iraf.net

I am doing time-series spectroscopy measuring pulsation velocities using fxcor. As such, our timing is incredibly important. I have 2 issues related to this:
The txtonly output of fxcor provides time in HJD. However, I want JD output because then I will apply a more accurate barycentric correction later. Can I get fxcor to provide JD rather than HJD output? I tried editing the keywords and just changing the HJD keyword to JD, but then the fxcor output became INDEF. I can use a script to change the HJD value to the JD value *prior* to running fxcor, but I'm concerned of what that will do to the velocities. Does fxcor use the HJD for anything or just the VHELIO value? If fxcor does not really HJD for anything, then this might be a solution.Issue #2: In order to run refspec, the image headers must have HJD, and so I run setjd before refspec. However, then when I later run rvcor, it seems to add half the exposure time again. So my pre-rvcor and post-rvcor HJD values differ by 1/2 of the exposure time. Thanks in advance for the help.
Mike Reed
Missouri State University

Here's a new issue. running setjd on the same image, same version of IRAF, different computers (using the same Linux distribution: FC3) output different results:On computer 1 I get:
# Image jd hjd ljd
# SETJD: Observatory parameters for Roque de los Muchachos, La Palma
# timezone = 0
test 2453843.37380 2453843.37616 2453843On computer 2 I get:
# Image jd hjd ljd
# SETJD: Observatory parameters for Roque de los Muchachos, La Palma
# timezone = 0
test2.fits 2453843.37380 2453843.36955 2453843The observatory information is the same:
Computer 1:
# Observatory parameters for Roque de los Muchachos, La Palma
observatory = lapalma
timezone = 0
altitude = 2327
latitude = 28:45.5
longitude = 17:52.8
name = 'Roque de los Muchachos, La Palma'Computer 2:
# Observatory parameters for Roque de los Muchachos, La Palma
observatory = lapalma
timezone = 0
altitude = 2327
latitude = 28:45.5
longitude = 17:52.8
name = 'Roque de los Muchachos, La Palma'and in the header, the UT and date-obs are the same:
Computer 1:
DATE-OBS= '2006-04-17' / End of observation
UT = '20:58:04'Computer 2:
DATE-OBS= '2006-04-17' / End of observation
UT = '20:58:04'How can this be?!!!!
Thanks,
Mike

Hi Mike,The HJD value is strictly an output value of the FXCOR task, basically the JD of the midpoint of the exposure corrected to the Sun. SETJD is what you would use to add the JD time to the header based on the date and exposure time.FXCOR does try to compute the midpoint of the exposure so if it is again adding half the exposure time then it probably means there is some confusion about the DATE-OBS keyword and the UTMIDDLE keyword, both of which are controlled by the KEYWPARS pset. If no UTMIDDLE is found, DATE-OBS is assumed to be the start-exposure time and half the exposure time is added to get the UT middle time. Note that SETJD will also set the midpoint time.As for the last SETJD producing different values: I can't say much without knowing the IRAF version and platforms (which linux involved)? The difference corresponds to about 40min which might be a clue if you look at the exposure time. Otherwise this is larger than what I'd expect from floating-point precision differences. I'd also try unlearning both the SETJD and OBSERVATORY tasks to be sure you're using and resetting the default parameters.Hope this helps,
-Mike

Hi Mike,
Thanks, that pretty much answers the problem. The bottom line is that FXCOR will not print JD as part of the output but only HJD.
The second problem it turns out I had stated incorrectly: yes setjd produces a JD and HJD at the integration midpoint, but I was wrong on the direction rvcor moves. When rvcor is run, it *subtacts* off 1/2 the integration time, making HJD at integration start time. Perhaps this is because we do not have the midpoint keyward and rvcor ignores the previous HJD value and overwrites it with its new calculated value.
The different setjd values have been cleared up- we had an incorrect RA calculation and that was messing it up.
So for me, the problems are essentially solved: Since FXCOR will not include JD as part of its output, the RVCOR change in HJD is irrelevent since we have to cut and paste the JD time we want from the images anyway (via scripts).
Thanks, I think I'm in good shape, I just wanted to point out the SETJD and RVCOR inconsistency because it could catch people unaware (like it did me).
Mike