NAOJ GW Elog Logbook 3.2
Yuhang, Aritomi
1. We tried to send 500mVpk-pk sine wave to main laser intensity modulation channel. It was not clear how much is the bandwidth for this channel. We will check this next time. In any case, the frequency was set as 10kHz.
2. We put ND1+ND0.5 filter for LO before homodyne. The voltage measured on the homodyne first eye is 220mV. We measured LO spectrum on homodyne eye one as H1ND2.txt.
3. We removed one ND0.5 filter. The voltage measured on the homodyne first eye is 659mV. We measured LO spectrum as well, named as H1ND1.txt.
4. We send 800Hz 700mVpp to IRMC locking servo noise injection point. And optimize noise subtraction manually.
5. We put back 500mV noise to laser intensity, then measure homodyne sub-DC spectrum as LONOI.txt. In this condition, no ND filter was used.
6. We measured LO sub-DC spectrum as LO0NOI.txt when there is no noise sent.
7. We measured homodyne sub-DC in this situation (from ~0.1Hz to 200Hz) (a temporary check indicates that it is about 70dB at least. We will inject more noise next time to see if we can achieve better CMRR.)
The CMRR was characterized by sending 10kHz 500mVpk-pk noise to laser intensity. To get the CMRR value, we compared two situation. One is when the LO is sent to only one eye of homodyne, but ND1 was used to make sure no saturation. Therefore, if the ND1 is removed, the noise should be increased by 20dB. The other situation is as usual way to measure shot noise, which is to measure sub-DC spectrum. The result is shown in Fig.1.
Then we also took a shot noise measurement at low frequency. The shot noise becomes flat until 3Hz (Fig.2). This is reasonable considering that the LO RIN measured in logbook 2988 shows almost 40dB RIN increase from 20Hz to 1Hz. Although we didn't have a clear number of CMRR in the past, we know that we achieved flat shot noise until 10Hz. Compared with the old usual shot noise, we should need more than additional 40dB CMRR to achieve shot noise like Fig.2.
Personal note: the 80dB CMRR is actually limited by the noise we can send to main laser intensity. Remind that the intensity noise modulation channel has an efficiency of 0.1A/V. So when we send 0.5V, we are actually modulating by only 0.05A the laser current. I think we can at least increase this noise by a factor of 10. But I checked the manual of our laser and I haven't found what is the damage threshold for this channel. This needs to be confirmed.