Kentaro, Lee, Daniel, Nutsinee

While tackling noise on high CLF power antisqueeze trace we found the source of mysterious CLF 'scatter shelves' to be the high SHG gain.

One might remember the weird looking shelves on the CLF error signal at 4kHz and 20kHz (alog46679 and alog47222).

Plot attached was taken today at 40uW transmitted CLF power, with high SHG gain being 30-31dB (~2kHz SHG transfer function UGF, see attached GIF file) and low SHG gain being 7dB (~200Hz SHG UGF). The 4kHz shelf isn't prominent here compared to the past alogs that's likely because the data was taken with relatively low CLF power, the 20kHz shelf however, is now gone. The broad peak at 5.5kHz is likely due to gain peaking as we were operating at higher CLF UGF than usual (that was not intentional, better CLF data is coming). We looked at the same comparison on Sunday with higher CLF power and saw that 4kHz shelf was eliminated (data was saved, but unretrievable).

At high SHG gain we had 35 deg phase margin on the SHG loop. We suspect the shelves were caused by SHG mechanical noise. Since then we have been operating at 7dB common gain where we no longer observe this coupling in CLF. This sadly put SHG UGF at 200Hz.  

Note that this still doesn't solve the problem with worsen squeezing at high CLF power.