The problem the telescope has been having (probably only since last
Thursday) has been that the pointing has a "memory".  That is where
the telescope points depends upon where it was pointed previously.
The pointing error could have been as large as 50 arc seconds in
right ascension and 16 seconds in dec for a long slew across the
zenith.  Adjacent 2mass scans would have had the correct scan to
scan RA change because a much larger ha motion would be required
to work out the "memory".

Yesterday Ed Hileman and I checked every bolted joint in the telescope
for motion and every bearing surface not measured by an encoder for
friction and slippage.  We found the following things:

1. Last March a truss was added to the telescope to strengthen
the fork when the telescope was pointed far over in hour angle.
Because the width of this truss is small the shear load where the
truss is bolted to the fork near the tail bearing and near the
declination bearing is very high.  We found evidence that the
bolted joint near the west declination bearing was moving.  Ed
tightend the bolts but we were not able to quatitatively measure
how much improvement we made since we could still detect slippage.
We have decided to weld the joints where the truss is now bolted
to the fork.  We will do this in the next day or so.  

2. We took one of the declination bearings that is not installed on
the telescope and measured the force required to move the
spherical joint which allows the bearing axis to align itself.
We found that a large force was required to overcome the friction
in the spherical joint.  An experiment on the telescope showed
that the declination bearing was not distressed by freezing the
spherical joint in the telescope zenith position.  This joint
is now loc-tited and shims have been removed between the halves
of the pillow block to increase the clamping load on the spherical
joint.

3.  We found that the "steering" adjustment of the new right
ascension axis rollers upon which the entire weight of the fork
and tube rests was far out of adjustment.  This caused a large
force to build up as the telescope was moved in hour angle
(probaly a ton or so) until the roller slipped on the horseshoe
track.  This force distorted the telescope structure appreciably.
These rollers were installed last Friday.  The adjustment Ed
made yesterday brought the force down by a factor of between 
ten and twenty.

4.  We remeasured the previously known end play in the declination
bearings.  The end play is between .003" and .006".  The measurements
are somewhat inconsistent.  Since this motion is a translation along the
declination axis it should not to first order affect the pointing
of the telescope.  However, because the bearing race is at an angle
the motion causes some tipping of the declination axle in the 
hour angle direction.  We have improved bearings with reduced end play
on Mt. Hopkins for the telescope.  The same improved bearings are on the
boat to South America for the CTIO telescope.  M3 is manufacturing a stand
that will allow us to change the Mt. Hopkins telescope bearings quickly.
We have not made a decision to change the bearings.

Last night Eric measured the pointing memory caused by a slew from
-1.5 hours to +1.5 hours and back to the star at -1.5 hours.  The
telescope returned to the star with a RA error of -1.7 arc seconds
and with a declination error of +1.4 arc seconds.  So the telescope
pointing is quite healthy again.

Rae Stiening
October 1, 1997