In this paper, we describe the details of control unit and GUI software for positioning two filter wheels, a slit wheel and a grism wheel in the ADFOSC instrument. This is a first generation instrument being built for the 3.6 m Devasthal optical telescope. The control hardware consists of five electronic boards based on low cost 8-bit PIC microcontrollers and are distributed over I2C bus. The four wheels are controlled by four identical boards which are configured in I2C slave mode while the fifth board acts as an I2C master for sending commands to and receiving status from the slave boards. The master also communicates with the interfacing PC over TCP/IP protocol using simple ASCII commands. For moving the wheels stepper motors along with suitable amplifiers have been employed. Homing after powering ON is achieved using hall effect sensors. By implementing distributed control units having identical design modularity is achieved enabling easier maintenance and upgradation. A GUI based software for commanding the instrument is developed in Microsoft Visual C++. For operating the system during observations the user selects normal mode while the engineering mode is available for offering additional flexibility and low level control during maintenance and testing. A detailed time-stamped log of commands, status and errors are continuously generated. Both the control unit and the software have been successfully tested and integrated with the ADFOSC instrument.
In this paper, we present the work on characterization of friction in the 3.6 m Devasthal optical telescope axes. The telescope azimuth axis is supported on a hydrostatic bearing while the altitude and rotator axes are supported on hydrodynamic bearings. Both altitude and azimuth axes are driven directly by high power BLDC motors and the rotator is driven by BLDC motor via a gearbox. This system is designed by AMOS, Belgium and tuned to achieve a tracking accuracy better than 0.1 arcsec RMS. Friction poses control related problems at such low speeds hence it is important to periodically characterize the behaviour at each axes. Compensation is necessary if the friction behaviour changes over the time and starts dominating the overall system response. For identifying friction each axis of telescope is rotated at different constant speeds and speed versus torque maps are generated. The LuGre model for friction is employed and nonlinear optimization is performed to identify the four static parameters of friction. The behaviour of friction for each axis is presented and the results are discussed.
We describe the details of telescope control system design for the 50/80 cm Schmidt telescope at the Aryabhatta Research Institute of Observational Sciences. The overall control hardware architecture features a distributed network of microcontrollers over controller area network for interfacing the feedback elements and controlling the actuators. The main part of the hardware is a controller whose final objective is to provide position control with 10 arcsec accuracy and velocity control with 1 arcsec/s accuracy. For modeling and simulation, the telescope parameters were experimentally determined. A linear proportional integral (PI) controller was designed for controlling the twin-motor drive mechanism of the telescope axes. The twin-motor drive is provided with differential torque for backlash-free motion reversal. This controller is able to maintain negligible rms errors at all velocities. At higher speeds over 2 deg/s , the PI controller performs with peak errors less than 1%. Whereas at fine speeds, depending upon the preload on bearings, limit cycles are exhibited due to nonlinear friction posing control related problems. We observed that the effect of nonlinear friction dynamics can be reduced by reducing the preload on the drive bearings and the peak errors at fine speeds using a linear controller can be maintained within 25%.
In this paper, we describe the details of telescope controller design for the 50/80 cm Schmidt telescope at the
Aryabhatta Research Institute of observational sciencES. The GUI based software for commanding the telescope
is developed in Visual C++. The hardware architecture features a distributed network of microcontrollers over
CAN. The basic functionality can also be implemented using the dedicated RS232 port per board. The controller
is able to perform with negligible rms velocity errors. At fine speeds limit cycles are exhibited due to nonlinear
friction. At speeds over 3.90 × 10-02 radians/sec, the PI controller performs with peak errors less than 1%.