...
- partial credit for imperfect observations
- look-ahead needed, or not?
- HA weighting, for 24-hr urgent fields. (HA weighted by night or annually)
- homogenization
- we should do single band version of this problem
- get field centers.
- 10 year Op Sim now takes 2 days to run- bottlenecks are look-ahead and slew rate computation.
- rotator and az cable wrap are the two variable things for slew computation.
- repulsion from moon, as a function of passband
- different aspects of look-ahead: sky, moon, homogenization, twilight, etc
- fallback plan of march along the meridian in a fixed band.
- partial credit for mediocre observations
- get around the "science proposal" mentality
- We need to request d(merit)/d(m5) for all science goals.
- They claim they will do actual real-time pointing decisions, that they can run at 2000x real time now, so they can keep up with telemetry stream.
- Predictive cloud motion.
- currently 120 sec for filter change.
- 18,000 sq deg is default area
- minimum slew is 5 sec
- dome is pacing item
- optics tweak takes 34 seconds, if elev slew exceeds some threshold
- az wrap is cumulative, just track it and add it up. Add penalty for az unwind
- assume rotator wrap is no issue
- SRD median slew is 5 sec.
- Field tessellation has closer distance between field centers than we can tolerate- depends on whether we dither in offset or rotation. In other words 18,000 sq deg/N_fields is substantially less than camera FOV.
review questions:
- How are clouds handled? minimum of two filter changes per night, due to twilight, etc.
- Kem claims we now have 94.2% of available visits.
- need adjustable readout time and noise for first vs. second exposure
- What is histogram of slew times between az,alt pairs?
...
SSTAR user documentation at NOAO
configuration properties for LSST (max rates, etc)
Field_table.txt - ASCII file with: fieldID fieldFov fieldRA fieldDec fieldGL fieldGB fieldEL fieldEB, GL abd GB are Galactic l,b and EL EB are ecliptic l,b.
...