During opening, we again saw the secondary E axis Galil fail to stop on a full 
step, despite further work on the Galil by Jon B. and Wendell this afternoon,
after which Steph and Russell exercised the secondary without incident.  The first
time we homed the axis, tcc_prim reported a 148 microstep error.  We immediately
saw the same problem as soon as we re-enabled the collimate process again.  This
time, homing showed a 1 microstep error.  We did not see the problem again during the
night.

In pursuit of the source of the difference between the two pointing models taken
this run, Robert and Steph verified that the same fiducial tables were used, and
Craig and Russell eliminated axisDtime errors and problems with the earth 
orientation data block as possible causes.  The mystery continues.

We opened at 20:30 under partly cloudy skies.  After fixing the problem with the
secondary, we attempted spectroscopy, starting with the pointing model of last
night.  

Plate 627, cartridge 1.  No stars in guide fibers after initial gotoField.  The first
fk5InFiber produced bizarre behavior.  The field was at (RA, Dec) = (313, -6).  The
track to the fk5 aimed at (248, 47)!!!  This occured between getclock times 997763607,
and 8997763798.  Another fk5InFiber sent the telescope to a star a few degrees from 
the field.  However, we could not find a star within a 30 arcsec radius of a
spiralPattern search.  We tried reverting to the previous pointing model without
success.  We decided to try a different cartridge.

Plate 636, cartridge 3.  With the latest (20-point) pointing model, we found an fk5
star at (deltaRA, deltaDec) = (12, 2) from the aimed position, and quickly got on the
field.  Then Jurek's sop, sopGUI, telrun and fsaoImage all locked up, just as we were 
about to close because of rapidly rising humidity.  We had to kill those windows.  This
may have been due to hitting the doNextScience button in sopGUI to start the second
exposure too soon.  SoS failed to record either the first or (aborted) second exposures,
so we acquired no data on this plate.

Spectro 1 fill faults reported by watcher at 22:19 did not apparently correspond to 
actual warming.  This is a know fault. 

We closed because of high humdity at 00:05 MDT, followed by dust levels above limits
after 01:00.

The imager rebooted itself while we were outside closing, causing us much confusion
on our return about what all the errors meant.  This is put down to the test of the
panic button during shakedown last week.  Many thanks to Connie for stetting us
straight again at 3am her time.