Mon, 06/15/2020 - 11:39
Hi everyone, I need help, I have problems sending spectra, every time I get this error "star not in database" I also do all the possible name changes but useless, I contacted the service center and I was helped to insert the work, it seems that my Isis software and Demeter generate a double JD key, honestly I can't disturb the service center every time, could you help me?
Massimo Di Lazzaro
While we at HQ are always welcome to help what would be useful for the community at large is a way to edit fits headers. Is there a reliable program for this? I know you can do it through various programming languages, but if anyone knows of one with a gui interface, it would be very helpful.
Staff Astronomer, AAVSO
F4W2HDU is a free FITS header edotor that I've found useful & easy to use. It can be downloaded from https://astroshed.com/software.html . It's intended for bulk changes of multiple file headers, but works fine for a single file.
I just checked that BASS also can edit the FITS header on both image & profile files. It also allows you to add new keywords (which can be dangerous if you don't format them correctly), Also, unlike ISIS, it doesn't discard unrecognized keywords from the existing header.
Hi Massimo, Bert,
The header can be edited in ISIS provided the keyword is already there. (See attached screenshot.) There should not be any need to edit the header produced using ISIS specifically for the AAVSO database however. Can you describe what exactly the problem is ?
If I remember correctly, Massimo's main problem was that the JD-MID keyword ended up being 0 even though he had a reasonable value for the date observed. If JD-MID isn't listed, then we will just calculate it, but if it's listed then we assume the listed value is correct. However, 0 isn't a valid time for JD-MID and so it fails.
Curious. I have not heard of this problem before with ISIS. (JD-MID has been in the ISIS header since 2012). Can you recall what the value for EXPTIME was? I am wondering if ISIS perhaps failed to calculate the exptime correctly due to invalid or corrupted values from a set of sub exposures which might then knock on to the JD-MID time. If not it could be raised as a potential bug on the ARAS forum.
I don't know how it was calculated, but I can show you the header values of the file:
<bound method Header.keys of SIMPLE = T / File does conform to FITS standard
BITPIX = -32 / Number of bits per data pixel
NAXIS = 1 / Number of data axes
NAXIS1 = 4335 / Length of data axis 1
CRVAL1 = 3700.08666992188 / Coordinate at reference pixel
CDELT1 = 0.876572669 / Coordinate increment
VERSION = 'ISIS V5.9.7' / Software version
OBJNAME = 'Y Cassiopea' / Current name of the object
DATE-OBS= '2020-04-24T20:37:04.3' / Date of observation start
EXPTIME = 150 / [s] Total time of exposure
EXPTIME2= ' ' / Exposure decomposition
BSS_INST= 'Spectra1' / Instrument
BSS_SITE= 'Max1Obs ' / Observation site
OBSERVER= 'OBSCODE '
CUNIT1 = 'Angstrom' / Wavelength unit
CTYPE1 = 'Wavelength' / Axis type
Q_CAL = '0 ' / Quality flag 4-excellent
CRPIX1 = 1 / Reference pixel
BSS_VHEL= 0 / [km/s] Heliocentric speed
BSS_COSM= ' '
BSS_TELL= ' '
BSS_NORM= ' '
SPE_RPOW= 0 / Spectral resolution power
JD-OBS = 0 / JD start observation
JD-MID = 0 / JD mid observation
JD-HEL = 0 / JD heliocentric mid-obs
GEO_LONG= 0 / Obs. geographic longitude
GEO_LAT = 0 / Obs. geographic latitude
GEO_ELEV= 0 / Obs. geographic elevation >
The main issue is that JD-MID at the bottom is 0. (Not the error the system is complaining about is the greek letter gamma in the name, but that is easily corrected for. Even when you do the JD-MID is the real problem).
I see JD-OBS is set to zero too, not just JD-MID. This is very odd as ISIS should have generated that directly from DATE-OBS which it picks up from the original spectrum image. It looks like there has been a problem somewhere in the reduction using ISIS. It is unlikely to be a bug but I am not sure what has gone wrong.
There is also this recent thread on the subject which turned up some suggestions though I have not tried them
I wrote a small (and hopefully intuitive) Windows program last summer to edit FITS headers, and I've been using it successfully for almost a year. I can't upload it here, but if anyone would like a copy, just email me at nathankrumm@Live.com. I'm happy to share either the *.exe file or the C# source code. The program includes a help file with instructions on how to use it. It works for both 2-D images and 1-D spectral profiles.
Thanks for the info I attach the header of the fit generated by Isis
But the strange thing is that I have the same problem even if I process the spectrum with Demetra, since I have the Alpy600 I also tried with this software, I don't really know what to do
That header looks ok to me. Both JD-OBS and JD-MID are there as expected.
All the other values that were 0 in the header that Bert posted are now filled too so it looks like ISIS ran correctly for that spectrum. Was this spectrum also rejected by the AAVSO database ?
Thanks Robin for the help, and thanks to all of you, yes, when I entered it again it didn't go well it was rejected: "star not in the database"
I think the object name has to be recognised in VSX. Sheliak was not recognised but it is under the name Bet Lyr or HD 174638
I am not sure what source VSX uses as a cross reference for names but the BeSS database and ELODIE archive for example uses SIMBAD to cross check and so recognises Sheliak and even obscure names like Renson 48890 as Beta Lyrae
Thanks Robin, so to avoid errors on names I should check on Simbad