(Duo, Nelson, Orion, Vladimir, Keith)

This is old news (going back to ER6 and ER7), but with lower-frequency lines improved since O1,
we are looking harder at the disturbances seen in both H1 and L1 DARM very near 100 Hz.
In both interferometers one sees lines / combs extremely near 100 Hz and several visible harmonics:

Computing average spectra of 30-minute Hann-windowned FScan SFTs from about the first two months of O2 data,
I see the following lines / combs:

H1: 

99.9762 Hz (visible to 2nd harmonic)
99.9989 Hz (visible to 5th harmonic)

L1:

99.9728 Hz (visible to 2nd harmonic)
99.9755 Hz (visible to 2nd harmonic)
99.9780 Hz (visible to 2nd harmonic) 
100.0000 Hz (visible to 2nd harmonic)

Having lines show up in both H1 and L1 so near 99.98 Hz has led to
spurious outliers in O1 CW all-sky searches and will likely do so again
for O2 searches. 

NoEMi shows correlations between DARM and end-station magnetometers.
Duo Tao followed up these clues for O1 H1 and L1 and more recently for O2 H1 and L1
with week-by-week coherences between DARM and auxiliary channels.
The lines near 99.98 Hz are seen in many magnetometer, suspension (and sometimes even accelerometer
channels) at both end stations at both LHO and LLO.

Given the prevalence of these lines in so many different types of channels, 
it's natural to suspect electronic coupling. Are there known power supply combs 
with ~100 Hz spacing? I see this LLO alog entry from Anamaria, which mentions
some 100-Hz detective work (and documents a dramatic mitigation of a 2.24-Hz comb).
Is there a better understanding of the 100-Hz artifact(s)? At either observatory?

Figugre -- Zoom of DARM for 100-Hz region, showing arithmetic and inverse-noise-weighted averaged spectra from H1 and L1 for December and January together