Hi Alessandro,
What Rob just said.
If you load up the observations for this object and shut off the 1950 data
and solve for an orbit, and then click on either 1950 observation, you'll
see that the summary of data for the observation says "dT = -33.1 minutes;
cross 0.38." (And close to that for the other observation.) Meaning that
the object passed within .38 arcseconds of that location 33.1 minutes earlier.
I notice also that the two data points are 31.5 minutes apart. (Easy
way to do this is again to click on each observation; the time is given
in HH:MM:SS form in the observation summary area.)
Ideally, you'll be able to dig up a corrected time for those observations,
and things will fit Just Fine.
This is a pretty handy thing to keep in mind. When I see an observation
with bad residuals, I'll usually check the dT and cross-residual data;
it's amazing how often I'll see that it's off by exactly an integer number
of hours, but with a sub-arcsecond cross-residual. For NEOs, it's often
off by just a few seconds, i.e., data taken by somebody who got good
astrometry, but either didn't set their clock well enough or had problems
with trailed data.
(Though why your two 1950 data points fit in the first place is a little
strange. Maybe you had the right times in the file fed to Find_Orb, but
not in the file sent to MPC?)
-- Bill
On 02/02/2013 03:12 AM, Alessandro wrote:
> I think I found a 1950 DSS precovery image of this asteroid.
> When I used Find_Orb to check the residuals of these observations, I got
> a reasonable result: they were well below 0".5, and the overall mean
> residual was 0".784
>
> So I sent the measurements to MPC and apparently all was fine, the
> observations appear in the MPC database. However, yesterday I
> realized that the 1950 observations are not used in the orbit
> calculation because their residuals are incredibly high.
> MPC calculates:
>
> 19500422 675(29.6- 10.2+)
> 19500422 675(29.6- 10.3+)
>
> I can not understand this.
> Did I make a big error when using Find_Orb?
>
> Thanks!
>
> Best wishes,
> Alessandro Odasso