Transcript Autonomous Mobile Observing Stations

Autonomous Mobile Observatory Stations –
Development in Progress
Bill Hanna
IOTA North American Annual Meeting July 29, 2016
Stillwater, OK
Issues With Multi-Station Deployments
• There is a lot of “stuff” to deal with
• It takes time to align to the pre-point target
– Reduces the number of stations that can be set up
prior to the event, especially if it occurs shortly
after sundown
– There are workarounds for this, but at the
expense of visiting the observing sites multiple
times over more than one day
Concepts Being Examined
• Self-contained unit
– No (or almost no) fiddling with “stuff” on site
– Requires only an initial coarse manual alignment
• Able to steer the optical axis as necessary
– To the pre-point target
– Track thereafter?
• “Inexpensive”
– Commercial mounts can align themselves
autonomously, but at significant expense
Self-Contained Unit
• A single housing to integrate all of the components
–
–
–
–
–
–
–
–
–
–
–
Telescope
Camera
Battery / Power Controller / (Solar Cells?)
Executive Computer
Recorder
GPS / Time Inserter
Dew-Suppression Heaters
Cables
Optical Axis-Steering Assembly and Controller
Rain Sensor?
Security Sensors?
All of the Physical Components Are Available
• The amateur robotics community is an excellent
source for the mechanical components
– Rotational and Linear Servos
• Analog and Digital
– Stepper Motors
– Gears, Belts, Pulleys, etc.
• Wide selection of executive computer platforms
–
–
–
–
Raspberry Pi
Arduino
Micromite
Single-Board Computers
Initial Manual Coarse Alignment
• Want to point the default optical axis to the nominal EL@AZ from
Occult Watcher
• Manual coarse alignment, possible even in daylight, based on three
components
– Level
– Elevation
– Azimuth
• Referenced to a celestial body
– Sun
– Moon
– Bright star
• Referenced to a ground target
– Mountain peak
– Transmission tower
– Building / Structure
• Should easily get the default axis within 5°, perhaps within 2°
Freely-Available Information
• The easiest way to do something is to get
someone else to do it for you
– Sun AZ and Moon AZ from the USN Observatory
– Star positions from C2A or similar
– Ground targets from Google Earth
• Will require a “pre-processor” to collect (and
compress?) the necessary information for a
given event
– Shared among multiple deployed units
Optical Axis Steering
• Pan/Tilt Mirror Control
– Two rotational servos
– Worm gears
• X/Y Mirror Control
– Two galvanometers
• Housing Alignment Control
– Two (or three?) linear servos changing the lengths
of the legs
Pan/Tilt With Two Analog Servos
• The cheapest option, but is it suitably accurate and
repeatable?
• Lynxmotion SSC-32U USB Servo Controller
– 32 servo channels, in two banks of 16 each
• Provision for separate voltages on the two banks
– Accepts ASCII text commands via USB or RS-232 serial
– Controls 500µs to 2500µs pulsewidths in 1µs increments
– Additional functions using minimal external components
• Hitec HS-422 analog servos
– Speed and torque are not significant factors
– Want a small deadband
Pan/Tilt Test Rig
• Pan/Tilt assembly mounted over an Orion 80mm Short
Tube OTA
• First-surface mirror to minimize false images
– Edmund Scientific
• Astrovid StellaCam3 (Peltier-cooled Watec-120N+)
• Pinnacle Dazzle to VirtualDub on a Lenovo laptop
• SSC-32U commanded via USB to a virtual COM port
using Tera Term
– 2000 1µs counts = 180 degrees = 10800 arc-minutes
– 1 1µs count = 5.4 arc-minutes
Pan/Tilt Mirror w/ Analog Servo Evaluation Rig
Pan/Tilt Assembly Close-Up
Pan/Tilt Target Ruler
Results
• Arbitrarily chose 30 arc-minutes (the apparent
diameter of the moon) as a repeatability goal
• Reasonably accurate and repeatable, to within ~16 arcminutes (3 1µs counts) but only after making a large
(≥ ~5-degree) slew
• Small moves (such as those necessary for tracking)
were essentially impossible due to the typical ~5µs to
~8µs deadband of a standard analog servo
• Good position holding after removing power to the
servo
– The mirror assembly has an extremely small mass
• Overall assessment: good, but not good enough
Next Test Will Use Digital Servos
• More expensive
• Need to be programmed individually prior to use
– Only need one programmer
• Higher power consumption
• The deadband can be programmed to zero at the
expense of continuous “hunting”
– This will have a minimal impact on the servo lifetimes
because the powered use will be relatively intermittent
• Significant parameters to be evaluated
– Position hold after power removal
– Small motion (i.e., tracking) performance
• In particular, the effect of any resulting vibration on image stability
Software Issues
• If using an analog camera, need a reliable
driver for the various available video digitizers
(e.g., eMPIA Technology em28xx chips) for
various small platforms (e.g., Raspberry Pi)
– Able to capture a single frame
• Plate-solving routines
Looking Forward
• Design for the future
– Internet of Things (IoT)
– 3D Printing
• The number of different cameras currently (and
projected to be) in use may mean that it will be
simpler to have a separate (and standardized)
camera for the alignment process
– Also means that only a single version of plate-solving
software will be necessary
I Invite Questions and Comments
• Best to contact me at: [email protected]