I've continued my work on improving BSC St2 blends, I think I have something fairly Hippocratic. I moved the St2 ellips down to 1.5 Hz and added a pole at 10 Hz. This affects the St2 CPS phase (~24 degrees at the blend frequency) but this doesn't seem to negatively affect the performance. I've talked to the commissioners about trying this during a lock, I think I can install and turn the new filters on without breaking anything.
First plot shows my modifications. Red is stock, green is the "old" again, blue is my current.
Second plot shows the performance. Pink is ground before, green is ground after, blue is before, dashed and solid are CPS and GS13 respectively, red is after. Does pretty good between 1 and 20 hz, doesn't seem to do any worse below 1 hz. I think all of the differences below 1 hz can be attributed to the ground.
Jeff suggested I check LLO' blends to see if they had made a similar change, and it turns out they have. A nice demonstration of convergent evolution. Attached plots show the comparisons. The LLO blend has more gain peaking, but keeps closer phase up to the blend frequency and has more high frequency roll off. My elliptics dont afftect the gain peaking, but don't roll off as fast around 10 hz, but at that point my performance plots show we are already hitting GS13 noise.
