            accelerometer, magnetometer
        )
        SensorManager.getOrientation(rotationMatrix, orientation)
        
        val azimuth = Math.toDegrees(orientation[0].toDouble())
        // azimuth: 0 = north, 90 = east, 180 = south, 270 = west
    }

    fun start() {
        sensorManager.registerListener(this,
            sensorManager.getDefaultSensor(Sensor.TYPE_MAGNETIC_FIELD),
            SensorManager.SENSOR_DELAY_UI)
        sensorManager.registerListener(this,
            sensorManager.getDefaultSensor(Sensor.TYPE_ACCELEROMETER),
            SensorManager.SENSOR_DELAY_UI)
    }

    override fun onAccuracyChanged(sensor: Sensor?, accuracy: Int) {}
}

The azimuth obtained through getOrientation is measured in radians and can take values from -π to π, where 0 corresponds to magnetic north. For on-screen display, values are converted to degrees (Math.toDegrees) normalized to the 0–360 range. It is important to account for magnetic declination — the angle between magnetic and true (geographic) north, which depends on geographic location and changes over time. Google Maps automatically adjusts declination through the GeomagneticField API.