hand-tuned demod phase:
P = 346.2
Y = 162.5

Very strange very-low-frequency (5 minute period) oscilations in the demod output.
Wasn't saturating anything...
Trying factor of 10 less drive
   OSC     Frequency [Hz]     Amplitude [ct]     Sin [ct]      Cos [ct]
    P         4.0              12000 (1.2e4)       1000           1000
    Y         4.0              12000 (1.2e4)       1000           1000

Much better... dunno why.
No perturbation
H1:SUS-SR2_LKIN_P_DEMOD_I_OUT    -1247.80963867    7316.11394178
H1:SUS-SR2_LKIN_P_DEMOD_Q_OUT    3.22766243837    45.1774166174
H1:SUS-SR2_LKIN_Y_DEMOD_I_OUT    -476.208041539    2759.93281889
H1:SUS-SR2_LKIN_Y_DEMOD_Q_OUT    -1.70296711835    12.9795196364

BIG PIT (use YAW outputs to measure YAW phase change)
H1:SUS-SR2_LKIN_P_DEMOD_I_OUT    -1199.64975791    37.1288197564
H1:SUS-SR2_LKIN_P_DEMOD_Q_OUT    1.70620713623    4.340853072
H1:SUS-SR2_LKIN_Y_DEMOD_I_OUT    -354.038072632    63.2362957799
H1:SUS-SR2_LKIN_Y_DEMOD_Q_OUT    -1.23790475251    2.7270743623

yaw phase change = 180/pi*atan((-1.70296711835 - -1.23790475251)/(-476.208041539 - -354.038072632))
                 = 0.218 (too small to care)


BIG YAW (use PIT outputs tp measure PIT phase change)
H1:SUS-SR2_LKIN_P_DEMOD_I_OUT    -978.093236321    160.949921027
H1:SUS-SR2_LKIN_P_DEMOD_Q_OUT    1.08216599345    3.16146457299
H1:SUS-SR2_LKIN_Y_DEMOD_I_OUT    -468.938958886    81.2320184814
H1:SUS-SR2_LKIN_Y_DEMOD_Q_OUT    -1.36779755241    3.51527231493

pit phase chase = 180/pi*atan((3.22766243837 - 1.08216599345)/(-1247.80963867--978.093236321))
                = 0.4558 (too small to care)


Needs WAY too much imbalance to make sense. I'm giving up for the night.