This year has been crucial for the Section because three experienced scientists have retired and have been replaced by the staff transferred from the Radiometry and Photometry section in 2003 and one new staff member recruited in May 2006. The training of these new members of the Section has been the priority during the last few years, to ensure the continuity and quality of our services.
During the last year all three calibration services: voltage, resistance and capacitance have been audited by external experts. Special attention was paid to knowledge transfer from the experienced staff to the new staff. All auditors expressed their satisfaction with the level of competence of the new staff. In addition, a first Josephson on-site comparison was successfully carried out by the younger staff. Three on-site comparisons of Josephson voltage standards have been successfully completed, with the INMETRO, the NMIA and the NMi VSL. The agreement between the BIPM and with both NMIA and NMi VSL is of the order of 1 part in 1010, which is an excellent result. In the case of INMETRO, a higher noise level is responsible for a ten times larger uncertainty. A voltage comparison using Zener voltage standards as transfer standards was carried out with the NML (Ireland). A new comparison of resistance standards at the level of 1 
with the NIST, to link a SIM comparison to the KCDB, is under preparation. Preparations have also started for a capacitance comparison of 10 pF with the NIST, to validate BIPM's uncertainty budget and to prepare for the determination of the von Klitzing constant at the BIPM.
The renewal of the Josephson equipment is continuing with the development of a compact and transportable probe holder and new electronic filters. We have also tested a compact programmable microwave source which will, in the future, replace the much more complex present system based on a Gunn diode. The new automatic system for the more efficient calibration of Zener diodes has been tested and needs some further improvement.
In the field of impedance measurements, a new double current source for the cryogenic current comparator built last year, was constructed, demonstrating that the Electricity section still masters these important technologies. Some of the crucial components linking the quantized Hall resistance (QHR) with conventional resistors have been verified at the level of a few parts in 109 by comparing pairs of QHR samples connected in parallel and series configuration with a single sample.
The measurement sequence from resistance to capacitance, using a quadrature bridge, has been optimized. This is important for the planned measurement of the von Klitzing constant with the calculable capacitor and for our capacitance metrology in general. Further improvement is planned by addressing the frequency dependence of the resistors in the bridge.
We have started a programme to renovate the impedance laboratories. This has allowed us to move the capacitance laboratory close to the QHR laboratory. A cable between both laboratories now gives a direct connection of the QHR apparatus, without moving any standards.
Our work on time series analysis has been brought to an end by addressing the problem of calculating the variance of the mean of a time series of stochastically correlated measurements of a weak stationary process. The variance of the mean of white noise measured through a low pass filter was determined in four different ways, which were shown to be in good agreement. The results can be directly applied to some of the measurements in the Electricity section but the scope of applications is much larger.
Summary
