On Sun, Jun 17, 2012 at 8:21 PM, Eduardo Unda-Sanzana
<[log in to unmask]> wrote:
> On 17 June 2012 23:11, Tim Jenness <[log in to unmask]> wrote:
>> If you converted from FITS then the epoch header should have been
>> filled in from the DATE-OBS header originally. The DATE-OBS in the
>> FITS header is just a memory of the original FITS header rather than
>> an expression of the WCS. I assume the header in this case did not
>> come from a FITS file.
>
> Yes it did. The NDF produced from the original FITS file has the right
> DATE-OBS but the EPOCH is missing (I think it was absent from the FITS
> headers). However, when I convert back to FITS this strange DATE-OBS
> is added. I'll try to get one of the original FITS files in case this
> is useful.
>
Yes. That is wrong. FITS2NDF is meant to parse the DATE-OBS and store
it in the AST frameset. From the odd "almost 2000" date you do get
it's almost like there is a TIMESYS issue as well (maybe it's using
2000.0 TDB and converting that to a UTC for the output date-obs).
--
Tim Jenness
----
Starlink User Support list
For list configuration, including subscribing to and unsubscribing from the list, see
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A0=STARLINK
|