The squeezer team has been unable to lock their TTFSS at full bandwidth, and in late August traced the problem to a line in the PSL FSS (alogs 64649 and 64653) at about 2.5 MHz.

Today I went to the PSL racks and had a look to see what knobs I could turn to affect the peak.  (N.B. the other 'knob' we can turn is slight tweaks of the FSS's EOM angle to see if that affects the peak, but that requires going into the PSL enclosure).  For now, I am leaving the PSL FSS gains somewhat lower, which removes the peak in the spectrum. 

The peak is very clear in the cable pulled from the enclosure labeled FSS TF IN2 (the one the squeezer team was looking at).  The peak is also present when connecting to the PMC's Mixer Out, although it's much smaller there, and is easier to see with averaging enabled.

When I unlocked the FSS, the peak goes away in the PMC Mixer Out.  This is expected, but confirms that the peak is somehow due to the RefCav / FSS.  Also as expected, when the RefCav is relocked, the peak comes back in the PMC Mixer Out signal.

I looked again at the FSS TF IN2 signal, and changed the gains of the FSS loop sliders.  Their nominal values are Common = 15 dB, Fast = 11 dB. 

• With Common gain at 13 dB (fast still at 11 dB) the peak is smaller.
• With Common gain lowered to 12 dB (fast still at 11 dB), the peak jumps all the way down to 162.8 kHz. 
• Common gain back to nominal 15 dB, lowering fast gain to -10 dB, no change to the 2.5 MHz peak.
• Common gain at 15 dB, fast gain up to 18 dB, get 162 kHz peak in addition to the 2.5 MHz peak.
• Common gain at 11 dB, fast gain at 8 dB, basically no peaks present, maybe a small broad bump around 160 kHz.
• Common gain at 10 dB, fast gain at 7 dB, somewhat smaller bump at 160 kHz, but not very noticable.
  • Peak also gone in the PMC Mixer Out.

I also measured the OLG of the FSS loop at several of these settings, to see what the effect on the overall loop would be.

• Nominal: Common gain at 15 dB, fast gain at 11 dB: FSS UGF = 320 kHz.  This is lower than what I thought the target was of ~450 kHz.
• Higer gain: common gain at 17 dB, fast gain at 11 dB: FSS UGF = 425 kHz. Note also that the little nub in the FSS OLG at 2.5 MHz is at -11 dB, so not at all close to unity gain crossing (but as expected, the 2.5 MHz peak is back and large)
• Lower gain settings: common gain at 11 dB, fast gain at 8 dB: FSS UGF = 230 kHz.  Since this setting does not have peaks in the spectrum, I am leaving the gains here.  The phase margin at 230 kHz is about 45 degrees.  We'll have to see if this is sufficient gain for main IFO locking and noise though.