Hello all,
The "in-testing" version at
http://home.gwi.net/~pluto/devel/find_orb.htm
now shows the U parameter in the orbital elements.
This is shown _only_ when one does a "full step". (It's not
particularly meaningful in other contexts). It currently replaces the
G magnitude parameter, since that is currently always 0.15 anyway.
U is shown to a precision of tenths. That's not because it is really
that accurate of an indicator; it was mostly to simplify testing. U
is not shown if the orbit is excessively inaccurate. (To be specific:
if the formal uncertainty in semimajor axis is greater than the nominal
semimajor axis, U is probably meaningless and isn't shown.)
U is also not shown for satellites or parabolic/hyperbolic orbits.
If you look in the file 'covar.txt', you'll see quite a bit of
debugging information from your last "full step". Most of this won't
be of interest to anybody but me. But the last few lines give sigmas
for various orbital elements, thus:
sigma_Tp 0.593 days
sigma_e 0.0175
sigma_q 0.00358 AU
sigma_Q 0.206 AU
sigma_1/a 0.0217 1/AU
sigma_i 0.131 deg
sigma_M 0.872 deg
sigma_omega 0.904 deg
sigma_Omega 0.435 deg
sigma_a: 0.103 AU
sigma_P: 83.7 days
P = 3.23 years; U=10.2
Similar data is stored in 'monte.txt' when one does a Monte Carlo
or statistical ranging run. The fact that the covariance uncertainties
closely match the Monte Carlo uncertainties is very reassuring to me.
(I realize that it would be extremely nice if the above sigmas were
shown in Find_Orb's on-screen element display. More about this below.
Also, of course, ephemeris uncertainties are _very_ much to be desired.)
A couple of other small changes:
-- The "Make Ephemeris" dialog now has a "Now" button, to address
the issue mentioned by Andrew Lowe. There's also a check-box to specify
that the ephemeris start should be rounded to the nearest step size. (By
the way, Andrew, I tried the fixed-size font, and it made the elements
too big to fit. It's possible this can be fixed eventually by rearranging
the main Find_Orb window.)
-- If the elements are deemed to be truly wacky (heliocentric e=2,
for example), they are shown in red. (In the console version of Find_Orb,
they are shown in _blinking_ red. I may try to do that in Windows
Find_Orb at some point. People new to orbit determination really need
need something to warn them when elements are bad.)
-- A new Find_Orb user ran into some problems with the Vaisala function.
To use this, you have to enter the perihelion/aphelion distance in the R1
field, which is far from obvious. I fixed this: click on "Vaisala", and
you are prompted to enter the peri/aphelion (apohelion?) distance.
The new uncertainty data raises some questions about on-screen display.
As far as I know, nobody really makes any use of the P and Q vectors shown
by Find_Orb. So far, I've slavishly followed the MPC's eight-line format;
it looks as if it's time to junk that, in favor of something like...
Orbital elements:
2002 RG23
Perihelion 2002 May 15.050409 TT = 1:12:35 +/- 1.18 days
Epoch 2002 Sep 6.0 TT = JDT 2452523.5 Earth MOID: 0.8646
M 30.950381 +/- .215
n 0.271614677 +/- .00041 Peri. 319.339168 +/- .313
a 2.361387318 +/- .0054 Node 338.982785 +/- .0311
e 0.20417200 +/- .00038 Incl. 3.186808 +/- .026
P 3.63 +/- .01 H 18.4 U 8.1 q 1.879258137 Q 2.843516499
From 19 observations 2002 Aug. 28-Sept. 14; mean residual 0".544.
I'd like to also cram in the sigmas for q and Q, but you get the idea.
Any thoughts?
-- Bill