[Miyakawa, Aso, Nakano]

I found that REFL45_ERR has huge noises, although it is the error signal for the CARM loop which has huge suppression gain. Futhermore, the CARM servo error point does not have such noise. This indicate that the CARM common mode servo has servo noise and it dominate the CARM noise. We started investigation of that.

As the first step, we found that there is huge 13 MHz peak at TEST_B channels, where is just before the slow output. Miyakawa-san put the notch at that frequeny into the IN1 port. The attached figure shows the spectrum for each channels with and without notch. Obviously, the noise of IQ demod output reduced.

However, still there is huge gap between the IQ demod output and servo mixer monitor port. We need more investigation.