I think the reason of the high mixer noise is the amplifier for the RFPD signal before the mixer. I remove it for now and the noise seems comparable with that from RCAV's RFPD. The beat measurement between with and without the amplifier is quite similar.
the result will be posted soon.
1)Here is the noise at the error point from ACAV loop during lock.
The mixer signal is really high. The sensitivity range on SR785 was on -50dB. To check that there is no weird coupling
between the signal and the spectrum analyzer I tried changing the range up to - 20 dB, but the noise level is still the same ~ 1uV /rtHz.
2) slope of the error signal,
a) connect the triangular signal from the function gen, 10Vpkpk, 2 Hz to piezo sweep ch on the PDH box, turn the switch on.
b) unplug the mixer signal from the PDH box, connect it on the scope, ch A. connect the triangular signal on ch B.
c) pk-pk value is 424 mV.
I also tried sweeping on the FAST channel. The triangular signal is sent to FAST on the laser, 10Vpkpk 20 Hz
and read out the mixer out signal again. I get the same pk-pk value, 424mV.
Then the width of the linear region is approximately FWHM of the cavity108kHz, (cavity pole x 2), the slope is then
108kHz/424 mV ~ 0.25 MHz/V
nice, much better, i only have two little things at the moment:
we should check the mixer signal, the noise level is too high. We should compare it with the other EP signal (of the other cavity). If there is a huge difference between both we should start searching there. We should also check the RF noise level of the RF-PD. Maybe something is broken there.