Today we have continued with the activity of locking the arm cavities with the green lasers. We started by making some measurements of transfer functions, to better understand the system.

First, as a sanity check, we measured the correction filtered applied, without surprises, see Figure 1. Then we have injected noise using the fast DAC and we have measured the Transfer Function between this noise and the output of the VCO (the demodulated signal). In this measurement we find an unexpected behaviour, see Figure 2, and see Figures 3 and 4 for the fit (north and west arms respectively). We observe a pole around 90kHz with a high quality factor (~5) and we measured a delay of about 5-7 us. Part of this delay might be explained by the demodulation process, but not the structure.

Investigating further using the spare VCO we have realized that this shape comes from the use of the differential output of the fast DAC. Using the single ended output of the fast DAC we don't see this behaviour. Moreover this behaviour could interfere with the sidebands at 70kHz.

We have made some locking tests as well, one typical example of what we observed today can be seen in Figure 5. We started with a small gain and we increased it. At the beginning things seem to improve, as it can be seen on the transmission. However, then the situation worsens, until we reach a "zone" where we start to experiences some "jumps", similar to a bistability, as seen in Figure 6.

Alain has added the transmitted power as a trigger for the loop, to try to engage it close to the working point with a higher gain. The normalization is also available now.