Interferometer_sensitivity_studies (General)swinkels - 19:03 Wednesday 31 August 2011 (30213)
Print this reportVela bump caused by non-linearity in end-mirror actuatorsYesterday evening, I used 1 hour of commissioning time to study the origin of the non-linearity that causes the Vela bump. Two very strong lines (~ 5-10 times stronger than the calibration lines around 350 Hz and the DARM line) at 290 and 313.3 Hz were added to the correction of the end-mirrors. This should cause a bump around 23.3 Hz. As shown in fig 1, the bump is indeed visible at the expected frequency, but only if both lines are applied to the same mirror (either NE or WE). If the two lines are applied to two different mirrors, the bump is much lower (note that when applying a line to one mirror, the DARM loop will apply the same loop with reduced amplitude to the other mirror). This makes it very likely that the Vela bump is caused by a non-linearity in the actuators, and not by some non-linearity in the ITF (e.g. scattered light) or in the detection electronics (which is known to exist).
To study the frequency dependency of the bump, the 313.3 Hz line was moved to 323.3 and 343.3 Hz, which should move the bump to 33.3 and 53.3 Hz Fig 2 shows that the amplitude of the bump scales more or less with 1/f^2 (note that the frequency was only changed by 3 and 10%). This might be compatible with some extra force applied to the mirror. To study the amplitude dependence, the injected lines were both increased and decreased by a factor 2, see fig 3. The height of the bump decreases and increases more or less with a factor of 4, as expected when the bump depends on the product of the two lines.
The source of the non-linearity might be caused 'simply' by the electronics, but we cannot exclude some exotic mechanisms like eddy currents or Barkhausen noise (as seen at LIGO). Finally note that the non-linear process that caused the Vela bump is probably not quadratic, but cubic, since it involves a combination of 3 different lines (DARM line, calibration line and some 0.2 Hz resonance that causes the horns).