daniel.sigg@LIGO.ORG - posted 13:25, Wednesday 08 February 2023 - last comment - 09:29, Wednesday 15 February 2023(67297)
Full Bandwidth DCPD Spectrum
Here are the spectra of the DCPDs with a 230 kHz bandwidth.
Non-image files attached to this report
Comments related to this report
The attachment shows a comparison of the 524 kHz channel H1:OMC-DCPD_A0_OUT against the downsampled 16 kHz channel H1:OMC-DCPD_A_OUT_DQ, which is used along with the corresponding B channel to form the 16 kHz h(t). With a 32 s FFT and 40 averages, aliased peaks appear starting around 760 Hz, 780 Hz, and 840 Hz, with many more peaks and a few broad features appearing above 1 kHz. The spectrum below 100 Hz is not shown, but no additional peaks were seen there.
This data lives in /ligo/home/evan.hall/Public/2023/02/DCPD_fast_test_points.xml
Images attached to this comment
To better facilite future design upgrades to anti-aliasing to try to mitigate what Evan points out above, I've found the DTT templates in these aLOGs and exported them to text file for easy consumption. For Daniel's data in LHO:67297, the two attachments are just linear and logarithmic displays of the same data set. I found the DTT template for this ASD in /ligo/home/daniel.sigg/Desktop/DCPD/DCPD_230208_B.xml and copied it into /ligo/svncommon/CalSVN/aligocalibration/trunk/Common/Electronics/H1/SensingFunction/OMCA/Data/ 20230208_H1_TIAxWC_OMCA_S2100832_S2300003_DCPDAandB_524kHzvs16kHz_asd.xml and exported the two A0 and B0 traces into one text file, /ligo/svncommon/CalSVN/aligocalibration/trunk/Common/Electronics/H1/SensingFunction/OMCA/Data/ 20230208_H1_TIAxWC_OMCA_S2100832_S2300003_DCPDAandB_524kHzvs16kHz_asd.txt with columns [freq, ASD of A @ 524 kHz, ASD of B @ 524 kHz]. For Evan's data in LHO:67328, I copied the data from his template from the location he mentioned above over to the same CalSVN location, /ligo/svncommon/CalSVN/aligocalibration/trunk/Common/Electronics/H1/SensingFunction/OMCA/Data/ 20230210_H1_TIAxWC_OMCA_S2100832_S2300003_DCPDAandB_524kHzvs16kHz_asd.xml and exported it into two files, because he saved the reference trace at 524 kHz, and the live trace at 16 kHz. /ligo/svncommon/CalSVN/aligocalibration/trunk/Common/Electronics/H1/SensingFunction/OMCA/Data/ 20230210_H1_TIAxWC_OMCA_S2100832_S2300003_DCPDA_16kHz_asd.txt 20230210_H1_TIAxWC_OMCA_S2100832_S2300003_DCPDA_524kHz_asd.txt with columns [freq, ASD of A] (at the respective rate in the file name). Unfortunately, these two sets of data were taken at different times, but I'm crossing my fingers and hoping that these features are not time-dependent and/or wandering. So, by these two data sets combined, I should be able to do the numerology to make plots and understand which features in the 524 kHz data are being folded down to the 16 kHz data -- and improve he design of the anti-aliasing filters such that we *don't* get these extra features. The above 524 kHz text files are too large to attach, so if you want them for yourself, download them from the location above.
For the record, the data for both these data sets from Daniel and Evan were taken with the detector in "nominal low noise." However, as we all know this can mean many different things. For Daniel's data, taken with a start time of 2023-02-08 20:53:25 UTC, the data was taken exactly at the time that Craig & team were sweeping the DARM offset. As such, the light on the DCPDs was stepping from 40 mA (sum, 20 mA on each) to 8 mA (sum, 4 mA on each). I also see that the BNS range is "only" 110 Mpc, which to my implies that there was now squeezed light from the FDS system being injected at the time. Evan's data, taken with a start time of 2023-02-10 16:48:13 UTC, the DCPDs had a rock solid 20 mA sum (10 mA on each), as is indeed the nominal amount. The BNS range was pushing ~140 Mpc, which implies that squeezing *was* being injected. It would be most interesting to capture / compare data when there is *no* light on the DCPDs vs. these kinds of times. I'll work to get that "dark noise" data.
Images attached to this comment
