Subject: new GRIM focus script and associated IRAF task!


Submitted: Wed, 6 Mar 1996 01:18:14 -0700 (MST)

Message number: 9 (previous: 8, next: 10 up: Index)

New focus script for GRIM, and an associated iraf task to 
help look at the data!!!!!

************************** GRIM focus script ***************************

Here's how to run a GRIM focus using the tcl script grimfoc.tcl:

first, in an mcnode type:  

		source ~bergeron/tcl/grimfoc.tcl

Then, acquire a star, center it vertically, and get an exptime that doesn't
put the star more than 28K above bias (~36k total for an IRAF image, ~26k
total for a fits image), but still has good signal. This can be done in any
f/rat you desire. Zero out your offsets (if any) by issuing a "sync ref" and
then re-slewing (even if you used inst offsets that aren't reflected in the
remark position window, its safer to zero before starting the focus).  Then
run grimfoc in the mcnode as follows:

grimfoc 5 20 -50. 50. 9 7.0 gfoc. 

The arguements in order are: f/rat (5,10 or 20), offset in *pix* between
images, starting focus value, focus step size, number of focus steps to take,
exptime, and rootname of output images.

The output images start with seq 1. There is a double sized gap before the
*last* image in the set, just like for DIS. Also note that the sequence will
start at the bottom of a frame for f/5 and at the top of the frame for f/10.

To reduce the data, use the grimcombine task in the new apo package on 

ap> epar grimcombine

Set the inmput images parameter to be all the focus images (eg. gfoc*.hhh)
Set the output image to whatever you like (eg. gall.hhh)

This task makes a sky flat from a median of your imput images, subtracts it
from each image, and then averages the images together, giving you a "sum"
off all the focus frames. (by the way, this task will work on any image set,
not just grim).

Then just display and imexamine the output image, just like you normally
would for a DIS image.

Some tips: Use a 50 micron step for focus in f/5. 25 is probably too small to
beat the plate scale and S/N in f/5. In f/10, 25 works well. I usually use a
9 step sequence. A 20 pixel step works well in f/5, and in f/10 if you are
using a 25 micron focus step and the seeing is reasonable. Probably safer to
use 25-30 pix in f/10. I haven't tried f/20 yet, but I guess you could get
really fine focus resolution there. Running the script takes about 2

Best tip of all - just ask the operator to do this for you :)




Louis E. Bergeron (Eddie)                   /\   *    *      *||  *
Apache Point Observatory                *  /  \    *    ^   ==()== 
P.O. Box 59                              * /  \  *   * / \   *||   *
Sunspot, NM 88349-0059               *    /    \       / \  *    *  
(505) 437-6822                          * /    \    * /   \   *
                                   ,_ o  *  ||    *     ||     *  *     / //\,       Ski New Mexico !!
APO APO APO APO APO  Apache Point Observatory 3.5m  APO APO APO
APO  This is message 9 in the apo35-grim archive. You can find
APO  the archive on
APO  To join/leave the list, send mail to
APO  To post a message, mail it to