The CW group noted in a recent alog (53439) and an email thread that one of our ADS lines in particular is extremely close to a pulsar that could be detected, and they reminded me that Keith Riles long ago put in an alog about what frequencies to avoid putting lines in at (alog 14836*).

The first attached plot shows a long average of the H1 DARM curve during a recent NomLowNoise stretch, with vertical lines overlayed to indicate the current values of our ADS lines, and with shaded regions to indicate where Keith's analysis says it should be okay to put lines (I use the 0.1 Hz veto regions for 2f, which I believe means that the non-green regions in my plot are within +\- 0.1 Hz of the 2f frequency of a pulsar). 

The second plot is similar, but it is a shorter average since I wanted NomLowNoise time when the ADS lines were off.  Here the green regions are the same as the previous plot, and the vertical lines indicate my candidate values to move the ADS lines to.  The red 'shoulders' on each of the ADS lines are +\-0.1 Hz of the ADS candidate frequency.  I need the lines to be at least 0.1 Hz apart in order to separate them with bandpasses that aren't too narrow.  These red lines are also a semi-reasonable proxy for the width of the upconversion shoulders on the current ADS lines, so I have put all of my candidate frequencies at least 0.1 Hz away from the edge of a green region.

This afternoon (if commissioning time is approved) I will work on moving some of the ADS lines.  My first priority will be the 22.347Hz line that the CW group called out as most 'dangerous', but if I have time I will also work on moving some of the others.  For each line that is moved I'll need to rephase and check the control loop, so I don't know that I'll get to all of them today.

* When re-searching for this alog, I find that EvanG put in an updated version in alog 44892, but my plots are using Keith's original version - I suspect that the okay regions are still fairly similar to their original values.