E.G. Adelberger, J.H. Gundlach, B.R. Heckel, G.L. Smith and H.E. Swanson

In this past year we have made several improvements, both to the apparatus and to our understanding of systematic effects.

We placed a hollow copper jacket surrounding the temperature sensitive torsion fiber inside the vacuum vessel. The very thin torsion fiber has a very small heat capacity, and comes into temperature equlibrium with the jacket quickly through radiation. This has two advantages: the large heat capacity of the copper jacket insulates the fiber from temperature changes outside the vacuum vessel, and we can control the temperature of the torsion fiber by controlling the temperature of the jacket. We use the temperature control feature to reduce the linear drift of the torsion fiber by temporarily warming it up. In the future, we plan to fill the jacket with liquid nitrogen and take data where the properties of the tungsten torsion fiber are believed to be superior.

We are able to position the composition dipole to within 0.05 mm of the turntable center in both the horizontal and the vertical planes using purely gravitational means -- an extraordinary accuracy considering the strength of the gravitational interaction. To find the vertical position, we place a large q₂₁ moment on the pendulum and arrange the uranium to have a large Q₃₁ field. When the pendulum is on the midplane of the uranium, the orthogonality of the multipole moments require the torque to be zero. Similarly for the horizontal position, the pendulum always has a large q₄₄ moment and the source has a large Q₅₅ field. When the pendulum is on the axis of rotation of the source, there again will be no torque. A theorem which describes how multipole moments transform under translations¹ allows us to measure the displacements.

We identified a previously unknown systematic effect coupled to our magnetic swing mode damper. The damper was connected to the vacuum vessel through a short (3 cm), thick (75 micrometer diameter) fiber, and connected to the pendulum through a long (90 cm), thin (20 micrometer diameter) fiber. We had always assumed the center of mass of the pendulum hung directly below the suspension point, and that the thick and thin fibers were colinear. We discovered that at our level of sensitivity this was not necessarily correct. The center of mass of the damper-pendulum system hangs below the suspension point, but if the center of mass of the damper is not colinear with the fiber attachment points, then the center of mass of the pendulum does not hang below the suspension point. A significant torque then arises about the thick fiber as the pendulum is attracted to the source mass. We improved the situation by reducing the mass of the damper by a factor of 3, and by increasing the torsional spring constant of the short, thick fiber by a factor of 16 by doubling its thickness.

We are currently taking composition dipole data, and expect to measure the differential acceleration between lead and copper towards uranium at the 5 × 10^-13 cm/sec² level.