Original System

The UKIRT Primary Mirror Support System, as commissioned, comprised 80 pneumatic ``bellofram" pistons arranged in three rings and operated as a single unit by a compressed air servo system. This was controlled by load cell readings from four so-called ``hard points" at 60% of the mirror's radius, which were maintained at prespecified values. The hard points applied significant forces (of order hundreds of Newtons) to the back of the mirror, but were so located as to produce minimal deformation and indeed, by careful elastic design in conjunction with the mirror and the cell, almost no astigmatism when in proper adjustment, albeit this condition could be tricky to maintain. They did generate appreciable higher-order deformations such as quattrefoil etc.

New Support System

A schematic diagram of the new system and its components, taken from a paper by Richard Bennett (rjb@roe.ac.uk) at ROE (1995). It was designed by a team from the RGO and installed by them in June 1994.

The modified system includes :

• A choice between operating the belloframs all together, as before, or in three independent sectors (now the default mode);

• A new servo placing zero loads on three peripherally-located axial definers;
  • This reduced most high-order aberrations (quatrefoil, cinquefoil, etc.) by factors of two or more

• New, peripherally-located tangential radial positioning system with load cells
  • offers effective diagnostics
  • allows location of the mirror w.r.t. cell to ± 0.1mm
  • freed important space in central hole for ventilation system

Enhancements to the support system also include an upgrade of the 24 radial support levers to equip them with new, minimum-friction ball-joints (which, it is hoped, will be less vulnerable than their predecessors to deterioration for lack of exercise).

The results of the upgrade - several aspects of which were quite bold departures - have been excellent, and no problems have so far been encountered, while the new system has provided a compatible operating environment for the active-optics figure control system (see below).

Active optical control of Primary Mirror figure

The new pneumatic support system enabled the installation in October 1995 of an active figure control system, the hardware and control elements of which were also provided by the RGO. Its control algorithms were developed by Nick Rees at the JAC and Richard Bennett at ROE.

The figure control system comprises twelve peripheral actuators which can apply axial forces of ±250 N to correct:

• astigmatism up to ± 6.9 microns P-V on the wavefront;
• triangular coma up to ± 1.8 microns;
• quatrefoil and cinquefoil up to ± 0.5 and ± 0.2 microns, respectively
• Spherical aberration up to ± 0.5 microns.

Initial results with active figure correction

In October 1995 the control system for the 12 peripheral actuators was installed and made operational (the actuators and their attachment points around the edge of the primary having been installed at the same time as the support system upgrades in June 1994). In a trial run the results of the FEA were confirmed: it was found that after a wavefront had been measured, simply applying the combination of forces indicated by the models corrected the surface very much as predicted. Indeed an entertaining video resulted, showing artificially-induced trefoil (low-order triangular coma) rotating smoothly under the control of the active system.

The figure control system was not brought into general use at that time as the dominant aberration even then was misalignment coma, which could not readily be corrected until the 5-axis controllable secondary mirror support system was installed with the new topend hardware.

First results with full active optical control

In August 1996 the major topend and bottom-end systems were installed, allowing coma to be removed by centralising the secondary mirror on the primary optical axis and aligning it to the correct tilt. The primary control system was then used in conjunction with the telescope wavefront sensor to tune the primary mirror, removing ~ 1 micron of astigmatism, ~ 350 nm of trefoil and ~ 350 nm of spherical aberration.

The Figure shows mildly-defocussed K-band images before and after this operation, which increased the intrinsic Strehl ratio from 0.23 to 0.53. The remaining limiting factors are mostly associated with the secondary mirror itself. (See: Description of UKIRT.)

A series of Wave-Front Sensor measurements are now underway - all-sky sets taken at roughly monthly intervals - which, combined with mechanical measurements of the alignment of the primary and secondary mirrors, will be used to construct lookup tables which are expected to provide near-optimum performance over most of the accessible sky.