PDS_VERSION_ID = PDS3 RECORD_TYPE = STREAM OBJECT = TEXT INTERCHANGE_FORMAT = ASCII PUBLICATION_DATE = 1997-08-07 NOTE = "Internal memo describing gain changes seen in Clementine bistatic radar data acquired at DSS 43. Since DSS 43 data are not included in this archive, the results are not directly applicable. Table 1 and figures were not available when the archive was constructed." END_OBJECT = TEXT END DSS 43 Amplitude Calibration Report (94/099-100) Characterizing Several Different Causes for Signal Level Changes Objectives ---------- The goal of this task was to characterize several types of signal changes observed during Clementine bistatic radar data acquisition. Overview -------- Several types of signal changes can be identified, in approximate order of increasing time duration: (1) telemetry mode changes (2) swaps between two-way and one-way (AUX OSC) radio tracking on the spacecraft (3) changes in attenuator settings at the ground station (4) changes in range modulation -- on the ground and on the spacecraft (5) discontinuities in ground receiver tuning -- introduction of frequency offsets (FROs) in the local oscillator (6) switching between high-gain (HGA) and omnidirectional spacecraft antennas (7) ambient load calibrations at the ground station (8) changes in pointing of the ground antenna (9) changes in pointing of the spacecraft HGA Each of the events above requires a certain amount of time and affects the amplitude and/or frequency of the signal in different ways. Telemetry mode changes are essentially instantaneous and usually amount to a few dB; a change in spacecraft HGA pointing can reduce the signal by more than 40 dB and be spread over several minutes. We are especially interested in signal level changes that can be attributed to changes in attenuator settings; we can compensate for these using the GAIN program. We are also interested in changes during ambient load calibrations; these allow us to calculate system temperatures for each channel. In addition to the data themselves, we have notes taken during most experiments by Dick Simpson; because Simpson was at JPL and the data collection was being carried out at DSN stations, the notes are an incomplete record of actual events. A few notes were provided by Sal Abbate, on site at DSS 14 for the first observations. Automatic recording of attenuator settings in DSP record headers is also useful documentation, but header information was infrequently updated and sometimes lagged real time by more than 30 seconds. Method ------ We looked closely at the power vs time plots for each file as generated by the PREPPOWER program. We looked at high time resolution around the points that seem to have the most abrupt changes. In order to accomplish compensations we have to know the beginning and end times of the transition regions for each attenuator change. We then use these times in the input file to the gain adjustment program. During some time intervals the signal either saturates the receiving system (or the analog-to-digital converters) or it is so weak that reliable amplitude estimates cannot be obtained because the analog signals are well below the digital quantization threshold. Identifying these as quantitatively invalid data means we can skip over them in subsequent processing. Knowing the beginning and ending times of other signal changes also allows us to flag them as questionable. Therefore the plots we obtained at several different resolutions serve a double purpose. One is to identify the cause of the signal level change, and the other is to determine the exact transition region for these changes. Results ------- This report is an analysis of the typical characteristics of these events; it is illustrated with example plots. Table 1 summarizes key characteristics of each event type. 1. Telemetry Mode Changes: When the telemetry modulation index changes, the amount of power in the carrier also changes. In general, a reduction in the telemetry modulation index increases the carrier power. When the modulation index is reduced to zero (as it was for bistatic radar observations), all transmitted power is in the carrier. Figure 1 is an example of the transition when modulation index is reduced to zero prior to a bistatic radar experiment. Time required for the transition cannot be resolved; the increase in carrier power is about 5 dB. There may be several telemetry mode changes during an orbit; but there should be only one per orbit in which the modulation index is reduced to zero. 2. Frequency Reference Swap (two-way to AUX OSC): Because the spacecraft cannot lock reliably to an uplink signal during bistatic operations, it is necessary to switch from two-way mode to using the onboard AUX OSC as the transmitter frequency reference. Although this should not, in principle, affect the carrier output power, there is an observed change in the data. This may be partly from a shift in the downlink signal frequency -- and appearance of the carrier at a different region of the baseband frequency window. The change shown in Fig. 2 is less than 0.2 dB in amplitude. The signal seems to disappear for about 5 msec during the transition. Quasi-periodic amplitude "spikes" at about 0.5 sec intervals may be from non-linear behavior of the maser while the spacecraft HGA was pointed directly at the DSN station. There should be (at most) one swap to the AUX OSC during each Clementine orbit. 