Skip to main content Skip to search Skip to main navigation

Knowledge · Navigation & GNSS

Antenna azimuth error: causes, cost and how an IMU helps

By the AMCORIS application teamNovember 20264 min read

Nazimuth errorIMU

QUICK ANSWER

In antenna pointing, an IMU does three jobs: it measures tilt against gravity with millidegree resolution, it stabilizes the beam on moving platforms, and it holds the heading when GNSS drops out, for example only about 0.13° of drift in 10 minutes at 0.8 °/h bias instability. The azimuth error itself comes from four places: the north reference, tilt of the mount, mechanical misalignment and drift over time. North comes from dual-antenna GNSS or a surveyed reference, because no MEMS gyro of any make can find north alone; that takes fibre-optic or ring-laser gyros.

0.13°heading drift in 10 min during a GNSS outage – M-G370PDG, 0.8 °/h bias instability
0.0014°tilt equivalent of 24 µG accelerometer bias instability – M-G370PDG
1.1 dBloss of a 1.2 m Ku-band dish at 0.5° azimuth error, 30° elevation

01

Where does an azimuth error come from?

Azimuth is the horizontal pointing angle measured from north. For a satellite dish, a radar, a directional radio link or a mobile mapping sensor, every error source ends up in the same number. Four sources dominate:

  1. 01North reference: the heading the system believes in. It comes from a compass, a gyrocompass, dual-antenna GNSS or a surveyed mark.
  2. 02Tilt of the mount: if the azimuth axis is not vertical, the error grows with elevation (section 04).
  3. 03Mechanical misalignment: offset between the sensor axis and the antenna boresight. It is constant and can be calibrated once.
  4. 04Drift: the heading wanders while no external reference is available, mainly through gyro bias.

02

How much does an azimuth error cost?

A pointing error θe reduces the gain of a directional antenna. Within the main beam, a common engineering approximation is L ≈ 12 · (θe / θ3dB)² dB, with the half-power beamwidth θ3dB ≈ 70 · λ / D degrees for a parabolic dish. An azimuth error contributes θaz · cos(elevation) to the pointing error.

Dish (Ku band, 12 GHz)Beamwidth θ3dB0.1° azimuth error0.25°0.5°1°
0.6 m2.9°0.01 dB0.07 dB0.26 dB1.1 dB
1.2 m1.5°0.04 dB0.26 dB1.1 dB4.2 dB
2.4 m0.73°0.17 dB1.1 dB4.2 dBbeam missed
012345600.250.50.751−3 dB0.6 m dish1.2 m dish2.4 m dishazimuth error (°)pointing loss (dB)
Ku band 12 GHz, 30° elevation; L ≈ 12 (θe/θ3dB)², valid in the main beam (own calculation)

Values at 30° elevation, own calculation with the approximation above. The larger and higher-frequency the antenna, the smaller the azimuth error it tolerates: a 2.4 m dish loses half its power with half a degree.

03

Can an IMU find north by itself?

A gyro can find north by measuring the Earth rotation (gyrocompassing). The horizontal component of Earth rate is 15.04 °/h · cos(latitude), about 9.9 °/h in southern Germany. Any gyro bias b is indistinguishable from that signal, so the azimuth error is roughly Δψ ≈ b / (15.04 °/h · cos φ) in radians.

Gyro biasAzimuth error at 48.7° NTypical class
3 °/h≈ 17°industrial MEMS
0.5–1 °/h≈ 3–6°high-end MEMS (all makes)
0.09 °/h0.5°tactical fibre-optic gyro
0.02 °/h0.1°navigation-grade FOG / ring laser
0.010.11100.010.1110100< 0.5°high-end MEMSfibre-optic / ring laser48.7° N (southern Germany)gyro bias (°/h, log)gyrocompass azimuth error (°, log)
Δψ ≈ b / (15.04 °/h · cos φ), own calculation; sensor classes typical

This is physics, not a weakness of one product: no MEMS gyro on the market reaches the 0.01–0.1 °/h needed for sub-degree gyrocompassing. Those gyros cost many times more and are larger. Systems with MEMS IMUs therefore take north from GNSS and use the IMU where it is strong.

04

How does tilt turn into azimuth error?

If the azimuth axis of a mount is tilted by an angle i across the pointing direction, the azimuth error grows with elevation h as Δaz ≈ i · tan(h). Surveyors know the same effect as trunnion axis error of a theodolite.

ElevationAzimuth error from 0.1° tilt
10°0.02°
30°0.06°
45°0.10°
60°0.17°

This is where an IMU or a precise accelerometer earns its place: tilt is measured against gravity, which is always available. An accelerometer bias of 24 µG, the bias instability of the M-G370PDG accelerometers, corresponds to only 0.0014° of tilt. In practice mounting misalignment dominates, and it can be calibrated once.

05

Which architecture works in practice?

  1. 01Dual-antenna GNSS for north: the heading error scales with the inverse of the baseline. With an assumed 5 mm relative accuracy, a 1 m baseline gives about 0.29°, a 2 m baseline about 0.14°.
  2. 02IMU for tilt and stabilization: roll and pitch from the accelerometers, fast motion from the gyros, for example on vehicles, ships or masts that sway.
  3. 03IMU for hold-over: when GNSS heading drops out under bridges, in harbours or under jamming, the gyros carry the heading. Error grows with bias × time.
  4. 04Magnetometer only as a fallback: steel structures, motors and cables distort the field, often by several degrees.

Simulate the hold-over case for your own outage time with our IMU drift simulator.

06

Which Epson IMU fits antenna pointing?

ModelGyro bias instabilityARWHeading drift in 10 min*
M-G366PDG1.2 °/h0.08 °/√h≈ 0.20°
M-G370PDG0.8 °/h0.06 °/√h≈ 0.13°
M-G570PR0.5 °/h0.04 °/√h≈ 0.08°

* Bias instability × time, after the bias was estimated while GNSS heading was available; ARW adds only 0.02–0.03° in 10 minutes. Values: AMCORIS product data, own calculation. For tilt-only tasks such as alignment of fixed antennas, a low-noise accelerometer like the M-A352AD10 is often enough.

FAQ

Frequently asked questions

Your question is not listed?

Ask our engineers →

QWhat is a typical azimuth accuracy target for antenna pointing?

It depends on the beamwidth. As a rule, keep the pointing error below about a quarter of the half-power beamwidth: then the loss stays below roughly 0.75 dB.

QWhy does my azimuth error get worse at high elevation?

Because tilt of the azimuth axis couples into azimuth with tan(elevation). At 60° elevation a 0.1° tilt already causes 0.17° azimuth error.

QCan I use a magnetometer instead of GNSS for north?

Only in a clean magnetic environment. Steel masts, vehicles and motors distort the field. A magnetometer is useful as a fallback, not as the main reference for narrow beams.

QDo I need a navigation-grade gyro?

Only if the system must find north alone without GNSS. If dual-antenna GNSS or a surveyed reference provides north, a MEMS IMU covers tilt, stabilization and short outages.

EVALUATION KIT

Test it on your antenna

We lend evaluation kits for Epson IMUs and accelerometers so you can measure tilt and heading hold-over on your own mount.