After Peter and I powered off the DBB, I ran PeterF's script to see whistles.
Probably because IFO noise was good, I was able to clearly see whistles in DCPD that correspond to the difference betwee IMC VCO and DIFF.
In the plots attached, white circles are the difference of the fundamental frequency of IMC VCO and diff VCO, yellow triangles are second harmonics.
There are visually apparent whistles in DCPD (leftmost top, 0-4 sec, 23-24 sec between 4000 and 6000Hz) and they agree pretty well with the fundamental of IMC-DIFF. There are less apparent whistles (12-13 sec between 2000 and 4000Hz, 15-16 sec between 2000 and 4000Hz) and they also agree with the fundamental of IMC-DIFF.
All visually apparent whistles in OMC PI channel (there are four such whistle traces) seem to agree well with fundamental and second harmonics of IMC-DIFF and IMC-COMM.
I went to the floor and started powering off the DIFF VCO, it was a paint to disconnect power connector from the back, and IFO lost lock (ASC instability?) about 80 second before I successfully removed both 17V and 24V connector.
It's not clear if turning DBB off did anything as far as whistles go. At first I thought whistle traces got simpler, but if you look at the plot before DBB was turned off, the number of apparent whistles in OMC PI channel was four, so no change there.
And attached is when it was really ugly.
https://alog.ligo-wa.caltech.edu/aLOG/index.php?callRep=47188
At that time IMC VCO was at a different frequency and I couldn't find anything that comes close to this. That one is still a mystery. (That's why I initially got suspicious about DBB locking oscillator.)
Why? There are 2 switches on the front panel to power off the RF sections...
Oops, will use that feature next time!
A note on the 15-20kHz feature (first figure) in OMCPI channel. This was not there in the past ( 2017 March second figure), and is not present at Livingston (third figure).
