While investigating the strange shape of the drive spectrum of the OSEM estimator from 84219. I found that the only plausible explanation for the discrepancy between the observed and predicted drives had to be a mistake on the calibration of the YAW transfer functions measured through the M1_TEST excitation point.

Turns out that both the Pitch and Yaw gains for the SR3_M1_TEST path are set to numbers different from 1 (they're set to 2.675 and 12.272 respectively, [see 2nd attachment]). Therefore, the exported transfer functions from the DTT template are miscalibrated for all the Pitch and Yaw drives for SR3. The factor of 12 discrepancy for Yaw can be directly seen in the SR3 M1 to M1 transfer functions that Oli took a month ago (see the .pdf attached to 83939, or the first attached figure).

These gains are correctly set to 1 for PR3, so the PR3 transfer functions are okay. However, we should fix the gains to 1 for SR3 so future diagnostic transfer functions are correctly calibrated.