Laser signal input for PDH lock

[kelly0303]

Hello! I have a laser with only one input for external signals. Currently I use that to send the error feedback signal when I measure the power transmission from my cavity (not using PDH). If I want to use PDH, I would need to modulate the laser frequency, which means that I need to send to the laser both this frequency at which I modulate, but also the error signal. Is it possible to do that i.e. send 2 signals through the same input to the laser (can I combine them somehow before sending them to the laser)? Or do I need to use an EOM (or drive my laser at the dither frequency and send the error signal to a piezo attached to a mirror)? Thank you!

 

[Baluncore]

If the error signal is a DC referenced signal with low frequency components, and the modulation signal is a higher frequency AC signal, then you might simply combine them with a simple resistor and a capacitor to the input.

 

[DaveE]

yes

 

[kelly0303]

If the error signal is a DC referenced signal with low frequency components, and the modulation signal is a higher frequency AC signal, then you might simply combine them with a simple resistor and a capacitor to the input.

Thank you for the reply! I really don't know much about the topic so in the situation you specified, how would the servo controller know which signal is which. As I understand, you lock to the derivative of the signal, such that the slope at the point you lock is zero. But if on top of that you have the modulation signal, the error signal will be zero plus some error (which the servo should correct for) plus the modulation signal. Are the electronics of the servo automatically subtracting this modulation signal from the overall signal, without me having to do anything?

 

[Baluncore]

Are the electronics of the servo automatically subtracting this modulation signal from the overall signal, without me having to do anything?

Hopefully, yes. The servo will employ a low pass filter that should attenuate the modulation. Where the error signal is from a synchronous detector, there will be greater separation of the servo and modulation.
To be more specific we need some idea of the bandwidth of the servo error signal, and the centre frequency of the modulation or carrier.

 

[f95toli]

Thank you for the reply! I really don't know much about the topic so in the situation you specified, how would the servo controller know which signal is which. As I understand, you lock to the derivative of the signal, such that the slope at the point you lock is zero. But if on top of that you have the modulation signal, the error signal will be zero plus some error (which the servo should correct for) plus the modulation signal. Are the electronics of the servo automatically subtracting this modulation signal from the overall signal, without me having to do anything?

Yes, it should.
How well it will work depends on the electronics. It is possible to implement a PDH using a lock-in amplifier; in which case you are efficiently band-pass filtering the signal meaning the servo should more or less ignore what is happening at other frequencies; you can even do "double PDH" using this method where you have two loops going at once.

 

[Twigg]

The fancy name for the resistor-capacitor circuit Baluncore mentioned is a bias-tee. It let's you add together a DC signal (servo output) and an AC signal (modulation).

Are the electronics of the servo automatically subtracting this modulation signal from the overall signal, without me having to do anything?

The servo will employ a low pass filter that should attenuate the modulation.

If you are doing PDH, it's not the servo but the heterodyne detection circuit that filters out the modulation from the error signal. There will be a lowpass filter after your mixer which will ensure that the servo never sees the modulation signal. I attached a quick sketch of the situation. The lowpass filter should reject the local oscillator frequency.