MROIFastTipTilt.TechMeet8 History
Hide minor edits - Show changes to markup
June 15, 2010, at 05:58 PM
by jsy1001 -
Added lines 4-5:
June 15, 2010, at 05:56 PM
by jsy1001 -
Added lines 20-22:
- Prefer to use floating point calculations for centroiding and servo
- EBS to find out if this means we must use Xenomai user-space real-time and hence tolerate context switches
Deleted lines 29-31:
- Prefer to use floating point calculations for centroiding and servo
- EBS to find out if this means we must use Xenomai user-space real-time and hence tolerate context switches
June 15, 2010, at 05:55 PM
by jsy1001 -
Changed lines 10-12 from:
- Frame acquisition and data processing should be a real-time task in *all* modes
- Not necessarily for FLC, depends on availability of RT camera data driver
to:
Added lines 17-19:
- Frame acquisition and data processing should be a real-time task in *all* modes
- Not necessarily for FLC, depends on availability of RT camera data driver
Added line 23:
- Same process as frame acquisition/data processing
June 15, 2010, at 05:54 PM
by jsy1001 -
Added lines 1-34:
Technical Meeting #8: Software conceptual design
This meeting took place on Tuesday 15 June 2010. Attendees: JSY, DFB, EBS, MF, ADR
Conclusions
- GUI should be separate process
- Allows display on console for best performance
- Can be started and stopped at user's convenience
- Frame acquisition and data processing should be a real-time task in *all* modes
- Not necessarily for FLC, depends on availability of RT camera data driver
- Thermal control
- Scope of this software includes thermal control of camera enclosure
- Thermal control should be active continuously, day and night
- Does this include Peltier?
- Separate process for thermal control (possibly with its own interface to the ISS) desireable, but requires dedicated I/O hardware
- ISS Interface
- Most commands (including mode switches) can be synchronous - avoids extra threads/callbacks
- Exception would be any commands to move actuators
- May need to account for future off-axis offsets when choosing initial position of subframe on FTT Mode start
- Prefer not to have ISS perform calculation
- Prefer to use floating point calculations for centroiding and servo
- EBS to find out if this means we must use Xenomai user-space real-time and hence tolerate context switches
New Tasks
Medium Priority
- EBS: Design real-time task in more detail
- JSY: Investigate use of multiple threads in ISS framework
- JSY: Ask Allen if systems are allowed to run indefinitely
|