Spar Vibration Damping

Spacecraft use active damping of vibration, via momentum wheels and other systems, to remove spurious motion. Normally, ground telescopes use the drive system to actively damp motions induced by mechanisms on the telescope and by wind buffeting. The Mees spar has an undersized drive system, prohibiting high frequency damping.

We use fast guiders in the various instruments to compensate for high frequency vibration. That scheme requires a complete control system for each telescope. The full disk and coronagraph systems do have fast guiders; the range of vibration can break the guider lock in the MCCD and IVM systems.

An alternate is to actively damp spar vibration with a separate system, not the telescope drive. Linear actuators driving test masses are almost ideal, working well at high frequency and being insensitive to low frequency. The telescope tracking is unaffected and the vibration to all telescopes is reduced with a single control system.

Links to system and component suppliers are:

Cymer ACX - vibration suppression, piezo based

Kinetic Ceramics - piezo actuators

Planning Systems - vibration control systems, electromagnetic

Taylor Devices - shock absorbers

BEI Kimco - Voice coil actuators

Progress to date:

Estimate of force, stroke - Dec 23 1998

Test fixture - Mar 8 1999

Spar Guider Upgrade

Proposal September 98

For some time the performance of the spar guider has been considered to be sub-optimum for several reasons, including poor disturbance rejection, hunting, and mechanical vibration. Several years ago a project to replace the present DFM system with an image-based system was begun but abandoned for lack of resources, though some hardware components were purchased and engineering begun. Recently discussions have suggested a more limited project may be feasible, replacing the signal processing and electronics with modern components to achieve better performance and reduced vibration.

Present System

The existing spar guider system utilizes a mechanical chopper and 4 solar cells to generate a pointing error signal from a solar image. The collecting optic has an aperture of about 1.5" and a 4m focal length. This results in an image of about the same size, and due to the small aperture nearly all wavefront distortion in the incident image is seen as tilt, making the system very sensitive to seeing.

Opposing pairs of solar cells are differenced and the resulting amplitude-varying AC signal is demodulated and converted to pulse and direction signals (via a voltage to frequency converter) which drive stepping motors actuating a tangent arm to correct the DEC error and driving a differential gearbox against a clock motor to correct RA errors. An analog filter is used to provide some stabilization to the closed-loop system.

Problems:

The control law is no longer well matched to the spar because of changes in the spar mass as instruments have been added. This type of feedback loop typically has a 1st order "Proportional-Integral" response (because generating an analog derivative signal is difficult). The lack of a derivative damping component in the control loop requires the closed-loop bandwidth of the system to be very low for stability (primarily due to integrator windup against the very large inertia of the spar).
The drive generates a large amount of impulse vibration into the spar from the DEC motor assembly due to the high cogging torque of the stepper motor. The motor is located in such a way that this vibration is well coupled into the IVM instrument package, potentially degrading the image quality of that instrument.
The system's sensitivity to seeing and poor disturbance rejection cause the spar pointing to hunt when seeing is poor. Because seeing-induced image motion is unrelated to actual spar pointing errors and too fast for the drive system to correct, these signals tend to be amplified and cause the spar to wander off of the sun (the "groaning" sound heard in the dome is the drive oscillating back and forth across the true pointing angle as the error signal is perturbed by seeing).

Proposed solutions

During the earlier project a PC, motor controller and micro-stepping amplifiers were purchased and remain available to the project. A modest amount of effort would be required to design a suitable signal digitization scheme to provide the signals from the limb sensor in the PC, and to install the motor drivers. This would address the problems as follows:

Implementing the control law/filter in a digital form allows much more flexible tuning than an analog system, and additionally allows design of much more sophisticated tracking routines such as incorporating dual time constant sliding mode control. This would substantially improve the spar tracking by allowing the controller to be matched to the spar inertial response.
By driving the stepping motors in "micro-stepping" mode at a factor of 10x smaller steps the vibration of the system should be reduced enormously. Because all 4 motor phases are driven the "cogging" of a standard step motor is eliminated and lower motor heating is experienced.
Digitizing the limb sensor signals and processing them in digital form allows filtering the high-frequency seeing-induced components off of the error signal used to correct the spar pointing. In addition these signals can be extracted and made available to other instruments for use in fast guiding applications, for example in the absence of a sunspot.

Implementation Steps

Review existing electronics schematics to determine if limb signal can be extracted in a useful place; look at signals with scope to make sure their characteristics are understood. (MW,LH)
Locate solar cells for alternative secondary detector, investigate design issues. (MW)
Analyze digitizer location (in spar vs in PC) impact on design & performance. (MW)
Rough out software design for guider system. (MW,EK, etc.)
Resurrect PC host platform (components, OS) (MW,EK)
Define cabling to interface drive amps to motors.

At this point, review feasibility of design and freeze parameters before proceeding to implementation/fabrication.

Design & build sensor assemble (or modify existing sensor package).
Fabricate chassis to mount Amps and power supply.
Make up cables.
Design & code data acquisition and motor control program.
Integrate, install, tune, document.

Last modified: Thu Sep 27 03:37:35 HST 2001