Polar Data Flow.

During normal operations the processor powered by side 1 of the system assumes the role of the active processor and the processor powered by side 2 is the secondary processor. Either processor, however, is capable of assuming the role of the active processor.

The active processor controls the instrument and processes the analog Magnetometer data to provide four digital data streams for the spacecraft as follows:

1. High rate (108.7 vectors/sec) onboard data
2. High rate (100 vectors/sec) internal snapshot data
3. Low rate normal data (8 or 4 vectors/sec)
4. Low rate sun crossing data ( 1 vector/major frame)
5. Instrument housekeeping data

The transfer function of the magnetometer is determined by the frequency response of 4 parts of the circuits: the basic magnetometer; an analogue anti-aliasing filter; a recursive digital filter immediately following the A/D conversion that limits the bandwidth of the 100 Hz data; and a digital filter that limits the bandwidth of the low rate data (8 or 4 vector/sec).

General discussion of timing

The MFE instrument has its own internal clock which is not synchronized with the spacecraft clock. Basic MFE system timing is derived from the MFE crystal oscillator that allows MFE to control the position of the ADC samples to minimize noise pickup. This high frequency clock is used to develop the 1 millisecond interrupt clock that starts the collection of data from the analog system and distributes this data to the spacecraft. By performing a vector sample on every other interrupt, we were able to keep the processing time required to collect, process and distribute the high rate data to between 500 and 750 microseconds. The remaining processing time is used to store the snapshot data, generate the housekeeping parameters and to filter, decimate and format the low rate data.

Data Flow and On-board distribution of MFE data

Analog Signals

The POLAR MFE basic magnetometer has a second order low pass response with a corner frequency of about 300 Hz. It introduces a phase shift that is smaller than the following anti-aliasing filter, second order analog Nyquist filter with a corner frequency of 239 Hz. The analog filters and basic magnetometer response are expected to have minimal impact on the spin tones.

The second order analog Nyquist filter with a corner frequency of 239 Hz introduces a phase shift of 0.0648 degrees and amplitude of .999999 at the spin period 1/6 Hz.

The 500 Hz Data Stream

The internal sampling rate of the MFE instrument is 500 vector samples per second. There is 150 microsecond differences between the x, y, z samples. The analog magnetic wave forms of the three sensors sampled at 500 vectors per second are then filtered with a recursive digital filter of the form:

On-Board Distribution of 108.7 Hz Data to CAP

The reordered and inverted 500 vector per second data stream is decimated to high rate data at 108.7 vector per second. Magnetic field values (Bx, By and Bz) are sampled within 150 microsecond. The sample time has a 2 millisecond jitter and is timed by the delivery of the first of the 4 words transmitted per vector. The Command and Attitude Processor (CAP) ingests and distributes the 108.7 Hz high rate MFE data. The CAP uses this high rate data to compute the Magnetic Field AZ (azimuthal angle) and EL (elevation angle). The CAP distributes these data to a number of other POLAR instruments (TIDE, TIMAS, CAMMICE, CEPPAD, HYDRA) for their use. We have noticed that there is a 4 deg jitter ( we would expect the jitter to be less than 1 deg) and a constant (5-10 deg) offset in the present AZ and EL values. This problem is currently being investigated.

Normal Data (8 or 4 Hz) and Burst Mode data (100 Hz)

The reordered and inverted 500 vector per second data stream is also, in parallel, decimated to 100 vector per second which can be buffered and stored for later transmission as "burst mode data". These data have the same attenuation and phase lag as the 108.7 Hz data used by CAP.

The 100 Hz data are then filtered with a recursive low pass filter prior to transmision to Earth using a filter of form

This filtered 100 Hz data are then decimated to a rate of 8.3 vector/sec (4 vector/sec if the burst mode is activated) and stored in the data buffer to be read out in the next major frame 9.2 s later. Thus, the data provided in a major frame were collected in the time period corresponding to the previous major frame. This 9.2 sec time delay must be taken into account when determining the time of acquisition of each sample.

Total digital transfer effects for both digital filters result in an attenuation of 0.99682 and a phase lag of 5.0308 deg at the nominal spin rate of 6 seconds in the normal data.

In the 8.3 Hz (120 msec) data stream that provides the standard magnetic field measurements to ground, the total tranfer function combined the analog filters and digital filters produces an attenuation of 0.99682 and a phase lag of 5.0956 deg at 1/6 Hz spin tone. This phase lag is equivalent to a time shift of approximately 85 msec.

The secondary processor is not normally used but could be used to sample the magnetometer not being sampled by the first processor. It could also completely replace the functions of first processor.

Low Rate Sun Crossing Data

The onboard sun crossing vector (1 vector/major frame) is a sample of the MFE data taken just after the suncrossing pulse distributed onboard the spacecraft.

Before 17:32 UT, October 2, 1996 the "sun crossing" sample of the magnetic field was obtained from 8 Hz (120ms) data stream by taking the first sample after the time of the sun crossing pulse. As a result of the lack of synchronization of the MFE and spin clocks the time of the sample can be as much as 120 msec after the time of the sun pulse. This introduces approximately 7.2 deg of jitter to the data. This jitter can be corrected by noting the actual time of the sample, in stead of the time of the sun crossing, and then rotating the direction of the field to account for the spin of the spacecraft between the suncrossing time and the sample time.

At 17:32 UT October 2, 1996 the flight software was changed to select the sun crossing data from 100 Hz (10 msec) data stream rather than the 8 Hz (120 msec) data stream. This will reduce the variability in the delay of the sun crossing data after the sun pulse occurs to 0.6 deg of s/c rotation.

Operations

The instrument is usually operated with each triplet of samples (X, Y, Z) being the field along the three spacecraft directions. However, there are times when the commands to do this were not properly executed. These times are noted under major events. This problem also occurs for 18.4 seconds (2 major frames) after each switch between the inboard and outboard sensor, and for 9.2 seconds (1 major frame) for a range change.

The onboard offset adjustment was activated beginning 22:02 UT, August 28, 1996. From this time, the MFE goal is to keep the zero levels being applied to this broadcast data to within 1 nT of the proper values by uploading them once per week.

A history of the updates to on-board offsets etc is available on MFE web page under major events.

There are two separate magnetometers mounted on the boom, that are designed to measure different range of the magnetic field vector:

Outboard Sensors: +/- 700 nT in low range and +/- 5700 nT in high range
Inboard Sensors: +/- 5700 nT in low range and +/- 47000 nT in high range

In order to properly accommodate the changing magnetic field and control when the instrumental change in operations occurs, there is a ground commanded range change four times per orbit.

The range change requires 1-2 major frames (each major frame is 9.2 sec) to set the zeros, coupling coefficients, and reinstall the sensor inversion patch. An additional major frame is required for power switching if the range change involves changing sensors. During this frame no magnetic field measurements are available.

The gain state is determined by looking at the mag scale factor which is provided in the instrument housekeeping.

The use of the scale word should read as follows:

    Inst mode  Engr Word    Actual     Full Scale Field
    Sens-Range 2*Cnts/nT    Cnts/nT       (32768 cnts)
               (MNTSCALE)
		    
       IN-HI     2          0.7018      46689.9nT
       IN-LO    11          5.59         5861.72
      OUT-HI    11          5.94         5516.33
      OUT-LO    88         47.44          690.70

In actuality, there are different scale factors for the different sensors but they all match to within 0.3% to the above values.