But I think there's some error (~ 10% ?) in calibrating the beatbox. In elog #6815, slope near zero crossing point is about 68 counts/MHz, but now, its 60 counts/MHz. Also, zero crossing point in elog #6815 was 47 MHz, but now, its 45 MHz. 5FSR scan was done between these two calibration measurement.
Calibrating beat frequency to Y arm length change;
I used L = 32.36 m (figure above, bottom plot).
dnu_g / dL = c / lamb_g / L = 1.74 MHz/m
Wow. This is way too short.
You don't need to use Albertoo's arm length as his measurement was done before the upgrade.