Paul Schwinberg, Peter King, Rick Savage

We found that when we increased the gain of the frequency stabilization servo (Table-top Frequency Stabilization Servo - TTFSS D040105 Rev. B S/N LHO03) to achieve the required bandwidth of around 400 kHz the servo lost lock frequently and had difficulty re-acquiring lock.
Investigations with the RF spectrum analyzer revealed peaks in the Open Loop Transfer Function (OLTF) presumably from resonances in the phase-correcting EOM (Pockels cell) near 793 kHz and 1.8 MHz. (See OLTF.TIF, attached)

We loaded C16 = 150 pF and tuned the existing notch in the Pockels cell path to 793 kHz using C50. (See NOTCH.TIF, attached)

We then added a 20 pF fixed capacitor and a 25 pF variable capacitor in parallel wired in series with a 220 uH inductor and installed between the upstream side of L2 and Ground.  This was tuned to give a second notch in the PC path at 1.8 MHz. (See 2NOTCH.TIF, attached)

Then, by locking the TTFSS with only the FAST path (PC path disconnected), we found a peak in the OLTF at 35.7 kHz which is likely from the FAST actuator (piezo) in the NPRO. (See FASTONLY.TIF, attached)
We tuned the notch in the FAST path to 35.7 kHz after changing L3 to 1000 uH and C49 to 18 nF and 2.2 nF wired in parallel. (See FSTNOTCH.TIF, attached)

After some tuning with the TTFSS installed, the OLTF was measured with the TTFSS locked using only the FAST actuator. (See OLTFFAST.TIF, attached).
With both the FAST and the PC actuators connected, the crossover was adjusted and the loop UGF was set to about 400 kHz where we have about 50 deg. of phase margin and close to 20 dB of gain at 100 kHz. (See OLTF3.TIF, attached)


The loop operates robustly in this configuration.


Prior modifications:
As found (and verified):
R50 = 499 Ohm
C62 = 1000 pF
R44 = 4.75 kOhm
R41= 24.9 kOhm