A late entry with data:  It doesn't look like the ETMx offsets that Keita (alog 16591) implemented in an attempt to decouple the large pitch of the main chain from the reaction chain was doing much.

Attached are Pitch transfer functions of the top, middle, and lower stages of the ETMx comparing various times when the large pitch bias and the Keita offsets were on and off.  Not much change in any of the configurations in any of the 3 stages of the ETMx.  Note, according to Jeff's alog 16824, the Keita offsets were all turned off last Thur night.