Figure 1

The atom cloud is launched vertically with a slight horizontal velocity into the illustrated four-pulse Raman sequence. The retroreflected Raman beams are indicated by the two horizontal blurred lines, separated by a distance consistent with the pulse timing. The desired interferometer path is indicated with the black lines, and the undesired paths are indicated by the gray lines. The two closed three-pulse loops lead to spurious signals. A second atom fountain (to the right, similarly oriented and sharing the same optics) is not shown.Reuse & Permissions

Figure 2

Experimental demonstration of multiple-path elimination. The transition probabilities from the two gyroscope outputs are plotted parametrically for $dT=100\mu \mathrm{s}$ (black, multipath eliminated) and $dT=0\mu \mathrm{s}$ (gray, multipath present). The transition probabilities are obtained from measurements of the number of atoms in each of the two hyperfine levels following the interferometer pulse sequence. The two interferometer outputs were tuned to be $\pi /2$ radians out of phase.Reuse & Permissions

Figure 3

Measurement of Earth’s rotation rate as a function of sensor orientation angle in lab coordinates. Solid points indicate outputs of the two interferometers (red and blue). The black sinusoid is a fit to the data. Also shown are the fit residuals. The phases were determined from fitting sinusoids to the interferometer outputs. The interferometer fringes were scanned by electro-optically adjusting the phase of the final Raman pulse.Reuse & Permissions

Figure 4

(a) Phase shift as a function of induced rotation about the axis mutually perpendicular to ${\mathbf{k}}_{\mathrm{eff}}$ and $\mathbf{g}$. Red and blue points indicate outputs of each of the two interferometers, respectively. The solid black line is a fit to the data by assuming a fixed prefactor of six and adjusting only the offset, while the dashed black line assumes a fixed prefactor of four. A fit to the data at higher rotation rates gives an estimated slope of $6.07\pm 0.2$. (b) Normalized contrast as a function of rotation rate about the vertical axis. Contrast was determined by fitting 100-point parametric ellipses. The upper curve is for a $dT=20\mu \mathrm{s}$, and the lower curve is for a larger time asymmetry $dT=500\mu \mathrm{s}$, demonstrating that the performance is limited in part by the acceleration sensitive term.Reuse & Permissions

Figure 5

(a) Output of parallel interferometers (red and blue) and the subtracted signal (black). (b) Output of a single sensor sitting on the side of the fringe (red), predicted signal from the auxiliary sensor (purple), derived from the measured rotation rates during the pulse sequence, and the subtracted residual (black). Corresponding histograms indicating 60.5% cancellation, limited by the auxiliary sensor noise floor.Reuse & Permissions