This is an additional explanation.

Originally, WFS used a DCcentring loop to eliminate the offset of the length signal. However, in the case of WFSf3, because it detects beat signals of RF45 and RF56, a centering loop using an RF signal at twice the RF45 should have been designed. However, the conventional DCcentring loop is now set by klog30793. Therefore, there is a concern that the offset of the length signal is not canceled.

However, as shown in FIG3, when an offset is set to the PRCL signal, the POPRF90 signal is at its maximum when it is 0, so the length signal is not detuned and is controlled by the LSC. The offset due to the length may be shown by DCcenteringloop, but according to FIG3, the offset may not be visible and may be small enough.　I would like to prepare a model with an RF90 centering loop, for that it may appear in PRFPMI.

It is certain that a mysterious offset has appeared in WFSf3 in FIGS. 1 and 2, but I don'tm know what it is coming from. I haven't done a test yet where I set an offset on the CARM and detune it, but it's indistinguishable from RFAM. It could be a coupling issue. BPC control of IMMT2 is possible from PRCL, so it may be worth trying.

Also, regarding the offset setting of WFSf3, the current DCcentringloop range is narrow and it is not possible to set a large offset, so I do not know which is the optimal offset.