Research Of Systematic Errors In G Measurement With Time-of-swing Method

The absolute value of G was first measured in the laboratory by Cavendish in1798,which gave G=6.67(7)×10-11m3kg-1s-2with a relative uncertainty (ur) of about1%. Over the past two centuries, nearly300different values of G have appeared, however, the measurement pre-cision of G value has improved only at the rate of about1order of magnitude per century. The newest recommended G value of CODATA2010(Committee on Data for Science and Tech-nology) is6.67384(80)×10-11m3kg-1s-2with ur=120ppm. The adopted G values in CO-DATA2010are in poor agreement in their claimed uncertainties, in which the largest difference yields487ppm. This situation testifies the great difficulty in G measurements, and there maybe some unfound systematic errors.Based on the HUST-09G measurement in our laboratory, we develop the new work on G measurement with the time-of-swing method by using a high-Q silica fiber. The main im-provements in the experiment are:1. A silica fiber with Q-factor of5.00×104is used as the suspension fiber to reduce the fiber’s anelastic effect and advance the stability of the pendulum’s period. Due to the fiber’s high Q, the anelastic effect is reduced from-211.80(18.69) ppm in HUST-09G measurement to only-6.37(0.50) ppm. This means that the anelasticity will not be a main error source in present G measurement;2. To reduce the potential electrostatic effect, the silica fiber is coated with8-nm-thickness of germanium and8-nm-thickness of bismuth layers in turn. After coated, the period’s stability is about0.05ms yielded from one set of3-days data of the pendulum;3. The background gravitational field of the environment is measured by using a dipolar pendulum. After compensated with-800kg lead blocks, the influence on the period due to the background gravitational field is reduced to1/5of that of before compensating. Besides, the equilibrium position of the pendulum in G measurement is located at the peak value of the background gravitational field which leads to the least contribution to G value;4. A copper tube with5-mm-thickness is installed around the fiber to decrease the temperature discrepancy, and6sensors are installed near the fiber to monitor the temperature change. In experiment, the largest difference between the sensors is less than0.02℃, which contribute an uncertainty of<5ppm to G value;5. The magnetic damper is designed to reduce its influence on the pendulum’s pe-riod. At present, the correction from the magnetic damper is reduced from17.54(0.31) ppm to only0.13(0.01) ppm;6. Aluminum is chosen as the coating material of the pendulum instead of gold and copper. After improved, the correction for G value from coating layer is reduced from-24.28(4.33) ppm to-1.81(0.91) ppm.Up to now we have finished the primary experiment, and the combined uncertainty of G value is16.95ppm. After repeating the G measurement time after time, we will report the final G value with the hoped uncertainty of <20ppm.This work is supported in part by the National Basic Research Program of China under Grant No.2010CB832801.