||Robotic Operation - Overview|
The observatory was designed from the beginning to support robotic operation. This means both that it is possible to operate it remotely, and also that observing sessions can be automated and carried out with little or no human oversight.
Cabling is a key part of the system. There is a lot of electronics on the telescope, and therefore a lot of wires that need to be routed from moving to stationary parts. This needs to be done very carefully, both to eliminate changing forces that can interfere with smooth tracking and to avoid interference that could be disastrous in an unattended setting. Here are some cabling details.
This has heavily influenced my choice of instrumentation. Everything utilizes either serial or Ethernet interfaces. Other interfaces, such as USB or parallel are simply not suitable for remote control operation.
Guiding a remotely operated telescope is also a tricky business. The best guiding is normally done using the same optical path that the imaging camera uses, either with an off-axis guider or with some type of self guided camera using either the main sensor or an adjacent sensor. But this is problematic for remote operation because the guider field of view is small, and rotating the guider is difficult. Also, there is a substantial loss of guider sensitivity that occurs when guiding through imaging filters. My solution to this is to use a piggyback guidescope with a very short focal length (F=200mm, f/5.1). This results in a guider that covers a wide field of view (greater than one degree) and never needs to be aimed or focused. The guide camera operates at 7.3 arcsec/pixel, and the imaging camera at 0.8 arcsec/pixel. I find that this ratio gives excellent guiding performance. There is a long standing rule of thumb among astroimagers that the guidescope focal length should be at least as long as the imaging scope, but this applies only to manual guiding, an obsolete technique that is seldom practiced. Cameras used for automated guiding can determine the location of a guide star to a small fraction of a pixel, so guidescopes can be quite short. To minimize problems with differential movement, I have rigidly mounted both the guider camera and its objective lens to a rail. My primary limitation is now movement of the primary mirror, which normally constrains me to subexposures of 30 minutes or less. This is not a problem, however, since I have found that subexposures of 10-20 minutes work best for me. Here is an image of my guidescope system. Normally there is a flocked tube between the objective and the camera to block stray light, but it has been removed for clarity in this picture.
The observatory is on my property-wide network, and all of the instrumentation is controllable from any computer on the network. I have had some problems with lightning ground strikes damaging the network, so I am considering switching to a wireless network, even though that will be somewhat slower.