Citation
Abstract
Given the Near-Earth Object Camera (NEOCam) Project’s stringent data return requirement of having virtually no frame errors during a 100-minute tracking pass utilizing a K-band (26 GHz) link, an analysis was conducted to ensure that the link assumptions were sound in realizing this requirement. The link was designed with sufficient margin to ensure near 100 percent data return in the case of quiescent weather conditions. In this case, any lost data is expected to occur in isolated rare instances, where it can be easily retransmitted using a protocol such as Automatic Repeat Request (ARQ). For the case when significant fading occurs beyond 95 percent weather, widely scattered frame errors would necessitate all or almost all of the entire pass to be retransmitted, thus a link margin above 8 dB is used to alleviate this. A theoretical frame error rate (FER) of 10−8 will result in virtually no frames lost during a typical 100-minute pass. The change in threshold E b /N 0 at FER = 10−8 relative to the Project’s adopted threshold of E b /N 0 at FER = 1.4 10−6 is only ~0.2 dB which lies well within the Project’s adopted 8 dB margin, providing confidence in the link design. Actual frame errors realized during operational Ka-band tracking passes were also examined in an attempt to provide observational evidence of the theoretical expectations. We used Kepler Ka-band (32 GHz) telemetry data from 66 tracking passes where specific criteria were satisfied. These criteria involved considering data where the minimum symbol signal-to-noise ratio (SSNR) exceeded 0 dB, which lies just above threshold for the coding/modulation configuration used. Out of the 66 passes examined, 60 had zero frame errors after removal of nonzero frame data with justified reasons for deletion. Out of ~388 million frames in the 66 passes examined, there were 6 tracking passes containing 14 time instances with 738 frame errors where no justification for deletion was apparent. Thus, after removing data points during periods of nonzero FERs with justified explanations, we have effectively found only 738 frame errors occurring in 14 discrete time instances that have yet to be fully explained.
Details
- Volume
- 42-216
- Published
- February 15, 2019
- Pages
- 1–16
- File Size
- 700 KB