Re: GSC 3493-1324

[Stupendous Man <richmond@stupendous.cis.rit.edu>]

  Sebastian writes (concerning TASS measurements of GSC 3493-1324)

> Tom, here is my concern regarding false "eclipses" one stumbles upon in TASS
> data. I can throw datapoints away from the other databases looking at the
> flags or rejecting discrepant points made in pairs just  seconds apart. But
> with TASS data I don't have any indication in the file I download. There are
> three points at JD 2453109 that are wrong and I would have used them as a
> true eclipse if only analysing TASS data. They seem consistent with each
> other but they are not.
> Isn't there a way to flag these observations?

  I've looked at the measurements in question.  On the night of
JD 2453109, there are three measurement pairs in V and I.  
Each pair has a magnitude fainter than normal in V and and I.
The size of this dimming is about 0.3 mag in V and about 0.3 mag
in I -- which is roughly the same size as the dimming which
occurs during "non-false" eclipses.

  I'm away from my office at the moment, so I can't track
down exactly where the telescopes were pointing at these
times, so I don't know exactly where the star happened to fall
on the chips.  It's possible that each of the two chips
had an undetected defect in just the right spot for the star
to strike on this particular night ... but that doesn't seem
very likely.

  I looked at the sample comparison star Michael mentioned, 
GSC 3493-1088, on the same night (and other nights).  It does
NOT become dimmer on the night of JD 2453109, in either V or I.

  It's certainly possible that Sebastian is right, and these
TASS measurements on this particular night are wrong, while
other TASS measurements on other nights correctly appear dimmer
during real eclipses.  However, at the moment, I can't see 
an obvious reason why these particular TASS measurements 
must be wrong; so I can't see how we could flag the measurements
in the database as suspect.  

  If I were back in my office (instead of being on the road),
I'd try to track down the NSVS data Sebastian mentioned to see 
if there is some way that one could include these particular
TASS measurements and get a different solution to the binary
system parameters ...

                                          Michael