NavList:

   

Reply

> Of course. Please do. A positive bias like that suggests a simple error in
> index correction. How do you normally test index error? I don't recall
> --what type of sextant do you use?

I'm using an Astra IIIB, purchased about ten years ago from
Celestaire, along with a porro-prism monocular from them, which is
marked "Cassens & Plath 6x30".  I recently discovered that it was out
of collimation by about 2 degrees relative to the sextant plane with
no visible damage or problems with the fork or attachment points, so
I've shimmed it pending a more permanent repair.  I'm kind of evolving
on index error measurement; there are many good suggestions in the
archives.  Unfortunately it's not a constant 0.5' bias; from session
to session it moves around from close to zero to close to 1', but
fairly tight within the session.  My current thoughts are:

- I don't like IE measurement with the horizon or single stars.  With
the horizon (a treeline about 10 km away, so sextant parallax is a
negligible 0.02 arcmin), contrast is poor because of the superimposed
images, and I'm concerned about possible bias when approaching
coincidence always with increasing angle, as recommended to avoid
backlash.  As for single stars, it's even worse:  the horns of the
dilemma are, on the one hand, with no side error the superimposed
images are aberrated mush even with good focus, and with some side
error, judging when the stars are closest is difficult with no wire or
reticle telling me when the images are in a line normal to the sextant
plane.

- IE measurement with the solar or lunar disk is better, but this
still gives information only near zero degrees on the arc, and I'm
starting to worry about arc errors and micrometer-worm-eccentricity
errors.

- So, recently I've been measuring IE with star-star sights, three or
four spread across the arc from 10 degrees to maybe 90 degrees.  For
me, the swinging motion of the star in the index-mirror path really
helps.  Just having some motion when the stars collide makes it seem
like I have higher precision---the relative paths are trackable by
eye.  I suppose I'd want the distances uniformly spread modulo 1
degree to try to average out worm eccentricity, but I haven't done the
planning for that yet.

> It's good you're using a 6x30 scope.

I'm coming to the conclusion that I really need a reticle.  I think
some of my remaining error is still collimation, because with a
6-degree FOV in the scope, it's hard to touch exactly in the center
without a guide.  Chauvenet mentions 7 arcmin as a maximum angle
between touch and collimation axis!  That's only 2% of my FOV.  Even
if I loosen that to 15 arcmin, it's still challenging with no wires.

> With that magnification, there's every reason to expect you'll be able to get
> your mean error down to about a third of a minute of arc on individual
> lunars, with no bias.

I'd be happy with that.  I'll try to post my plots later today; random
errors seem to be about that large, but all that's showing is my
tangency repeatability.  The systematics are the hard part.  Of
course, at some point one has to just stop, given the diminishing
returns, or else get a better instrument (repeating circle perhaps).

Cheers,
Peter

   

Reply