Testing the Principle of Equivalence in an Einstein Elevator

Abstract. Improving the level of accuracy in testing the principle of equivalence (PE) requires reliably extracting a very small signal from an instrument's intrinsic noise and the noise associated with the instrument's motion. In fact, the spin velocity required to modulate a PE-violating signal produces a relatively high level of motion-related noise and modulation of gravity gradients at various frequencies. In the test of the PE in an Einstein elevator under development by our team, the differential acceleration detector free-falls while spinning around a horizontal axis inside an evacuated, comoving capsule released from a stratospheric balloon. The accuracy goal of the experiment is to test the PE at an accuracy of a few parts in 10^15, a limit set by the expected white-noise sources in our detector. The extraction of a very small signal from the prevailing noise sources is necessary for the experiment to succeed. In this paper, we discuss different detector configurations and describe a particular design that is able to provide a remarkable attenuation and frequency separation of the effects of motion and gravity gradients with respect to a PE-violating signal. Numerical simulations of the detector's dynamics in the presence of relevant perturbations, realistic errors, and construction imperfections show the merits of this configuration for the differential acceleration detector.