J. Kissel

With today's new infrastructure, which down-samples the output of the 524 kHz test banks to 16 kHz in the same way as the primary GW channel, we confirm that last week's addition of an extra 65 to 16 kHz AA filter improves the GW data stream by reducing several down-converted line features. 

Using the infrastructure configured as shown in the 1st attached screenshot, I compare a ~17 average, 64 sec FFT (~0.02 Hz bindwidth), ASD of
    - RED The 524 kHz version of the A1 path, which has NO anti-aliasing.
    - BLUE The 16 kHz version of the primary GW path, which now has 1x 65k and 2x 16k AA filters
    - GREEN The 16 kHz version of the A2 path, which has the 1x 65k and 1x 16k AA filter configuration used prior to 2025-01-23
    - BROWN The 16 kHz version of the A1 path, which has NO anti-aliasing and
See 2nd attachment, focusing on the region where I see the most improvement between 400 and 1500 Hz.
The The 3rd attachment shows the identical data set, but without DTT binning adjecent FFT bins that we normally used to aesthetically clean up the broad-band noise floor. 

One can see GREEN features that are not there in RED and also no in the improved BLUE that line up with features in the unfiltered, but down-sampled BROWN.

More, fruitful, data to come from this infrastructure, I'm sure.
I'm also sure that the CW group and DetChar must be happy with last week's change!