From: "Pete Newman"
Subject: RE: 2.5m obslog 20041029 (MJD 53308)
Date: Sat, 30 Oct 2004 09:02:02 -0600
Mike--
> Lots of fiducials too large to correct tonight.  Most of them were
> in altitude
That's to be expected, given the adjustments to the altitude fiducial blocks
and their read head this bright time.  These have thus moved relative to
their positions in the current MCP altitude fiducial table.
As long as the blocks and head haven't moved since the sweeps, I think
there's only one annoying side effect until the tables get built and the
pointing model gets done: you might not get to instrument change position by
There's at least on side effect you may run into until you build new tables:
trouble inserting the altitude locking pin.
Try slewing to alt=90 with the MCP and see if the pin is well centered.  If
so, you're ok.
If the pin is NOT well centered, then it's easy enough to adjust.  Move the
telescope as close to true zenith as you can by tweaking the position in MCP
until the altitude locking pin is well centered in its hole.  Ignore where
MCP thinks it is at this stage (may be either side of nominal 90deg) but
instead tell MCP it is at alt=90 by using the 'p' command in mcpMenu.  Any
residual mean error will get fixed in the pointing model.
I think you need to be sure the MCP is working with the correct position
before rebooting the MCP as the new tables are implemented, since the angular
position of the fiducials are NOT recorded in the built tables, only their
mean encoder positions.
Except even worse fiducial crossing errors until you rebuild the tables, and
beware that if you later accept a fiducial crossing error with the old
tables, you'll lose the true position again.
Cheers,
--Pete
p.s. I _hope_ I got this procedure right - it's sometimes confusing trying to
understand how the telescope knows where it is at all, given the number of
subsystems involved!  If it IS right, I'll folklore it.