Phil Massey, Tony Abraham, Bruce Bohannan, Chuck Claver, Richard Green, George Jacoby, and Richard Wolff
Over the past few years we have been working hard to improve the delivered image quality (DIQ) at the 4-m, the "seeing" the astronomer measures. Starting in early 1996, we began nightly seeing measurements using a dedicated CCD camera and PC-based software, implemented by G. Jacoby. We have used these data, along with the recorded temperatures, to determine how to best spend our resources in trying to improve the situation. We were encouraged by the eventual success of similar efforts at the CTIO Blanco 4-m, although conversations with the staff confirmed our own impression --- one has to do everything before there's any noticeable improvement.
We appear to have finally made it past that hurdle. The median seeing has dipped below the psychological 1.0" barrier, and we see corresponding gains in both the 25th and 75th percentiles --- in other words, the best seeing, the worst seeing, and the average seeing have all gotten better. The significance of this is striking when we compare the data semester by semester.
Caption: The median seeing at the Mayall 4-meter dropped below one arcsecond in semester 1999B, the best seeing has gotten better, and the worst seeing has too.
Below is a table:
Semester Number of Points Delivered Image Quality Improvements Implemented (1) (arcseconds) Median 25% 75% Prior to 1996A (2) 1996A 116 1.10 0.92 1.36 1996B 142 1.03 0.88 1.38 1997A 120 1.11 0.95 1.34 1997B 185 1.08 0.88 1.37 April 1997 (3) 1998A 211 1.08 0.89 1.32 November 1997 (4) 1998B 169 1.11 0.88 1.44 1999A 202 1.12 0.91 1.42 1999B 123 0.98 0.77 1.22 Summer of 1999 (5)
Notes: (1) Semesters "A" include the period 1 Feb through 1 Jul, while semesters "B" include the time period from 1 Sept through 31 Jan of the following year. The values correspond to a 10-second exposure through an "R" filter. We aim to obtain three measurements at night when the telescope is at f/8 and f/15; in practice we average about two. (2) Prior to the 1996A semester when we began regular DIQ measurements, we had begun using two large dome fans for sucking in outside air; were using a chiller to cool the horse-shoe bearing oil, had begun performing rigorous optical collimation at the end of each summer shut-down, had insulated the windows of the 4-m control room, and had begun cooling the 4-m primary during the day. Of course, the observing floor had been chilled during the day from the beginning of the 4-m operations. (3) In April 1997 we implemented a dome air mixing fan to bring cold air from the chilled floor to the upper part of the dome and prevent stratification during the day. (4) In November 1997 we began using 22 louvered dome vents to provide passive air mixing during the night. (5) During the summer of 1999 we installed a new mirror cover, and began extracting air over the primary. We also began running the new mirror support system in "passive emulation mode."
Of course, it could just be that the site seeing was spectacularly good this past semester; we have no way of knowing this as WIYN has discontinued nightly seeing measurements while waiting on a new wave front camera to be installed.
Caption: What made the difference in the delivered image quality at the Mayall? Was
it the new mirror covers that allowed air to flush over the primary? Was it
the dome vents that blew fresh air into the dome that could flush over the
primary? Was it the air extraction system that exhausted air from over the
primary, breaking up any residual thermals? Or was it a combination of
everything we have done? Shall we experiment?
We are pressing ahead with additional improvements --- a better primary mirror cooling system is being installed in late January, and this summer we will implement slow control of the tilt of the f/8 secondary to correct for systematic decollimation around the sky. Plus, our new 4-m active primary support system "4mAPS" is now fully operational, and we have begun running using look-up tables at all three foci in order to remove low-order aberrations.