Overview of HRC Secondary Science Corruption
Secondary Science Corruption (SSC) refers to an occasional byte-shift anomaly in HRC data which causes a portion of the housekeeping information to be corrupted. Symptom are dropouts in the dead-time-factor (DTF) which can be seen in a trend of the DTF file’s DTF values vs time, where the DTF is significantly ( > 10%) below the median value (∼ 1.0), very high total evt count values in the DTF file (> 2000 cps), and a large number of GTI intervals over the times of the anomaly.
When SSC is noticed during Verification and Validation (V&V) of pipeline products, a comment is entered into the V&V report to the effect that all events are good.
Relevant Documentation
Some background and low-level information on SSC can be found in the following links:
- “Proposers Observatory Guide”, §7.12, “Instrument Anomalies”
- HRC Secondary Science Anomaly on Day 2005:349
- Secondary Science Byte Shift
- HRC Secondary Science and Housekeeping, Invalid data
SSC Mitigation
There is no a priori determination that SSC has occurred at an HRC-frame level. Instead we rely on a statistical treatment based on inspection of about 20 observations, with and without SSC. The primary discriminant seems to be the DTF file’s total_evt_count, a quantity integrated over 2 sec bins (by de- fault, in standard product DTF files). Bright, extended sources, such as Cas A and N 132D, can have high total evt count (∼ 1000–2000) and low DTF (∼ 0.1) and not have SSC; they are simply telemetry saturated observations. When SSC occurs, the total_evt_count can be very large, in excess of 10,000 cts. Empirically, values larger than 4000 cts seems to be a reliable indication that SSC has occurred.
Examples
Example of SSC in ObsID 19016, taken with HRC-S/LETG. The top panel shows the total_evt_count against time. Note the cluster of high values at about y = 18000. These are symptomatic of the occurrence of SSC. The middle panel shows the DTF, with original points in black, showing drops to 0.9 or 0.0 during SSC, and corrected values in red (all at 1.0). The bottom panel shows the old GTI intervals (upper line, with colored “+” marks), and below it, the new GTI in black. The y-scale is arbtrary. The “Old” GTI successive intervals have alternating offsets and colors, just to show the multitude of intervals.
Example of saturation, but not SSC, in ObsID 3698, taken with HRC-I/NONE. (The very small differences in exposure and counts are due to an 85s difference between the Level 1 GTI on the events, and in the flt1 file.)