UCAC-3 vs USNO-B1 is a bit of a puzzle. I'm running Astrometrica for my astrometry, and average reference star positions are 3 times closer when using UCAC-3. While I'm certainly not an expert, this suggests that UCAC-3 better matches my camera/scope setup, and that the plate solve is just a wee bit more accurate.

Clear skies,

Pete Peterson
Wishing Star Observatory I15

--- In guide-user@yahoogroups.com, Bill J Gray <pluto@...> wrote:
>
> Hi Pete,
>
> > UCAC-3 seems to be the optimum catalog for astronomical plate solves...
>
> Be warned that there is some controversy over this. UCAC-3 has some
> very nice properties; it's a big leap forward over GSC, its magnitudes
> are decent, and it doesn't have as many "fake stars" (or missing stars)
> as other catalogs. This is, in part, because it's based on CCD
> imaging, and matched to 2MASS. So you don't get the plate scratches
> and other artifacts common to GSC or Ax.0 or B1.0.
>
> Also, the match to 2MASS means there's some passable color data
> available. And if a star doesn't have a 2MASS match, it's probably
> not "real". (Guide 9 will omit the 1.5% of UCAC-3 stars that lack a
> 2MASS match; almost all of them are bogus.)
>
> So what's the down side? The positions and proper motions have some
> errors, almost entirely negligible for most purposes. For a printed
> or on-screen chart, for example, the errors will usually be about
> .001 pixel or less; nobody would ever know. But for plate solving,
> you really need to pick nits. Something appears to have resulted in
> anomalously high proper motions for some stars. (The usual cases I've
> seen involve close pairs of stars; in UCAC-3, it looks as if they're
> zooming apart from one another.) There's also concern about a systematic
> offset of about 20 milliarcseconds.
>
> There were some other concerns about UCAC-3. For example, there are
> some missing brighter stars. The really bright ones get picked up in
> Tycho-2 anyway, but some around mag 11 or 12 or fainter are just plain
> missing. There are also some spurious stars and false doubles. Good
> news there is that they can be immediately spotted by the fact that
> they weren't matched to 2MASS; drop them, and the problem goes away.
>
> You would never notice this under most circumstances, which is why
> I'm reasonably content with UCAC-3 as the base catalog for Guide 9.
> But for plate solving, if you're really trying to get the sort of
> precision required for asteroid astrometry and such, you may be better
> off with B1.0. (Especially if you're going down to mag 20; at that
> point, I'd expect a lot of UCAC-3 stars to be saturated.)
>
> Also, be aware that the above should not be taken as gospel. There
> was some rather heated debate about all this on the Minor Planet Mailing
> List a while back. Nothing conclusive, but one does need to be aware
> that UCAC-3 has strengths and weaknesses (just like any other star
> catalog.)
>
> > ...The star display still allows me to turn on the SAO numbers on,
> > but they don't display any longer. Right clicking only provides the UCAC-3 info.
>
> There are some odd things there, because UCAC-3 isn't linked to the
> other catalogs. In Guide 9, I'll be able to know which UCAC-3 stars
> appear in Tycho-2; that lets Guide show bright stars from Tycho-2 (taking
> advantage of the fact that Tycho-2 has SAO and PPM and other numbers),
> and shows fainter stars from UCAC-3. But that requires the specially
> processed dataset that's on the Guide 9 disk. At present, if you show
> UCAC-3 stars, then the SAO and PPM numbers will be suppressed.
>
> (At least, they will be until you zoom out a bit and UCAC-3 isn't
> shown any more. At that point, they'll snap back into view, and you'll
> also see them when you right-click on a star.)
>
> -- Bill
>