[Joan-Rene Merou, Alicia Calafat, Sheila Dwyer, Robert Schofield]

We have analyzed the evolution of the near-30 Hz and near-100 Hz comb amplitudes, continuing with the work on aLOG 87414. In that alog was shown how there appeared to be a correlation between the ITMY ESD bias and the amplitude of the near-30 Hz and near-100 Hz combs.

In this study, we first listed all time intervals between ITMX and ITMY bias changes from the study Robert Schofield did and reported on aLOG 78925 and the associated entries. Robert changed the biases in a number of occasions "for determining the ITM biases that minimize coupling of electronics ground noise". We wanted to see if this was also true for minimizing the amplitude of lines in DARM.

The full list of constant-bias segments in-between changes is shown in the following table, which defines the start and end GPS times, along with the corresponding ITMX and ITMY bias values. We have added a few data points in between the ones shown in the related Robert aLOGs.


  

    

      
const_start_gps
      
const_end_gps
      
ITMX_bias
      
ITMY_bias
    
  
  

    
1403650000
1403651100
0
0
    
1403651220
1403653448
0
-39
    
1403653569
1403655679
0
170
    
1403655799
1403657640
0
-222
    
1403657760
1403804070
0
0
    
1403804192
1403804365
0
77
    
1403804489
1403807629
0
-80
    
1403807751
1403809555
0
-20
    
1403809816
1403812204
-40
-40
    
1403812324
1403816907
36
-40
    
1403817127
1403820880
36
0
    
1403821000
1403912193
0
0
    
1403912517
1403914518
20
-20
    
1403914684
1403917205
-19
19
    
1403917405
1404317823
0
0
    
1404317979
1404319278
-20
-22
    
1404319453
1404320829
20
-20
    
1404320940
1404341414
0
0
    
1404341526
1404342991
0
-20
    
1404343117
1404344533
0
-40
    
1404344655
1404346187
40
-40
    
1404346300
1404413221
0
0
    
1404413147
1404416077
0
-40
    
1404416200
1404417200
0
0
  


For each of these intervals, we performed custom Fscan analyses using DARM and the 5-26 ADC channel as input data with 128 s Tsft. Each Fscan job covered one of the constant-bias segments, allowing us to investigate whether higher ITM biases corresponded to larger spectral peaks in the resulting spectra. The attached plots show the relative amplitude of the first harmonic for the two strongest combs identified near 30 Hz and 100 Hz.

In these color scatter plots the X and Y axes represent the ITMX and ITMY bias voltages, respectively. The color scale encodes the relative amplitude ratio (or signal-to-noise ratio) of the selected comb harmonic. These plots correspond to DARM (GDS-CALIB-STRAIN-CLEAN).
From the visual inspection, there appears to be a consistent trend where higher absolute ITMY bias values correspond to larger peak amplitudes, the same could be true for ITMX values, but the data does not appear to be conclusive. It would be interesting to get data points with much higher absolute values of ITMX bias.
Here is one example for the first harmonic of the highest near-30 Hz combs:



Although these results currently refer only to the first harmonic of each comb, this correlation already suggests a relationship between ITMY bias magnitude and the peak amplitude.

We also generated the same set of plots for the H1:PEM-CS_ADC_5_26_2K_OUT_DQ ADC channel, to test whether the observed trend in DARM is also visible in the ground current monitor. From these additional plots, the correlation appears less clear. For the near-30 Hz comb. It appears that changing ITM bias values does not change the comb amplitude in the ground monitor. Showing an example for the same comb as before:



The current analysis only considers a single harmonic per comb. The next steps are to extend the study to higher harmonics of these combs and check whether the same behavior is observed in other comb families beyond the 30 Hz and 100 Hz series.