[K. Izumi, Y. Arai, R. Koz]
BS oplev's QPD values were around 300. So we suspected that gain swithch[high/low] was low.
However, it was high. So we instead used the binary output modules. We installed bit switches (s1605438* and s1807678*) for the TILT and LEN signals, respectively.
The switch module has eight stages. B0, B1, B2 and B3 increase the gain by a factor of 15.85 (B0:24dB), 3.98 (B1:12dB), 2.00 (B2:6dB) and 1.41 (B3:3dB), respectively. On the other hand, B4, B5 and B6 make a pole at 10 Hz and a zero at 1 Hz. B7 is a logic enable(LE) switch (Please check: Universal whitening filter V2).
Experimentally we determined that B7 needed to be off position. In this condition, we toggled B0, B3 and B4. As a consequence, all the QPD values became as high as 10000 digital counts.
With these changes, we succeeded to remove noise from the spectra of OpLev pit and yaw(see below).
