Copyright 1995 Kluwer Academic Publishers.
Reprinted by
permission of Kluwer Academic Publishers.
An extensive series of amplitude calibrations, frequency responses, and instrument performance checks were carried out for the PWI before integration on the spacecraft. In addition, phase calibrations were performed on the HFWR, LFWR, and SFR receivers.
Amplitude calibrations for each of the receivers were accomplished by providing an input signal of fixed frequency at the center frequency of each filter. The amplitude of the stimulus was decreased in two dB increments to cover the full amplitude range of the receiver. The amplitude response curves are used to construct look-up tables that convert the telemetry data value to the true input signal strength. For the WBR and HFWR, the amplitude response tests were repeated for every possible gain setting and for every filter and conversion mode. Using a gain setting of 0 as the baseline, the gain of the input signals for the remaining gain settings was correspondingly decreased to maintain a constant signal strength into the instrument.
The MCA and SFR channels use a compressor with a piecewise-linear approximation to a logarithmic amplitude response. Over the full range of amplitudes, the response of each compressor consists of a series of five distinct linear segments that deviate somewhat from a true logarithmic response. Because the compressors have different amplitude sensitivity characteristics, the amplitude response of each compressor must be individually measured. The amplitude response of a typical logarithmic compressor is shown in Figure 8 for the middle step of the first SFR band at 69 Hz.
Figure 8: The amplitude response calibration curve for the middle step of the first SFR band at 69 Hz. The curve illustrates the five-stage, piecewise-linear response of the SFR logarithmic compressor. Amplitude calibration curves for the PWI filters are used to construct look-up tables which convert the telemetry data values to the true input signal strength.
The slight deviation from a true logarithmic response is caused by the piecewise-linear response. Similar amplitude responses are measured for the middle frequency step of each SFR band and for the 20 MCA channels.
Since all of the frequency steps of a given SFR band utilize the same logarithmic compressor, it is only necessary to measure the amplitude response at one frequency step in each SFR band. Amplitude responses for the remaining frequency channels are derived from this amplitude response by using the results from the channel-to-channel gain test (see the next section). This combination of calibrations is sufficient to determine the input signal strength for all frequency steps in each SFR band.
Frequency response and phase calibration tests were performed simultaneously by applying an input signal of fixed amplitude and sweeping from the lowest frequency to the highest frequency. For parallel receivers that simultaneously receive signals from two antennas, the phase difference in degrees was measured using the same input signal to both receiver inputs. The measured phase values were assembled in a phase calibration table for each receiver.
Figure 9 shows the frequency response of the five SFR channels.
Figure 9: The frequency response plot for the five SFR bands. The frequency response test gives the channel-channel-gain adjustment, which is used to calibrate a fixed-amplitude signal for the SFR frequency steps within a given frequency band and to calibrate a fixed-amplitude signal across channel boundaries.
This plot provides the channel-to-channel gain relationship that is used to calibrate the amplitude response across channel boundaries. Similar channel-to-channel gain plots were created for the MCA channels. A combination of amplitude response and frequency response measurements are used to create a complete set of calibration look-up tables for each receiver. In the case of the SFR receivers, the channel-to-channel gain measurements provide the necessary adjustments to create a complete set of calibration tables for every SFR frequency step.
To verify that the amplitude and frequency response of each filter are independent of temperature, all of the above tests were conducted at room temperature (25°C), at 0°C, and at 40°C.
Because the electric antennas were constructed and calibrated by the EFI team, all tests prior to spacecraft integration were performed by applying the input signals directly to the main instrument package. A similar procedure was used for the magnetic antennas. To obtain frequency response and phase calibrations for the sensor-receiver system, tests were independently performed on the sensors and combined with the calibrations for the main electronic package. In addition, prelaunch end-to-end calibrations, in which the input signals were directly applied to the electric and magnetic antennas, were selectively performed to verify the in-house calibrations.
To compute the effective noise bandwidth for the various filters used in the instruments, a white noise input signal of known spectral density (in V²/Hz) was used. The spectral density of the noise generator was adjusted to remain flat across the detection bandwidth of each filter. The DC output of the log compressor was then compared to the corresponding output for a sine wave stimulus (from the amplitude response calibrations). The ratio of the sine wave amplitude squared to the voltage spectral density gives the effective noise bandwidth of the filter.
For the search coil and loop antennas, the magnetic field sensitivity, the frequency response, and the phase response were calibrated at the Goddard Space Flight Center Magnetic Test Facility. The transfer function measurements were performed using a large solenoid, a Helmholtz coil, and two transmitting loops, all of which were driven in various configurations by known AC current sources. The results of these four methods were compared and found to be in good agreement. The phase response was determined using a transmitting loop driven by an AC source of known phase. The magnetic noise levels were measured by placing the search coil in a µ-metal container to reduce the effects of external noise sources. Figure 7 shows the estimated noise level curves for the loop and search coil antennas based on the initial examination of the calibration data.
