As agreed with Michal, I reduced the ETM RHs power by 50%, to allow arms astigmatism measurement.
13.06 UTC NE RH reduction to V = 11.7 V (3.7 W)
13.07 UTC WE RH reduction to V = 12.3 V (4.07 W)
As a second step, the ETM RHs have been switched OFF for the measurement of arms astigmatism.
17.29 UTC NE RH OFF
17.30 UTC WE RH OFF
Figure 1. Following the decrease in RH power the arm power increased by about 2%, peaking around 14:00 UTC. Shortly after the interferometer unlocked and then relocked with the power in the arms decreasing. This indicates that a better mode matching by ~2% is possible, and it corresponds to flatter end mirrors.
Figure 2. Shows in purple the position of the order 2 mode in the arm FSR before the RH change, and in blue at 14:00 UTC close to the peak in the arm circulating power. The mode is at 10.9kHz-11.1kHz when the power in the arms is maximum, this looks similar to the previous test in June of common RH change when the power in the arms was maximum for 11.15kHz, https://logbook.virgo-gw.eu/virgo/?r=67078
Figure 3 shows the position of the order 4 mode, it moves from 21.6kHz to 22.3kHz. Note that the position of the 56MHz sideband (assuming 56436993Hz frequency) is at 22.36kHz (assuming an arm FSR of 49'969Hz based on the O4 OptChar paper), so the change in EM RoC makes the order 4 mode and one of the 56MHz sidebands overlap.
Figure 4 shows the position of the orde 5 mode. It is not clearly visible before the RH change, and it is clearly at 27.6kHz at the peak of arm power. For the other 56MHz sideband (upper or lower) the peak in the arm FSR is at 27.60kHz, exactly the same frequency.
In conclusion the change in EM RoC increases the power circulating in the arms, but it also corresponds to making the 56MHz USB and LSB co-resonant with the order 4 and order 5 modes. So just changing the EM RoC doesn't seem like a good solution to improve mode matching, and the input beam needs to be change to modify the beam radius at the level of the input mirror, which requires acting on two actuators of the input beam. However, it is much simpler, and it doesn't seem like it creates any real problems to have the sidebands overlapping with the carrrier HOM.
