Reports until 14:18, Wednesday 12 June 2013
H1 SEI
hugh.radkins@LIGO.ORG - posted 14:18, Wednesday 12 June 2013 - last comment - 11:08, Thursday 13 June 2013(6715)
HEPI LVEA Pump Stations found off-Problem understood mitigated 
I was up on the Mezzinine measuring for cables and noticed that the pumps were not turning.  No level trip, no VFD trip, servo running pushing max volts.  Richard reminded me of a problem that we should have corrected some time ago.  The output from the servo has 4 BNCs to the four pump stations.  There was a bad BNC that has shorted in the past and it did so again.  Once it shorts, no matter what the servo does, the outputs are shorted and the VFD gets zero volts hence no revs.  As soon as I unplugged the bad BNC they rev'd back up.  We took it back down and Filiberto replaced the BNC.  This should not happen again.  I may have bumped the cable and caused the short but I'm pretty sure it happened 10s of minutes before I was in the area.

The pumps went down at ~1818utc and the BSC2 HEPI moved a fair bit, several .001".  After the BNC repair I got the Pump Output back to operating pressure ~2040utc with the HEPI Actuators reaching equilibrium position a few minutes later.  The difference in before & after position at BSC2 was ~0 to ~600 counts, all less than 0.001".  Obviously a need for alarms--working on it!
Comments related to this report
hugh.radkins@LIGO.ORG - 11:08, Thursday 13 June 2013 (6736)
Link 

Attached is a plot during the pressure loss event.  The lower left graph shows the pressure drop; the other graphs are the shifts seen of HEPI at BSC2.  The sloped section in the second quarter is a DAQ data loss period.  About halfway thru I found the problem and disconnecting the shorting BNC allowed the pumps to restart and pressure & positions to recover.  The third quarter is while Filiberto is repairing the cable.  Finally all back on and settling back into position.

The main thing here is the huge shift of the positions during the pressure loss.  A few hundred counts or even a couple mils (1310 cts) could be tolerated. But here the best is v3 (Ch 8) with a shift of just 1200 counts.  Most shift are quite a bit more and H4 shifts more than 7000counts almost 11 mils more than 25% of the range.  This suggest to me a problem with the Actuator.  Either the Large Pin Valve or the Parker Valve is miss positioned or calibrated.  It seems we should be able to tune these in situ to reduce/minimize this affect but I'm not sure I heard about it.

Hopefully, HEPI Pressure losses will be far and few between but it would be nice if we really weren't badly impacted.
Images attached to this comment