J. Kissel, A. Staley, S. Dwyer

In order to trouble shoot her ALS noise model, Alexa took an equivalent measure of the MCL / MCF crossover model using the fast path, where as I chose the slow path. She discovered that the crossover transfer function was a factor of two higher than mine (see 10064). This was notably before I changed the MC2 compensation filters (see 10043). After I changed the filters, I've taken the slow-path version of the IMC_L / IMC_F crossover transfer function again, and can confirm Alexa's result. 

This confirms that the crossover is border-line stable up at 25 [Hz], uncomfortably close to the 29 [Hz] Roll notch. This may explain why ALS CARM handoff has been so difficult last night and today.

Suspecting utter debauchery between the two measurements only two days apart, Sheila and I confirmed the following:
- All digital settings are identical to what I'd described in LHO aLOG 10042
- The 9 MHz modulation depth is set at the nominal +10dBm
- The overall open loop gain has not changed

We tried to check if the demodulation phase had rotated, but the test was inconclusive because we couldn't get any I signal from the demodulator for unknown reasons. We also measured the raw, peak-to-peak voltage of the RF output of the photodiode, and got ~250 mV.

Clearly a more thorough investigation is in order. :-(.
In the mean time the mode cleaner remains locked...