July 2001
J. Monnier & W. Traub

Feedback on MC/OT System
========================

From JDM:

1)  

Conceptual changes when dealing with magic constant issues:

1. internal offsets: (a: Fixed: AB , AC; b:fixed AB, BC)

2. instrumental offsets (FIXED: opd, SD1: opd, SD2:opd -- NOT
   AB,BC,AC)

3. get rid of the baseline delta vectors of the form AB,AC, BC, but
rather use position delta vectors of thef orm detla_A, delta_B delta_C
positions.  [For a while C will probably not move, so we can define
this center position as excactly 0,0,0 in position]

2) There should be a way to point the telescopes at the zenith.  This
is useful for debugging magic constant issues].

3) The CONTROL windows are nice -- with the pulldown
commands. However, one signficiant improvement would be for them to
reflect the current values of the parameters.

4) When various heartbeats are dead, it should make the system
unusable.  At least not when in some kind of debug mode.  When
taesting subsytems its really bad to require the WHOLE system be up
and running.  e.g. It seems that if one doesn't actually run the
startracker, then I can't do tests of the delay lines!

5) Is there a way to zero the star tracker offsets?  It doesn't seem
so.  Also -- it would be nice to be able to move the total offset
value into the user offset and then zero the paddle and ST offset.
This way once one paddles the telescope to the cneter of the
acqusition field, one can remember it when looking for the star on the
startracker.

6) the roll/tilt colmns are reversed on the Tracker_* windows compared
to the TELESCOPE monitor.  These should all be consistent order
ideally.

7) the spiral search refers to Tel 0,1,2. not sure what this
means. Plus its not obvious watching that the 'background search' is
working.

8) there is this problem with the roll and tilt paddle where it moves
in big jerks sometimes the first clicks in both roll/tilt.  Not sure
why.

9) We need to be able to put the delay lines in an Autocollimation
mode for testing.

10) The -- shell -- doesn't seem to work. or at least not very well.

11) Sometimes the cbcom server gets screwed up in debugging things,
and it seems the serial port gets stuck and can't get reset. We should
write a small program which can clear ports on the serial board -- and
have it invoked during a cbcom_clean.

---

From WT:

- The LD commanding from the OT needs to be implemented.

- The ST code might need some clean-up, but it seems to be working at
present so the need is not great.

- Data acquisition exists in the sense that ST and science data are
being placed in shared memory by the FPGA.

- Real-time data analysis needs to be implemented.  JDM has written
some display modules using IDL.  FPS has an interest here.  We all
agree that basic display of 6 raw or filtered fringe plots is needed,
that 6 power spectra are needed, and a cumulative waterfall display of
both is very desirable.  Headers and storage need to be addressed.

- The slow response of the OT is a big problem.  UMass sent revised OT
code with faster window-drawing after our run, and this needs to be
installed and tested.

- A picomotor controller is needed for many tasks.

- A frame-grabber would be useful for several tasks, living in the
Sun, not the VME rack, in order to get a lower-cost board.

- A yaw control program will be needed so that we can quickly move the
LD this Fall.  Components are controlling the flip-in mirror with yaw
light on it, using the ST CCD to note initial yaw light position,
commanding LD to new location, noting yaw light motion, restoring yaw
to original angle, turning off flip-in mirror and yaw light.

- The ADC board needs to have several inputs provided, perhaps via
BNCs.  Some signals to be fed to it are the SD currents and error
signals.  A display of these needs to be arranged on the OT.

- Can we operate with 2 monitors, for example 1 with the
interferometer functions displayed, and 1 with real-time data results?

- An OT programmer's manual has been started by JDM, and should be
completed under joint authorship of FPS and JDM.

- Backups of the Ultra-10 are needed.  There is a spare disk in the
machine which might be used to help here.