NavList:

Michael, congrats on getting your Mk IX running.

I have thousands of shots on various Kollsman bubble sextants using a tripod and printed base.  I find these to be very accurate with LOPs missing by 1nm or less most of the time but every so often I will get a poor LOP.  Most of the time, a poor LOP comes out of a dim star being shot thru the murky air or while under attack by critters (not sure if the skeeters or the coyotes are the bigger distraction), but bubble technique matters if you are chasing accuracy. 

I try to keep the bubble "medium sized."  Too small and the bubble might 'stick' and not get to the top of the chamber.  Too large and, well, it's hard to find the center.

When I setup the sextant, I will jiggle the tripod a bit to settle the bubble.

I let the sextant adjust up or down to ambient outdoor temperature for at least 15 minutes before I take the first shot.  This lets the bubble chamber settle and keeps the bubble from adjusting itself larger or smaller all on its own.

I run the averager for the full 2 minutes.  As I am stationary, I only account for MOB (Motion of the Body) and not MOO (Motion of the Observer).  I check my final altitude against the averager output - final altitude +/- the MOB should be pretty close to the averager output for a 2 minute sight. I use this as a quality test of the shot - if the two do not match, I will take the arithmetic mean of the two and use that in the reduction.

You can find MOB (and MOO) tables in the back of Pub249v1 as Table 2 (and Table 1) - or use a Polhemus computer which is a lot faster.

I also run the index knob in one direction if at all possible as I find that some old mechanical systems work a little differently if the knob goes up vs the knob goes down. I have different ICs for rising bodies vs falling bodies for a few individual sextants.

These are the things I do for max accuracy - not really applicable to "real world" shots.  

Just thought I'd share.

Art