hide
Free keywords:
-
Abstract:
All breadboards for the LTP interferometer showed an extra noise term that was, until recently, not fully understood. In this report that noise term is investigated in detail. It turns out that it is caused by sidebands on the light. In our lab, these sidebands were caused by nonlinear mixing processes in the power amplifiers that drive the AOM, if electromagnetic interference at a frequency near the operating frequency (ca. 80 MHz) is picked up by the power amplifier. The disturbing nearby frequency is the frequency of the other AOM, with a difference of exactly f_het, causing multiple sidebands at integer multiples of f_het from the carrier. They appear as pairs with a phase relationship that corresponds to phase-modulation (PM).
Experiments with a very different electrical setup (in Glasgow) also showed sidebands which demonstrates that they are not caused by peculiarities of the Hannover setup.
While the effect of a pair of first-order PM sidebands cancels and causes no harm, only one of the second-order sidebands produces noise which cannot be cancelled by its second-order mirror image. Hence the second-order sidebands are the dominant noise source.
Various strategies of mitigation are also investigated. The two most important ones, both of which are already implemented as baseline for the LTP interferometer, are (1) to reduce the sidebands by careful EMC design and dedicated testing, and (2) to stabilize the optical pathlength difference (OPD) between the two fibers with a Piezo device. The combination of these two measures will reduce this error term to insignificance.
We have also investigated other noise sources such as laser amplitude noise and beam jitter noise. Laser amplitude noise does have an influence on the total performance of the interferometer. Using a laser amplitude stabilization (part of the baseline), its influence can also be sufficiently reduced. Contrary to earlier worries, we did not find a significant noise contribution from beam jitter noise in conjunction with quadrant photodiodes.
As part of this investigation we have also developed a mathematical model for the sideband coupling that fully describes their effect and has been experimentally verified. Furthermore we have developed various numerical procedures to find correlations between auxiliary data streams (such as alignment signals) and the main interferometer output. They are useful for diagnostic purposes, but in general too complex to implement on LTP.
Using only those procedures that are the baseline for the FM, the noise performance of the LTP EM interferometer in the lab is now well below its specifications at all frequencies, with remaining noise sources mainly driven by ground-based disturbances, such that we are confident that the LTP interferometer will perform well on orbit and will enable the detailed study of the behaviour and noise performance of the inertial sensor and DFACS systems, which indeed is the primary job of the interferometer.
Comment of the Author: Version 1.2 2008/07/01
