After having disconnected the potentiometer of the ramp in the actuator section of the rampeauto (see previous entry # 875) we have further modified the rampeauto to reduce its noise. Following Pierre's instructions, we have soldered a 2.2K resistor in parallel to the resistor R24, in order to reduce the gain of the PA83 amplifier and then the effect of its input noise. In practice, since the resistor R24 was difficult to access, we have put the resistor in parallel with C23. 

- The attached PDF shows the electronic scheme of the rampeauto (actuator part). 

- Figure 1 and 2 show the rampeauto noise taken at the PZT output without any input signal. We see that the rampeauto noise is reduced by a factor ~ 20 in the kHz region, as expected. 

- Figure 3 shows the error signal of the filter cavity (green) with this new configuration, compared with the old ones. Note that, since now the gain of the P83 is reduced, we should increase the gain of the loop increasing the signal at the input of the P83 with the potentiometer "PZT gain". With respect to before, the error signal at high frequency is reduced by a factor ~ 8. 

- Figure 4 shows the error signal of the filter cavity (green) compared with the noise projection of the laser frequency noise (7.5 kHz/f [Hz/sqrt(Hz]) and rampeauto noise. Both the laser frequency noise and the rampeauto noise have been filtered by the pole of the cavity (2.4 kHz) and multiplied by the open loop transfer function of the loop, which assumes a 1/f behaviour at f> 3 kHz and a unity gain frequency of 10 kHz. We see that with this new configuration the error signal of the filter cavity (green) is explained very well by the laser frequency noise. 

Note that after the modification of the rampeauto it was not possible to obtain a ugf larger than 10 kHz. This should be understood better.