iraf.net

The deblending of lines can be slightly tricky. However it does fit
simultaneous Guassians of common width but DIFFERENT centers and
strengths. There is a bug when using "fluxed" spectra with intensity
values of 10e-13 and smaller. This is due to arithmetic operations
which can underflow. If this is your problem (though the symptom is
usually a crash) the work-around is to first multiply by a large factor
(using the splot 'f' for example) to obtain values near unity. If one
line is very weak it may be that the fitting is moving the position of
the weak line very close to the strong line. Are the reported centers
the same? You can try fixing the separation of the lines (admittedly a
bit of a pain) and see if you get a better result.The other equivalent widith routines do not perform least squares fits and
blending is accomodated by the one-sided options. You will have to consult
the help page to determine the algorithm appropriate for your needs.If you would like me to look at the problem in more detail we can
arrange to login via span or you can use the task "listpixels" with
output redirection to a file and then send me the pixel values by email
(if the spectra have many points you can use an image section or edit
the file to just send the region of interest).Yours, Frank Valdes 602-325-9219