Summary: There appear to be periodic transients in h(t) that are similar to (but nonetheless different from) those seen in the past from the infamous blinking LEDs in the timing electronics. From looking at magnetometer channels, I think the strongest (only?) coupling sources are near or in the EX_EBAY_SUSRACK and/or EY_EBAY_SUSRACK electronics.

More info:

It has taken a while, but the old Michigan data-folding infrastructure is working again (thanks to undergraduate Reilly Penhorwood). 

Figure 1 shows the May-August H1 DELTAL_EXTERNAL folded upon itself in 8-second folds (observing mode only). The top plot is the raw average over those folds, the middle left plot is the full spectrum of the top fold, the middle right plot is the "band-passed" zoom of the left plot, and the bottom plot is the inverse Fourier transform of the band-passed FFT coefficients. One sees what appears to be a combination of 1-second and 2-second periodicity.

Figures 2-3 show the corresponding plots for the EX magnetometer channels (no restriction on observing mode) EX_MAG_EBAY_SUSRACK_Y_DQ and EX_MAG_EBAY_SUSRACK_Z_DQ (the corresponding plots for the X direction are not yet available).  

Figures 4-6 show corresponding plots for the EY channels (x, y and z directions):  EY_MAG_EBAY_SUSRACK_X_DQ, EY_MAG_EBAY_SUSRACK_Y_DQ and EY_MAG_EBAY_SUSRACK_Z_DQ 

One sees in all of these figures evidence of transients close to integer seconds and about 0.15 second later, with in many cases (including in h(t)) a slight difference between even and odd integer seconds (but not nearly as pronounced as what we saw in O1 from the original blinking LEDs). 

Reilly has a web page with many more plots for individual months/days and for other magnetometer channels (and for corresponding L1 channels) here. Other magnetometer channels also show disturbing transients, but not with the same time structure as seen in h(t).

The direct links for drilling down to month/day for the six channels above are 

H1:CAL-DELTAL_EXTERNAL_DQ

H1:PEM-EX_MAG_EBAY_SUSRACK_Y_DQ

H1:PEM-EX_MAG_EBAY_SUSRACK_Z_DQ

H1:PEM-EY_MAG_EBAY_SUSRACK_X_DQ

H1:PEM-EY_MAG_EBAY_SUSRACK_Y_DQ

H1:PEM-EY_MAG_EBAY_SUSRACK_Z_DQ


We are still working to include the April and September data, to produce different band passes to exclude the low-frequency calibration lines for cleaner plots and to produce completer magnetometer data for L1, but wanted to give a heads up early in the commissioning break that EX and EY electronics seem to have blinking devices.