The Vector Magnetometer

There are a number of factors which affect the magnetometer accuracy. A residual constant offset, usually of the order of a few nT, is introduced by amplifier offsets. Corrections for this effect are determined by preflight calibrations. The instrument gain, that is, the ratio between the applied field and the output voltage resulting from the combined effects of amplified gain and A/D conversion, is also subject to small constant deviations from design specification and is different for the three sensors. Gain corrections are also obtained as part of preflight calibrations.

There is also a small non-orthogonality (< 0.1 degrees) between the sensor orientations leading to a mixing of field components. Some "cross talk" between sensors at the 0.1% level may also occur in the electronics. The alignment and electronic "cross talk" effects are also calibrated before flight.

Preflight calibration was performed at the Goddard Space Flight Center Magnetics Test Facility. Offsets were determined to within 1 nT, and gain corrections and cross talk effects were characterized to 0.1% accuracy. These calibrations are applied to the data in the form of a calibration matrix Gsubij such that

Both local fields from the spacecraft and sensor orientation uncertainties (principally due to uncertainties in the final boom orientation) adversely affect the VMAG data and must be corrected for by analysis after launch and deployment.

The VMAG sensor is mounted on a boom extending 5.77 m from the spacecraf t center line, vertically upward in flight, at the aft end of the spacecraft. The boom is an Astromast type which rotates around several times during deployment and is rated to repeat the final orientation to within 2 degrees. This uncertainty means that the sensor will be misaligned relative to spacecraft coordinates. To obtain high precision measurements, this alignment effect must removed by analysis of data obtained on orbit.

There are two primary sources of local magnetic fields on UARS: (1) permanent magnets in ion detector instruments near the VMAG sensor, and (2) electromagnetic torquer bars in the spacecraft Attitude Determination and Control Subsystem (AD&CS).

The HEPS units mounted at the end of the Astromast boom contain permanent magnets to prevent electrons from entering the ion detectors. The magnetometer sensor is separated from the these detectors by a 26 cm stub boom to reduce interference, but the field strength from the permanent magnets is still some 400 nT at the sensor. Since this field is constant in the spacecraft frame, it is readily removed to at least +-5 nT accuracy by preflight calibration of the particle detector package as well as analysis of in-flight data (see section below).

The second source of spacecraft fields, the AD&CS torquer bars, produces a field which is variable in time and therefore more difficult to remove. To minimize the torquer bar fields at the sensor, the magnetometer sensor and torquer bars are located at opposite ends of the spacecraft. The center of the closest torquer bar is 7.8 m away from the sensor . Even so, the field from the torquer bars at the sensor ranges up to 200 nT. To correct for this contamination field, a field model of the torquer bars was constructed using measurements made at the GSFC Magnetic Test Facility using an engineering replica of the torquer bars. By using the torquer bar currents monitored on the spacecraft, the model provides absolute prediction of the torquer fields to better than 2% accuracy.

To ensure maximum accuracy and precision of the magnetic field data it is necessary to verify instrument performance in flight. An internal instrument calibration is provided which consists of switching the amplifier inputs from the null current signal to a reference signal. Hence, only the conversion from null current to output voltage and subsequent digitization is monitored by on-orbit calibration. However, analysis of the data returned by the instrument allows end-to-end validation and monitoring of its performance.

Throughout most of the UARS orbit, the torquer bar currents are changed relatively slowly, with a period of about one orbital period. But rapid, (five-minute) cycling of torquer currents also occurs during normal operations and provides an opportunity to correlate stepwise changes in the measured field with the predicted torquer fields to verify the model on orbit. Analysis of such intervals has provided confirmation that the local field model is correct to within +-10 nT, and hence there is no evidence that the local field model derived from preflight calibrations is in error.

The UARS spacecraft is three-axis stabilized to 60 arc second pointing accuracy and 2.5 km position knowledge. The high precision of the spacecraft platform position and orientation allows one to use a geomagnetic field model to verify the preflight calibrations and to determine the actual sensor orientation relative to the spacecraft reference frame. For purposes of determining offsets, sensor orientation, cross talk, and relative sensor gain, the accuracy of the model field is not as important as is the fact that the accurate spacecraft attitude and position ensures that the model field presents a fixed reference to which the measured data may be compared.

Intercomparison of the data with model fields proceeds as follows. The model field is first transformed into spacecraft coordinates. All of the validation analysis is preformed in spacecraft coordinates. The residual field B = BUARS - Bmodel is constructed for a suitable period of time, typically 24 hours providing some 15 complete orbits of data. The extent to which B correlates with the model field reflects systematic errors in the data due to orientation or gain errors. The residual fields are least square fitted to the form B = M(Bmodel) + C . The constant offsets are given directly by C, and the matrix M provides the orientation, cross talk, and gain correction factors.

The gain corrections obtained in this way only assure that the gains are consistent among the three sensors, but do not provide a check of the preflight absolute calibration. Hence, it is only the precision rather than the absolute accuracy of the data which is validated and adjusted by comparison with the model field. With this technique it has proven possible to determine the sensor-to-spacecraft coordinate transformation to within about 0.1 degrees and to verify the constant offsets to within +-5 nT. The resulting data give measurements of the geomagnetic field to an absolute accuracy of +-1% and a precision of at least 0.1%.

Return to Instrument Overview