(Peter Shawhan, Eric Thrane, Corey Gray)

Using waveforms created by Chris Pankow, we have set up a schedule of burst hardware injections to occur once every 10000 seconds (~2.7 hours).  The schedule runs from now through the end of ER8, but we intend to turn them off once we've obtained 10 coherent (successful at both sites) burst injections.  This series is primarily intended to check the infrastructure, the low-latency analysis, and the EM follow-up alert generation machinery; we're not yet attempting to verify the calibration.  The injections are all scheduled for GPS times 1000*n+500, i.e. the first one should happen at GPS 1125350500, the next one at 1125360500, etc.  However, an injection will be automatically skipped if ANY of the following conditions is true:
   * The detector is not locked (L1:GRD-ISC_LOCK_OK==0)
   * We are not in Observing mode (L1:ODC-MASTER_CHANNEL_LATCH bit 0 is off)
   * There has been a GRB or SNEWS alert within the past hour (L1:CAL-INJ_EXTTRIG_ALERT_TIME is a GPS time less than an hour before the current time)
   * Injections are currently disabled by the operator (the Disable button on the TINJ / Transient Injection Control medm screen has been clicked, which sets L1:CAL-INJ_TINJ_ENABLE = 0)
   * Injections have been temporarily paused by the operator using the Pause button on the TINJ / Transient Injection Control medm screen (which sets L1:CAL-INJ_TINJ_PAUSE to a future GPS time when the pausing should end)
So, it's anybody's guess how long it will take to obtain 10 coherent burst injections.

To get this started at LHO, we needed to start the 'run_tinj' process on h1hwinj1, which hasn't been running since August 25.  I talked with Corey, who asked that we not do any hardware (strain) injections right now because the PEM injection folks were about to start doing tests.  However, we worked out that the injections can be disabled for now using the Disable button on the TINJ Control medm screen.  LHO had an old version of the screen, but Dave updated it to the latest version which has the Disable button (along with Enable and Pause).  Corey clicked Disable, and I double-checked that it's disabled by doing 'ezcaread H1:CAL-INJ_TINJ_ENABLE'.  I then did "nohup run_tinj &" at about 12:44 PDT.  So tinj is running now, but injections will be skipped until an operator clicks the Enable button on the TINJ Control medm screen.  Please do that when the PEM injection folks have finished their work - thanks!

There was one side effect that I didn't expect: when Corey clicked the Disable button, it took the detector out of observation mode.  Corey said that looks like it was done by sdf, i.e. this status (injections disabled) is being considered nonstandard.  I'm not sure if that is due to checking the value of L1:CAL-INJ_TINJ_ENABLE, or to some read-back bitmask value.  We should follow up with Jamie to find out, and take the injection enable/disable out of the configuration check if possible.  However, this isn't a problem right now because the PEM injection folks wanted to be out of observation mode anyway.