Hi Detlef,
Stefan beat me to it: you need to make sure the 'heliocentric only' check-box in
Settings is unchecked. And when you load up 2008 TC3, not all observations are
initially turned on. This happens rather often with longer arcs (2008 TC3 was
observed for less than a day, but the arc on the sky was relatively large).
In such a case, Find_Orb selects a sub-arc and does an initial orbit determination
for it.
Also, I see the documentation on the Web site mentions getting data for 2008 TC3
from MPCOBS, but doesn't mention that one can get it from NEODyS at
http://newton.dm.unipi.it/~neodys2/mpcobs/2008TC3.rwo
I'll fix the docs to mention the NEODyS link and the need to clear the 'heliocentric
only' box.
In re not all observations being turned on: in recent weeks, I've been thinking
about how Find_Orb might do arc extension (figuring out that adjacent observations
could be included, leading to still more observations being fitted). "Auto-Solve"
can sometimes do this, but it's a little bit limited. I've had some ideas for
improvement. It is possible the next update will be able, in about 99% of cases,
to load up data and completely figure out the orbit without human intervention.
(The 'problem cases' being certain artificial satellite orbits and SOHO comets;
these will probably still require some human thought.)
-- Bill