Filter cavity RF synthesis (particularly green lock)

As I presented in the research meeting on 25-03-05, there was some issues with the green control RF electronics. Lately the limiting factor to long term squeezing operation is green lock stability. It seems to be coming from the unlock of the green mode cleaner, and when the error signal is inspected it is found to be in Q-phase, despite optimizing for I-phase during the daily activities. I decided to check the RF signals to see if it was an electronics issue. The relevant signals to compare are modulation (SHG EOM from DDS1 DAC0) and demodulation (GRMC DEMOD from DDS2 DAC2). Both of these signals are sent to ~ +14 dB RF amplification on the amplifier rack, though SHG EOM is annoying to unplug from the amplifier so I just measured it straight from the synthesizer box. I looked at the spectra of the signals. Modulation SHG EOM (first image) is fine, although the amplitude of the modulation is quite low due to the ongoing master clock issue. Demodulation GRMC demod has many noise peaks including a rather concerning one at ~ 250 MHz, about -20 dB below the main signal (shown below before sending to +14 dB RF amp).





I then looked at the signals in the oscilloscope to see if the phase difference stays stable over a long time. I found that GRMC demod drifts with respect to SHG EOM. The separation changes from 3.8 ns to 2.8 ns after about 40 minutes of waiting. I checked SHG EOM vs SHG DEMOD and in that case it works fine, no significant change.





Looking at the PDH signal, I locked the SHG and then checked GRMC error signal. It seems to noticeably change in shape over roughly the same timescale. So it looks like the RF signal is not very good. There was some discussion in the research meeting that the frequencies are not exactly the same and this may lead to some slight phase error. There could also be a possibility that the PLL of the DDS board is broken leading to desync between boards. I checked an unused port on DDS2 DAC0. In this case, the spectrum looks better than DAC2. The main peak has +4 dB signal and there is only a faint noise peak at -60 dBm. After sending through the RF amplifier I see some 2f and 3f spurious peaks but these are not as large.





I then looked at SHG EOM vs DDS2 DAC0 on the oscilloscope and there was no significant change in phase delay after 40 min of waiting. So it looks like DDS2 DAC2 is emitting a dirty signal.





I changed the GRMC DEMOD channel to DDS2 DAC0.

Also, someone took my wiper bucket and the other one next to the IMC vacuum tube had a broken seal so couldn't be used. I took one from on top of a toolbox near the end of the West tunnel.