iraf.net

I find that when trying to extract echelle spectra with dispersion
correction using arcs taken at periodic intervals during the night,
the rms of the residuals to the comparison line fitting are much
larger (~300 times for this case) for the arc after the stellar
exposure compared to the residuals from the identifications of the
arc lines taken before the stellar exposure. An example follows:ECREIDENTIFY: NOAO/IRAF V2.12.2a-EXPORT willmart@makalu Thu 16:20:38 10-Nov-2005
Reference image = a0002.ec, Refit = yes
Image Found Fit Pix Shift User Shift Z Shift RMS
a0004a0003.ec 902/902 889/902 -0.0655 -0.264 -4.5E-7 0.00407
Extract and reidentify arc spectrum a0007
Nov 10 16:20: DATABASE - 24 apertures read for a0004 from database
Nov 10 16:20: DATABASE - 24 apertures for a0007 written to database
Nov 10 16:20: EXTRACT - Aperture 1 from a0007 --> a0004a0007.ec
Nov 10 16:20: EXTRACT - Aperture 2 from a0007 --> a0004a0007.ec
Nov 10 16:20: EXTRACT - Aperture 3 from a0007 --> a0004a0007.ec
Nov 10 16:20: EXTRACT - Aperture 4 from a0007 --> a0004a0007.ec
..etc.ECREIDENTIFY: NOAO/IRAF V2.12.2a-EXPORT willmart@makalu Thu 16:20:41 10-Nov-2005
Reference image = a0002.ec, Refit = yes
Image Found Fit Pix Shift User Shift Z Shift RMS
a0004a0007.ec 899/902 899/899 -0.14 -0.566 -9.7E-7 1.24
Dispersion correct a0004
a0004.ec.imh: REFSPEC1 = 'a0004a0003.ec 0.85733211'
a0004.ec.imh: REFSPEC2 = 'a0004a0007.ec 0.14266792'
a0004.ec.imh: ap = 1, w1 = 5360.408, w2 = 5436.76, dw = 0.037336, nw = 2046
a0004.ec.imh: ap = 2, w1 = 5412.471, w2 = 5489.561, dw = 0.037697, nw = 2046
a0004.ec.imh: ap = 3, w1 = 5465.556, w2 = 5543.398, dw = 0.038065, nw = 2046
a0004.ec.imh: ap = 4, w1 = 5519.692, w2 = 5598.302, dw = 0.03844, nw = 2046
...etc.Note that a0002.ec is the master arc which was identified initially.
I thought I recalled from a few years ago that setting REFIT to "no" was
a workaround, but I don't seem to be able to turn it off. In fact if I extract
this second arc spectrum, a0007, and just ECREIDENT it with refit = no, it
seems to work:ECREIDENTIFY: NOAO/IRAF V2.12.2a-EXPORT willmart@makalu Thu 16:02:13 10-Nov-2005
Reference image = a0002.ec, Refit = yes
Image Found Fit Pix Shift User Shift Z Shift RMS
a0007.ec 899/902 899/899 -0.135 -0.546 -9.3E-7 1.24ECREIDENTIFY: NOAO/IRAF V2.12.2a-EXPORT willmart@makalu Thu 16:03:36 10-Nov-2005
Reference image = a0002.ec, Refit = no
Image Found Fit Pix Shift User Shift Z Shift RMS
a0007.ec 899/902 899/899 -0.135 -0.546 -9.3E-7 0.00454I confirmed that the pix shift is correct for a0007 by measuring the same
line in the two arcs and get 0.14 pix difference. The similar task DOSLIT
does seem to work with refit=yes.Thanks for any help wth this problem.Daryl