Reports until 15:28, Monday 28 April 2014
H1 AOS (SUS)
richard.mccarthy@LIGO.ORG - posted 15:28, Monday 28 April 2014 - last comment - 11:21, Tuesday 29 April 2014(11609)
ETM-X ESD Problems with Power Supplies
The ESD was non-functional this morning when we tried to operate it.  I went down to EX and found the negative supply in the off state (indicating a trip probably not someone shutting it off).  Turned it on went out to ESD drive to turn it on.  As soon as I turned on ESD amp the Neg. Power supply tripped off.  Went back out to power supply to see if I had not set the over current.  The unit was off again. This time it would not turn on.  Took this opportunity to swap out bot the positive and negative supplies.  The positive due to problems noted last week with sync. and the neg. do to this failure.  Power the system back up.  The system once again was in an odd state all of the supplies present but sitting at a neg rail.  I verified voltage everything was properly connected but still at a neg. rail.  This was the same state the unit was in on Friday when I had Kiwamu come out to help.  I disconnected the DAC inputs and reconnected them.  Everything came up just fine.  This is a very troubling state as the DAC was not sending out a voltage and the unit was behaving as though it was.

The unit is now running and seems to be working again.
Comments related to this report
kiwamu.izumi@LIGO.ORG - 11:21, Tuesday 29 April 2014 (11623)ISC

This is just a followup of the ETMX ESD power supply failure.

It seems that the power supply became non-functional (or at least the HV amp started generating zero voltage) in the evening of this past Friday which is the day we activated the ESD for the first time on ETMX.

  • Apr.25 18:00 UTC  - We started using the ESD with a bias count of -125 k counts.
  • Apr.25 18:45 UTC  - The swept sine measurements were over. The DC bias was left on at -125 k counts.
  • Apr.26 00:29 UTC  - The readback ADC started reading zero voltage while the DAC still kept sending -125 counts. The readback remained zero until Richard did some activities yesterday.

The DC bias was the only channel which had a non-zero value and all the other quadrants was not doing anything at the time of the failure. A trend of the DC bias and its readback is attached.

Images attached to this comment