As part of our continuing effort to improve the imaging performance of the 4-m Mayall telescope on Kitt Peak we have implemented a variety of changes including a cooling system for the primary mirror and oil pads, dome ventilation, and most recently active control of the primary mirror support. Past wavefront measurements at the 4-m Mayall have shown that low order aberrations are degrading the delivered image quality by 0.10-0.20" FWHM, depending on seeing and where the telescope is pointed (see Figure 1). The modeling in Figure 1 clearly shows that under typical site seeing, significant image quality improvements can be obtained by removing the low order aberration from the delivered wavefront. We ultimately aim to have the optical performance of the 4-m Mayall telescope contribute no more than 0.05" FWHM to the delivered image quality in 0.60" seeing at each of its three foci. To this end, during the recent 1998 Summer Shutdown on Kitt Peak we have installed independent pressure control to the primary mirror's 33 axial support pads, similar to the system installed at the CTIO 4-m Blanco telescope (Schumacher, G., et al., 1995, SPIE Vol. 2479, 389).
Figure 1: The three imaging performance models for = 0.65
m shown are
computed by convolving the diffraction PSF from actual 4-m wavefront
measurements and a Gaussian seeing model. The FWHM of the resulting image is
measured using the IRAF task PSFMEASURE and compared against the input
seeing. The wavefront data for the uncorrected PSF was obtained at the
Mayall Prime Focus while pointing far into the South and represents a
relatively extreme case. In this instance the bulk of image degradation
comes from astigmatism. The remaining low order aberrations represent an
incremental change in image quality.
The axial support of the 4-m primary mirror is facilitated through 36 locations arranged in two concentric rings of 12 inner and 21 (plus three defining units) outer air pads. Originally, support pressure in the two rings was determined using two passive regulators, which compensated for changes in the telescope zenith angle. By manipulating the pressures in the array of axial support pads independently, we have gained control over the low order bending aberrations, astigmatism, trefoil and quad-astigmatism, in the delivered wavefront. Forces necessary to obtain a best fit of the low order bending aberrations have been modeled by Earl Pearson at NOAO using Finite Element Analysis. Because of its older style geometry, the 4-m primary mirror is relatively stiff to forces applied to the inner ring of pads; hence only the outer array is used to manipulate the bending shape of the mirror. This is one way the system as implemented at KPNO differs from the CTIO system, which used thin plate theory to determine the force patterns. Other differences include wavefront measurement using curvature sensing rather than Shack-Hartmann and the algorithm used with the pressure controllers. The other low order aberrations present in the optical system, coma and spherical, will be corrected using other methods as discussed below.
First "bend" with the new system came in September during initial system
checkout and verification. However, true commissioning of the new system
began during the recent October T&E time. As a demonstration of the system's
capabilities, we obtained the focus sequences shown in Figure 2, with known
amounts of the three bending aberrations induced. These data were obtained
with the telescope pointed near zenith using T1KA as a direct imager at the
4-m f/8 focus through a Harris R filter. From left to right the sequences
are: unperturbed for reference where best focus is 0.68" FWHM,
1.0 m RMS
astigmatism, 0.67
m RMS trefoil,
0.4
m RMS quad-astigmatism.
Figure 2: Four focus sequences showing the new 4-m support system in action (see text for details).
During the October T&E run we also measured the precision of the FEA modeling against recovered wavefront error. To do this we made multiple wavefront measurements for a given induced error over the full dynamic range of allowed pressures and forces. The results of these measurements are shown for the Zernike sine term of astigmatism, and trefoil in Figure 3 below. In the case of astigmatism the plot shows the semi-amplitude of the system dynamic range. The linearity of response is exceptional, as is the agreement between the predicted induced and recovered errors. It is also important to note the isolation between sine and cosine terms in each aberration, indicating that the system is orthogonal.
Figure 3: Calibration between predicted induced wavefront errors and those recovered using curvature sensing. In most cases the error bar for a given datum is within the point plotted.
Our active support system will have three basic modes of operation: 1) Emulation of the previous support system, which compensates only for the effects of zenith angle. 2) Position dependent correction of low order aberrations based on mapped look-up tables. 3) Correction of low order aberrations as determined from the look-up tables plus near real-time wavefront data. We expect to have modes 1 and 2 commissioned and operational at all three 4-m foci following the T&E time scheduled this semester and available for use in the 1999A semester. The third mode will become available following the commissioning of a new dedicated Cassegrain wavefront camera being built by KPNO and is expected to be operational for the 1999B semester. In addition, we are investigating additional active control of the three primary mirror defining units and/or the secondary mirrors to tune collimation and control of comatic aberration. With the result obtained thus far, we expect the new support system to result in significantly higher image quality for the 4-m Mayall telescope.
Chuck Claver
(for the 4M Active Primary Support Team)