I have re routed DC coupled cables from various transmission and reflection PDs to a single BNC feed-through on the north side of the table. The HV amp and FGs have also been moved there so that there is now a central point at which signals can be observed. This should make some minor diagnostic tests much quicker, especially as the slow controls are operated from this side of the lab. Eventually we can migrate the South path PDH control over to the rack as well, the only issue will be finding a ribbon cable of the appropriate length. Most of the key points of alignment an polarization tuning before entering the cavity can be reached from that side, so it makes sense not to have various control loops activated from every corner of the bench.
I misalignment the north cavity alignment. The visablity was down to 13%, so it needed fixing. I was tweaking this up by tuning the laser crystal temperature close to the refcav resonance and then fast scanning over the resonance. Unfortunately the temperature of the cavity moved slightly and I ended up walking off without realising that it wasn't spacial mode match but frequency mismatch. I hope this won't take too long tomorrow to fix, but should be doable; the main issue will be finding the correct frequency to get at least some resonant action. The beat upstream beat note detector can be helpful with this, giving some absolute reference of frequency to work with.
While I am optimizing the PDH loops to hopefully improve the quality of the PLL beat note there is another test measurement that might be good to do. We can use the RF signal from the newly setup upstream beat-note detector to perform a measurement of free running laser frequency noise of the two lasers. I'd imagine we can just calibrate to Hz/sqrt[Hz] and adjust by a factor of sqrt OR opt to lock one laser to a reference cavity and leave the other free running. This is a measurement that we can't do with the other beat note detector. It is also a good way to test the PLL loop electronics with a sizable beat note signal.
I'm not sure why the PLL loop didn't have an RF amplifying stage before the phase detector for the beatnote signal. The mini circuits ZRPD-1+1 data sheet seems to suggest that it should ideally be run with +7 dBm at both RF ports for a ~1V signal. All the references I could find on the eLog suggested a previous beat note magnitude on order of -20 dBm to -30 dBm. I don't know if there is an advantage to placing all the amplification after the mixer and LPF stage or whether an low noise RF amplifier would improve the signal to noise. This might be old ground people have gone over, I haven't come across it yet in the elog. Important to note is that the isolation between RF ports is about 58 dB at the 70 MHz we have been operating at. This is enough that with the +13 dBm LO (that the present setup was configured to), that a leakage couple back into the PD would be at the same level as our actual beat note. This would be a concern if back reflections were to create parasitic circuits.
The best situation would be to obtain a better, cleaner beat note and then think about ideal RF powers for our circuitry.