Today I increased the peal-to-peak voltage from 0.4Vpp to 0.8Vpp which is applied to the EOM for modulation.
This voltage corresponds to ~0.2[rad] modulation.
In addition, I adjusted the phase shift for demodulation.
The first picture shows before changing, and the second one shows after.

• Blue line : DC REFL
• Green line : error signal
• Red line : control signal (but not yet optimized)

On the other hand, the dip got worse than before.
In order to lock the laser, I have to improve the alignment.

Next step is to improve the alignment with two STMs, and increase the dip of reflected beam.
Then I will try to lock by feedbacking to PZT of the laser.