The examples below show typical wavefront images taken on the night of 25 May 1998.
As the dynamic range of the primary mirror actuators is not sufficient to correct simultaneously all of the astigmatism, r3 trefoil and spherical aberration currently present, a compromise setting is being used. This leaves some uncorrected trefoil and spherical aberrations, for an overall rms error of 360 nm on the wavefront.
The image to the right shows the residuals after Zernike polynomials up to Z23 have been fitted to the wavefront and removed analytically. After this process the rms error is still 140 nm, equivalent to the diffraction limit (formally, lambda/14) at 2 microns, which is the performance goal of the Upgrades programme.
The main features of the residual image are three high spots close to the central hole and ~ 120º apart. These are induced by differential contraction of invar mounting units, which are cemented to the zerodur secondary, between glue-setting and operating temperatures. This produces the excess spherical aberration, which currently uses up much of the dynamic range of the primary active optics system, and r5 and higher-order trefoil, which are not correctable by the active optical system. Also in the residual image are small “dimples” in the wavefront (a.k.a. the “waffle”), due to print-through of the light-weighting honeycomb pattern on the back of the secondary, which have delayed implementation of Adaptive Optics on UKIRT. Lastly, the current secondary has a turned-down edge, which is not very apparent on these images, but impairs the ultimate achievable performance.
The replacement secondary, expected early in 1999, will not have a turned-down edge, will be stress-relieved to remove print-through and will be equipped with athermal mounts to remove the high-order trefoil and excess spherical aberration. A substantial further improvement in intrinsic optical performance should result.