Main Page | Modules | Data Structures | File List | Data Fields | Globals | Related Pages

focus_spectro

Compute the proper spectrograph focus from a series of arc exposures, using a gaussian fit of the auto-correlation peak.

Parameters:
-in|cat focus.cat Input catalogue or single image
-dev ASCII Output device DISLIN devices (e.g. CONS,EPS,etc.) are supported if configured with --with-dislin
-seq start,step Linear focus ramp to be used if the focus keywords are missing
-range 3. Compute the focus range at the 3%-level
Returns:
Error status (in particular, status=5 if focus position is outside the input range)
The input frames should have a FLOAT `SPECFOC' keyword giving the position of the spectrograph moving element (camera on the red channel, collimator on the blue one). If this keyword is not there, you can use the option -seq start,step to describe the focus ramp actually used.

The default output device is `ASCII', i.e. stdout:

 #   Focus    Intens     SigX     SigY    <Sig>   d<Sig>
 1   750.0  2.39e+10    1.242    1.574    2.005    0.041
 2   800.0  2.53e+10    1.205    1.506    1.928    0.039
 3   850.0  2.66e+10    1.174    1.459    1.872    0.038
...
If you want a graphical output, you should compile the Snifs project with the DISLIN library (http://www.linmpi.mpg.de/dislin/):
 ./configure --with-dislin=$DISLIN 
The available output devices are then the DISLIN ones: CONS (X-window), EPS, PDF, PNG, etc. (see complete list at http://www.linmpi.mpg.de/dislin/kap6.html#METAFL)

The value used for the best-focus determination is <Sig> (RMS of SigX and SigY). It appears that the focus position is not very dependant on the quantity used (e.g. intensity)

Note:
Use with the -quiet standard option for minimal output.

Todo:
Test (native SNIFS format) multiple-extension input frames

Test different zones in the arc frame (size and position, but beware of 0th and 2nd orders)

Fit multiple zones in the arc frame and derive a CCD-tilt

Check out why the Y-error bars are so different between the red and blue channels. The formal error estimates derived with nllsqfit_bnd seem theoritically correct.

Warning:
The auto-correlation method cannot deal with chromatic focus (since everything is naturally all mixed up in the auto-correlation peak)

Generated on Tue Nov 23 18:04:23 2004 for Snifs by doxygen 1.3.3