Don Mickey (mickey@ahinahina.ifa.hawaii.edu)
Wed, 24 Mar 1999 21:10:59 -1000
The Focus job has been re-implemented. It takes an image at each of a
sequence of L2 settings, recording the center 256 x 256 area of a bin-1
image. For each, the variance in data and geometry images are computed.
When the task completes, the data are sent to the XUI working directory
along with a template file giving the exposure time, L2 start and step
values, and the number of steps. An IDL routine, ivmfoc.pro, is in
lilikoi:~ivm/idl. It looks for the temporary files and fits quadratics
to the variance lists--then prints out the apparent best focus for each
camera and plots both the data points and the fit.
A sample plot from today is attached. The abcissa is in steps, but each
step was 5000 units in L2 position and there were 30 steps--from 160,000
to 310,000. About this many steps are probably needed to smooth out
seeing variations. The centers of the curves are at 244,000 for one
camera and 246,000 for the other, and it looks like if you're within
20,000 of the best setting you won't degrade the image much. Perhaps
this tolerance will be tighter if the seeing is really good.
--Don Mickey <mickey@ahinahina.ifa.hawaii.edu>
