Reports until 13:13, Tuesday 15 April 2014
H1 SEI (ISC, SEI)
hugo.paris@LIGO.ORG - posted 13:13, Tuesday 15 April 2014 - last comment - 14:43, Tuesday 15 April 2014(11351)
HEPI Update allows 60s worth of saturations on L4Cs
 
LHO commissioners recently complained of having to recover from too many HEPI-L4C trips. 
 
HEPI master model and MEDM screens were updated to allow 60s worth of saturations on the L4C signals before tripping HEPI. The script used by commissioners to temporarily disable HEPI-L4C WD was turned off.
 
 
The Saturation Counters Work as Follow:
   - There is one saturation counter per sensor group (i.e. L4C, IPS, ACT, ...)
   - Counters get incremented every time a sensor of the group overpasses its WD safe threshold
   - The HEPI overview screen border turns orange when any of the platform's saturation counter is superior to 0. 
   - The platform trips whenever one of its saturation counters reaches the maximum amount of saturations allowed.
   - All channels, but HEPI L4Cs, are allowed 10 saturations.
   - HEPI L4Cs were temporarily allowed to take 60s worth of saturations (122880) before tripping HEPI. 
   - One does not need to wait for a WD trip to reset the saturation counters. Saturation counters can be reset anytime by pushing the WD reset button. 
 
 
Update Details:
   - HEPI master model and MEDM screens were updated
   - Every HEPI model as recompiled.
   - Every chamber was turned off (~11:30am)
   - HEPI safe.snap files were updated, and commited to the SVN.
   - HAM4-HEPI was restarted.
   - Hugh manually shook HAM4-HEPI so we could see the L4C saturation counter going up, until it tripped HEPI.
   - Every HEPI model was restarted.
   - HEPIs were turned back on (~12pm).
 
Work was performed under WP 4567, which is now closed.
 
Images attached to this report
Comments related to this report
jameson.rollins@LIGO.ORG - 13:36, Tuesday 15 April 2014 (11352)
Link

To be absolutely clear, my understanding is that the new threshold is NOT actually 60s worth of saturations, but

60 * (model sample rate in Hz) samples worth of single-sample saturations that will now cause trips.

This is a big and important distinction.  The first will trip after 60s worth of continuous saturations; the second will trip in an indeterminate time after the first saturation after a reset.

Please correct me if this is not right but this is my understanding of the new behavior, and it's important that everyone is clear on it.

hugo.paris@LIGO.ORG - 14:43, Tuesday 15 April 2014 (11362)
Link

60 x (model rate) = (number of saturated samples allowed) is what I coded into the master model, and what I meant by "60s worth of saturations" without realizing it could be misunderstood.

HEPI-L4C saturation counter will keep accumulating regardless of whether the saturations are continuous or not, until the reset button is pushed, or HEPI trips.

Images attached to this comment