3. Changes in Attenuator Settings at the Ground Station At low resolution attenuator setting changes are characterized by a power level jump of (almost) an integer multiple of 1 dB. These changes are recorded in the headers of DSP records but with a delay of up to 30 sec. Therefore we can verify them by reading the headers, but only if sufficient time elapses between attenuator changes to permit propagation of the new values. We have observed that the header values match the actual jumps closely except for the cases when the signal becomes too high or too low. In both cases the observed jumps are less than the jumps that would be expected by the attenuator change. In the first case this is due to saturation and in the second it is due to quantization. At high time resolution, attenuator setting changes appear to last 4-10 msec. In the transition region the signal drops suddenly and completely disappears as seen in Figure 3. The attenuator switch apparently breaks its old contact before making the new. Attenuator setting changes can occur frequently, though usually spaced by a few minutes so that engineers and others could observe results before requesting new values. Each attenuator change should be recorded (albeit with some delay) in the DSP record headers. Except when the signal reached saturation or fell below the sampling quantization threshold, each change in attenuator setting should appear as a change in signal level by an integral number of dB. 4. Change in Range Modulation Changes in range modulation state should affect the carrier level in ways that are similar to changes in telemetry modulation index. The range modulation accounts for only a fraction of 1 dB in total signal level, however. In DSS 43 data we have observed two events which MAY be associated with changes in ranging modulation. In Fig 4a, the ground ranging modulation was apparently turned off and there was an expected increase of about 0.5 dB in carrier level. A short time later, the spacecraft ranging channel was apparently turned off; instead of a further increase in carrier level, the signal actually dropped about 0.2 dB. This needs further investigation, but no other explanation has been forthcoming. 5. Introduction of Frequency Offsets (FROs) in Ground Receiver Tuning When the signal drifted toward either edge of the baseband filter, operators could introduce frequency offsets in the local oscillator tuning to bring the signal back toward the center of the passband. These discontinuities often resulted in a sudden change in apparent signal level because the carrier was moved from one part of the baseband filter to another. 6. Switch Between HGA and OMNI on the Spacecraft Switching between the high-gain antenna and omnidirectional antenna could produce a signal level change on the order of 40 dB if the HGA were pointed toward Earth. This event was usually schedule toward the end of a bistatic radar observation, however, when the HGA was pointed toward one of the polar regions on the Moon. In those cases, the change in signal level was much less. And there are circumstances when switching to the OMNI might actually increase the signal level observed on Earth. Figure 6 shows one such switch -- about 1 dB in amplitude and spread over a few hundredths of a second. The time dispersion may result partly from the fact that echo signal scattered from the lunar surface is extinguished a fraction of a second later than the direct signal -- because of the added propagation time. 7. Ambient Load Calibrations Conducting an ambient load calibration requires that a resistive load be attached to the receiver input on each channel. The resistor is at the ambient environmental temperature (nominally 290K) and produces broadband radiothermal noise. This is then compared to the noise level observed when the ground antenna is pointed toward a "dark" part of the microwave sky. A comparison of the two measurements allows estimation of the "noise figure" or "system temperature" of the receiver. Connecting the ambient load requires that a physical change be made in the receiver-antenna circuit. Figure 7 shows that connecting and disconnecting an ambient load takes a few seconds and that three distinct phases can be identified. First is the receiver output with antenna connected. Second is a transition lasting perhaps 2 sec when the receiver output drops; this is presumably a time when the receiver is not connected to anything. Third is the receiver output with the ambient load attached, with a power output about 12 dB higher than at the beginning. Higher resolution examination of the ambient load connection shows a short amplitude spike as the switch moves from the antenna to the "no-connection" condition. The rise as the ambient load is connected is seen, in high resolution, to take about 0.5 sec. When the ambient load is disconnected, the process is reversed. During ambient load calibrations on Channel 1 (RCP) we have seen changes in signal level of about 1 dB on Channel 3. These occur simultaneously with connection of the ambient load to the Channel 1 receiver. 8. Changes in Pointing of the Ground Antenna During some observations, the ground antenna was commanded to point toward the center of the Moon. This was intended to provide both receiver channels with an identical radiothermal noise source in the sky. The beam of a 70-m ground antenna subtends only a fraction of the Moon's diameter. Such tests typically resulted in a change in signal level on the order of 3 dB and spread over a few seconds (Figure 8). Return to pointing at the South Pole was a reverse of the process except that the antenna generally slipped off the Moon for a few seconds and signal levels took a significant dip. At the end of bistatic radar observations for one orbit, the antenna was usually returned to the stow (vertical) position for a round of ambient load calibrations. The receiver output typically dropped 10 dB in going from the South Pole to stow. Most of the change occurred over a couple of seconds, but an extended taper probably resulted from reduction in atmospheric radiation captured as the antenna went from its tracking elevation to zenith. During some portions of each observation the ground antenna moved between pointing at the spacecraft and pointing at various parts of the Moon. Signal level changes are obvious, but quantitative analysis may be impossible because the strong directly propagating signals from Clementine overloaded the maser front-end of the receiving system. 9. Changes in Spacecraft High-Gain Antenna Pointing Prior to each bistatic radar observation, the spacecraft HGA was pinted at Earth. In order to get scattering from the South Pole, the spacecraft was turned. As a result, the signal level on the ground dropped by as much as 40 dB. In most cases, the sidelobes of the HGA are apparent in the power vs. time plots of ground signal level. These events took tens of seconds.