050901 R Vanderspek HeteLessons

HETE Operations&Lessons Learned: 

HETE Operations & Lessons Learned Roland Vanderspek, MKI September 1, 2005

Overview: 

Overview Overview of HETE Operations Application to Eclairs

HETE Operations: 

HETE Operations Spacecraft in 600 km equatorial orbit Commanding done from any of three Primary Ground Stations (PGS): Cayenne, Kwajalein, Singapore Uplink and downlink are completely automated “Duty Scientist” (DS) is responsible for responding to spacecraft anomalies and burst events (one week at a time).

HETE Instrumentation: 

HETE Instrumentation Spacecraft bus includes basics (power, RF) and attitude control (momentum wheel, three orthogonal torque coils). Science instruments are Fregate (wide-field NaI scintillators, 6-400 keV), WXM (wide-field X-ray imager (2-25 keV), SXC (wide-field CCD imager (2-10 keV). Spacecraft aspect is calculated using two 5°x5° optical imagers coaligned with the instruments. Aspect is calculated every 1.0 seconds. Precision of aspect is ~5 arcseconds.

Automated Operations: 

Automated Operations Spacecraft commands are generated automatically: simple shell scripts define the parameters of the observations (start and stop times, instrument configurations, etc.) Basic satellite and instrument housekeeping data are reduced by automated scripts. Abnormal situations (e.g. low voltages, anomalous attitude) are flagged and the DS is notified (by email or beeper, depending on severity).

Burst Operations (I): 

Burst Operations (I) Fregate (or WXM) burst trigger leads to full on-board localization analysis and downlink of full set of high-resolution burst data from all instruments (time-tagged photons, detailed optical aspect). On-board burst localization analyses typically complete within 30 seconds Burst localization calculated on board is broadcast over the VHF antenna to the Burst Alert Network: the coordinates are then relayed to the GRB Coordinates Network (GCN) at GSFC via MIT.

Burst Operations (II): 

Burst Operations (II) High-resolution burst data are reduced by automated scripts shortly after downlink. HETE scientists, alerted by beeper, perform manual analyses of the downlinked data: improved localizations are typically available within 60 minutes. Discussion of manual analyses occurs via the HETE chat line (hchat).

Flight Operations: 

Flight Operations Spacecraft pointing is nominally anti-solar. Exceptions are handled by “nodding” the spacecraft away from antisolar by up to 40° Exceptions include Full moon avoidance (monthly) Sco X-1 avoidance (April-July) Galactic bulge (May-August)

Lessons Learned from HETE (I): 

Lessons Learned from HETE (I) Flight software development should be done under conditions as close to flight as possible (software uplink, data downlink, inter-processor communications). Each HETE instrument team was given a copy of the flight processor hardware (“pizza box” = spacecraft processor + instrument processor) for software development. A hardware copy or high-fidelity software simulator of the UTS (and EGCU?) is needed.

Lessons Learned from HETE (II): 

Lessons Learned from HETE (II) Each instrument must be end-to-end and GSE testable, both in the lab and on the spacecraft. This means: Laboratory tests need high-fidelity version of the flight computer (“pizza box”) Instruments should be equipped with “eavesdropping” connections (for GSE) If possible, instruments should be operable at room temperature (for on-the-rocket tests).

Lessons Learned from HETE (III): 

Lessons Learned from HETE (III) HETE flight software was developed in US (MIT + Los Alamos), Japan, and France. The Fregate + WXM transputer software was developed at Los Alamos, but required input from Japan and France: inadequate communications between teams caused problems and introduced delays. Flight software modules should be as distinct as possible. Software interfaces should be defined clearly and strictly adhered to.

HETE Lessons Learned (IV): 

HETE Lessons Learned (IV) HETE SXC ground analysis software relies on full information about spacecraft pointing during the burst, i.e., spacecraft aspect every 1.0 seconds. The performance of the ESXC flight software depends on the frequency of delivery and the precision of attitude information (plus temperatures) from the spacecraft. Good thermal/mechanical model of the spacecraft is necessary to calculate an aspect error budget. More temperature sensors is better than fewer temperature sensors.

HETE Lessons Learned (V): 

HETE Lessons Learned (V) HETE SXC mylar/aluminum Optical Blocking Filter (OBF) was destroyed by atomic oxygen on orbit; Beryllium CCD covers have been unaffected by orbital debris. ESXC will use a single Be cover over the CCDs.

HETE Lessons Learned (VI): 

HETE Lessons Learned (VI) The primary limitation in the precision of the HETE SXC is the knowledge of the SXC pointing. HETE spacecraft aspect is known to a precision of ~5” at 1 Hz. The aspect cameras are co-mounted to the SXC focal plane SXC localizations of bright, on-axis, orbit-midnight sources have ~20” accuracy; however, the inability to measure the systematic thermal distortions of the SXC+OPT system have led to uncertainties in SXC localizations of ~80” radius. Precise spacecraft aspect and knowledge of the spatial relation between the star trackers and ESXC pointing are essential for precise burst localization.

Lessons Learned (VII): 

Lessons Learned (VII) Reliable communication between team members is essential during burst analyses. The HETE team uses a text-based “chat line” during time-critical burst analyses. Team members in Japan, US, Europe can discuss burst details in real time with no delay (only the those who cannot type quickly suffer). The text of the chat is stored in the HETE data archive for reference.

Lessons Learned (VIII): 

Lessons Learned (VIII) Sco X-1 and the galactic bulge sources are very bright, resulting in higher background rates from April through August; this affects instrument sensitivity and downlink mass. HETE uses the “nod” capability to move the FOV of the WXM away from Sco and the bulge sources. E-SXC may also ignore CCD columns which are contaminated with Sco X-1 photons: this requires aspect knowledge in the E-SXC.

Lessons Learned (IX): 

Lessons Learned (IX) Estimates of PGS contact durations which assume that contact begins at AOS and ends at LOS so not properly account for signal acquisition times: nominal 10-minute contacts are actually 8- or 9-minute contacts…

Questions for Eclairs (and some answers): 

Questions for Eclairs (and some answers) How many star cameras are there, and how far are they from the ESXC? Two, about one meter. What is the PSD of spacecraft stability? .004°/.3s, .0075°/1s, .01°/5s, more answers are expected. Flight localization analysis requires steady flow of spacecraft aspect: what rate can we expect, what accuracy? 4 Hz, .02° (?) Can we nod? I think Bertrand said “yes”. Will it be possible to have a flight processor simulator, hardware of software?