The B2 8MHz feed-forward into the SSFS did not give the hoped for improvement in frequency noise.

Conclusion

• B2 8MHz feed-forward to SSFS is not improving frequency noise
• The feed-forward work in itself, and is able to improve the B2 6Mhz signal that is used for PRCL
• B2 4MHz is coherent with B2 8MHz so will not work any better
• The nature of the 2MHz might be different at different port of the interferometer.
• Next idea to try is to demodulate B4 at 4MHz and use that for a feed-forward to the SSFS

Details

FIgure 1 shows PRCL. Where the same result was obtained as during the CARM NULL 1F test. With a weight of 3e-2 (yellow line), the FF improves the PRCL spectrum compared to the reference case (blue and cyan).

Figure 2 however for h(t), which had a significant contribution of frequency noise at that time, there is no improvement for any of the test compared to the referecne (blue and cyan). And both 3e-2 (yellow) and -3e-2 (green) FF degrade the sensitivity in a comparable way.

Figure 3, shows SRCL. There is also no improvement, and degradation for 3e-2 and 3e-2 FF. The degradation is higher for 3e-2 (yellow line) than for -3e-2 (green line).

Figure 4 the feed forward adds an offset into the SSFS error signal, as the FF signal is added without any filtering (and it is not possible to add any for this fast FF in the SSFS).

Figure 5 shows that B2 4MHz and B2 8MHz are coherent (but the quadratures are swapped), and B2 4MHz has no coherence with h(t) or SRCL despite the poor CMRF. So a feed-forward of B2 4MHz is unlikely to work any better than B2 8MHz.

/users/mwas/ISC/medianSpectrum_20230804/medianSpectrum.m