Marc and Yuhang

Recently, it was realized that back scattered noise is more than what we had (elog2350). According to the fact that back scattered noise is almost the same when end mirror is misaligned, also it can be reduced when lens (inside arm of homodyne) are misaligned, we suppose the reason of this back scattered noise as following. Firstly, homodyne's LO hits on two lens in the two arms. After that, some power is back reflected and propagates to many optics (such as PR, BS, INPUT...)

To understand better how the mirror motion is causing back scattered noise, a line (for example, 20Hz sine wave) is sent to input mirror to excite a fixed frequency back scattered noise. When sending this line to INPUT, we would like to know how much we are moving INPUT mirror. But we couldn't find this information, so we did the following calibration.

1. We use diagui to send amplitude 100 to input mirror channel: INPUT_ZCORR_fil_exc.

2. Since FC is locked, we check correction signal sent to main laser. The correction signal is 103uV.

3. The calibration factor is 100*1e6*2/1e12 [V/m]

4. We get 20.6nm, which is less than half wavelength.

Since we could see the injected line in correction signal, this means we are exciting INPUT more than its residual motion. If the motion of input mirror is dominating back scattered noise, we should be able to see a peak at the same frequency with the line sent to INPUT.

Therefore, we think the back scattered noise is limited by motion of in-air bench.