Rough outline of investigation into E2 lock losses     10/24/00
      --------------------------------------------------

    (Participants: David Chin, Dick Gustafson, Keith Riles)

Pre-run:

  Ensure that channels likely to be relevant to understanding
  lock losses are present. Perhaps will need a handful of new
  channels (e.g., DC output of 1-2 OSEM's to check for photodiode
  saturation). We need to do some homework and testing.

During run:

  Keep our eyes and ears open! 

  Run a dmt monitor to record all lock losses. Follow up
  with on-the-fly investigations of suspect channels with
  data viewer (playback mode) and diagnostic test tool.
 
  If testable hypotheses formed from initial investigations,
  try testing them, perhaps during a special shift, depending
  on whether or not test is invasive (e.g., raising laser power
  to induce lock loss).

After run:

  See where we stand and address losses of unknown origin:

  * More investigations with data viewer and test tool

  * Check correlations with other DMT monitors, including the
    TID program.

  * Using record of lock losses, use the DMT to sift through time 
    intervals well before and immediately preceding losses, scanning
    systematically through all conceivably relevant channels, looking
    for systematic differences (e.g., larger rms, drifting mean,
    spectral distortion)

Data samples/channels needed:

  We will want access to most of the data taken during that week
  and would prefer to do analysis on fortress (or other LHO machine).
  Given the open-ended nature of the investigation, we will want
  the most comprehensive reduced data set available. However,
  once we have recorded all lock losses, we will look at only a
  modest number (we hope!) of very short stretches of time.

  The baseline reduced data set channel list provided by the
  Oregon group is a good starting point. For our purposes, all
  of those 16kHz channels could be decimated down to 1 or 2 kHz.

  We would like to add to that list, however, the following:
  (with suitable translation of names):

  Additional LSC signals:
  (would like at 1-2 kHz, these are nominally 16 kHz floats)

   H2:LSC-MC_AO
   H2:LSC-MC_L
   H2:LSC-DARM
   H2:LSC-CARM
   H2:LSC-MICH
   H2:LSC-REC

  Optical levers: (sensors & control signals)
  (would like at 256 Hz, these are nominally 2 kHz int16's)

   H2:ASC-QPDX_Y
   H2:ASC-QPDX_P
   H2:ASC-QPDY_Y
   H2:ASC-QPDY_P
   H2:ASC-ETMX_Y
   H2:ASC-ETMX_P
   H2:ASC-ETMY_Y
   H2:ASC-ETMY_P
   H2:ASC-RM_Y
   H2:ASC-RM_P
   H2:ASC-BS_Y
   H2:ASC-BS_P
   H2:ASC-IB_Y
   H2:ASC-IB_P

  If wave front sensing is running on some cavities,
  we may wish to have some of those signals too, again at 256 Hz.