Camilla C, Sheila D, Jenne D, Jeff K

While investigating unknown locklosses, we have noticed spikes in DARM <1 second  before we loose lock.

 

This has happened in 5 out of 24 unknown slower locklosses in Feb ( 20% ). Similar spikes were seen in the 2 locklosses where the DARM offset was changed (alog 55204).

We are suspicions that the OMC-DCPD's are saturating. I have attached images of the lockloss with the OMC-DCPD OUTPUT channels, the input channels are slow so don't currently have good data but it is important to note that the filters applied reduce these ~100Hz features by a factor of 10³!

Jenne captured the features on fast input channels from the higher DARM offset lockloss very close to saturating the DCPCs (saturates at value of+/-32768) although these features look small on the output channels (photo of OUTPUT CHANNELS and INPUT CHANNELS ). The other lockloss with higher DARM offset doesn't look as bad (photo).

 

There is lots of examples of other locklosses with spikes before that I have attached:  here, here, here.

The bigger (~200,000 in ETMX_L3) spikes seem to occur without us loosing lock around 5-9 times a day. Not very often compared to how many times we loose lock just after seeing them!