Sarah,
> #35 Err: !! No positions were matched between any dataset
> #35 Err: ! FINDOFF: Error determining position list matches
> #35 Err: !! SAI__ERROR: Error
Although these are error messages, the recipes expect this, say if you
have too few common stars to match and register the individual frames.
Then it has to fall back on telescope offsets. I don't think you can
readily suppress the messages.
> So although orac mosaics the data, the errors are telling me that I can't
> really trust the final mosaic as the offsets are not correct. This occurs
> with my data (JITTER_SELF_FLAT) and my standard (BRIGHT_POINT_SOURCE_APHOT)
The telescope offsets do not provide as good a registration as the stars
used as fiducial marks. There used to be about 0.5 arcsec uncertainty
(I don't know how good it is now). If it's important for your research
to obtain the best registration you could use CCDPACK's interactive tool
(PAIRNDF) to mark corresponding sources to dettermine the offsets
between frames, and then use REGISTER to apply the offsets, and
finally MAKEMOS to create the mosaic. Details are in SUN/139.
> Is this usual? If not, is there anyway I can persuade orac to mosaic the
Yes it is especially nearer the celestial poles and in short
integrations. I did try to use the world co-ordinates to enable
matching of fewer than three objects. As far as the pipeline is
concerned, it's main goal is to provide reasonable, if not
publication-quality, mosaics to present to observers in near real time,
so they can assess the data at the telescope.
> stars in my field of view I can't really mosaic the files any other way -
> most interactive software requires at least 3 field stars in the image -
> something I don't always have.
Exactly. That's what FINDOFF is complaining about. Have a play with
the _ff files in PAIRNDF to see if you can do significantly better
(comparing the PSFs of the automated and hand crafted mosaics).
Malcolm
|