Cavity locking using Lockin+PID technique

[Hashirama]

TL;DR

In summary I am trying to lock the cavity using the modulation demodulation technique, but I need help with the signals I am getting.

Hello, So I am trying to lock an X cavity to a 1064 nm laser. Laser has a linewidth of ~100KHz, cavity linewidth is 37.65KHz, and the FSR is 3.125 MHz. When I try to scan the cavity, the peaks usually shift and after a few seconds, the resonance comes off. I am scanning the cavity at around 100 Hz, and I assume it has some high-frequency noise that contributes to the jitter.
To lock, I am using Lock In+PID method, but with very slow modulation (20-30 KHz), my piezo where I am sending the modulation (attached to one of the cavity mirrors) can go up to 340KHz.

In the scanning image, the lock-in output is in green, the scanned peaks are in yellow. In PID image, blue one is error monitor and green is the output, I am not sure I am using it correctly.
When I use the modulated signal, the lock-in output behaves strangely, like at the peaks, it gives some signal, but whether it is the derivative or not cant be said. Same when it goes to PID.

I am attaching some pictures of what I got. Can someone please help me identify if it is right?

Thanks a lot!

 

[berkeman]

@kelly0303

 

[Hashirama]

@kelly0303

Hello, Greetings.
I have another question. If the cavity is locked at a certain point, What should I see as the cavity transmission from the photodiode? A "constant" voltage signal or the peaks? (While the ramp is off)

I think I managed to do it decently up to the error signal; I am attaching a picture of it.

Blue: cavity transmission, Green: Error from lock-in, yellow: error monitor from pid, setpoint as zero V.