SRM M3, try re-balancing after coil driver swap

Former balancing numbers with SRM M3
H1:SUS-SRM_M3_COILOUTF_UL_GAIN 0.972
H1:SUS-SRM_M3_COILOUTF_LL_GAIN -1.014
H1:SUS-SRM_M3_COILOUTF_UR_GAIN -0.972
H1:SUS-SRM_M3_COILOUTF_LR_GAIN 0.964

Hand-tuned phase
PIT 145.200
YAW 154.300

Demod output with balance gains set to previous values:
H1:SUS-SRM_LKIN_P_DEMOD_I_OUT    834.140155364    984.572993454
H1:SUS-SRM_LKIN_P_DEMOD_Q_OUT    0.498583179486    1.96366386161
H1:SUS-SRM_LKIN_Y_DEMOD_I_OUT    -539.285394612    384.238087629
H1:SUS-SRM_LKIN_Y_DEMOD_Q_OUT    0.0392783470605    2.23174625352

BIG PITCH perturbation to refine YAW demod phase
+P is the wrong direction
BIG PTICH for YAW phase tuning
H1:SUS-SRM_LKIN_P_DEMOD_I_OUT    883.963894996    1.8593546841
H1:SUS-SRM_LKIN_P_DEMOD_Q_OUT    12.4137009488    3.2195018471
H1:SUS-SRM_LKIN_Y_DEMOD_I_OUT    -798.020498359    4.02128074456
H1:SUS-SRM_LKIN_Y_DEMOD_Q_OUT    644.880931431    1.6094056463

Calculate new YAW phase
DETLA =  180/pi*atan( ((0.0392783470605)-(644.880931431))/((-539.285394612)-(-798.020498359)) )
 -68.1375

Try 154.300 - 68.1375 =  86.1625 ~ 86.2

BIG YAW for PITCH phase tuning
H1:SUS-SRM_LKIN_P_DEMOD_I_OUT    1398.52565511    367.977293334
H1:SUS-SRM_LKIN_P_DEMOD_Q_OUT    -827.285357445    598.334033454
H1:SUS-SRM_LKIN_Y_DEMOD_I_OUT    -828.15270257    19.3013835775
H1:SUS-SRM_LKIN_Y_DEMOD_Q_OUT    647.178536304    2.30723444202

Calculate new PITCH phase
DELTA =  180/pi*atan( ((0.498583179486)-(-827.285357445))/((834.140155364)-(1398.52565511)) )
 -55.7137

Try 145.200 - 55.7137 = 89.4863

restored COILOUTF gains to previously balanced values. Scale of Striptool is huge, but Q phases didn't change much for a large I phase change... might be good.
PIT Q phase changed by 10
YAW Q phase changed by 20
... hrm

Change back to hand-tuned phase
PIT 145.200
YAW 154.300

Perturb YAW by 5%, P I&Q change as much as before. No surprise. 

Restore to COILOUTF gains to previously balanced values.
Let's try the other way with the phases,
YAW 154.300 + 68.1375 = 222.400
PIT 145.200 + 55.7137 = 200.900

Hrm -- that didn't work either.Going back to hand tuned phases. But with hand-tuned phases, Q phases change much more with perturbation. I'll stick with 
PIT 89.5
YAW 86.2

Balanced ...
H1:SUS-SRM_M3_COILOUTF_UL_GAIN 0.934242
H1:SUS-SRM_M3_COILOUTF_LL_GAIN -1.00269
H1:SUS-SRM_M3_COILOUTF_UR_GAIN -0.979864
H1:SUS-SRM_M3_COILOUTF_LR_GAIN 0.999795
or
H1:SUS-SRM_M3_COILOUTF_UL_GAIN 0.934
H1:SUS-SRM_M3_COILOUTF_LL_GAIN -1.003
H1:SUS-SRM_M3_COILOUTF_UR_GAIN -0.980
H1:SUS-SRM_M3_COILOUTF_LR_GAIN 1.000