I have compared the specturm of RFC Tra before and after replacement of the photodiode. To compare the spectra, I have used the spectrum of the new photodiode as it is, as it should be calibrated correctly into mW based on the value of the resistors used in the photodiode. For the old photodiode, I have adjusted the calibration to match the two spectra at 1501Hz, the frequency of the PSTAB line. They also matche well below 10Hz after the adjustments, which adds confidence that the comparison is done in the same units.

Figure 1 shows the two spectra. They agree well above 200Hz. Below that frequency there are bumps that are different frequencies, and the overall noise floor before the PD change seems lower. From a few mW on the photodiode one would expect a quantum shot noise of a few 1e-8 mW/rtHz, so at high frequency the spectrum is likely dominated by quantum noise. Depending on which data I look at I conclude the power is 2.3mW or 5mW on the PD. Hence it is difficult to be sure if it is only quantum noise.

At lower frequency one can try to explain the noise floor with a 1/f^0.66 shape. I honestly don't know if it is really the noise floor, or just the effect of the many bumps give a floor which looks to have the same slope as the mystery 1/f^{2/3} noise.

Yesterday morning, during the maintenance time, we made a visual inspection to the RFC tra setup. In particular, we tried to estimate which beam splitter is actually used to split the beam between RFC_TRA photodiode and camera without unmounting it. By only using an IR sensor card, we expect few mW of light going onto the camera, which would probably mean a 50/50 beam splitter or similar, and not the 1/99 as in the optocad design.
To be noted that, at the end of the intervention in LaserLab, we found that the RFC tra photodiode was almost completely misaligned. Since we didn't touch any optics, we couldn't identify which component moved.  Anyway, the realignment procedure was quite difficult since the RFC doesn't lock without the transmission trigger, and the new photodiode surface is much smaller than the previous one.

After the last maintenance, we swapped back the EAB_EOM_TRA_audio cable on its original photodiode and added a new one spefically for the new RFC_TRA photodiode (see fig. 1).

The cable, named 12 LL-EER has been connected to the dbox in eeroom "EER_DBOX3_ADC0" #40 (SN30) in place of the unused (and commented till today) channel INJ_SIB1_PELTIER.

The correspondent dbox configuration has been restarted today at 14h52m57 UTC, and the new channel INJ_RFC_TRA_AUDIO is now available.