The goal of atomic physics parity violation (PV) experiments is to test the "Standard Model" of particles and interactions. Weak interactions are known to violate parity and the exchange of Z° bosons between nucleons and electrons gives rise to a very small parity mixing between the S and P states of an isolated atom. For this reason, forbidden transitions have been chosen for atomic physics measurements of parity violation [1]. The electric dipole transition moment is measured by interference of the corresponding amplitude, either with the magnetic dipole amplitude, as in the measurements in Tl, Pb or Bi, or with a Stark induced electric dipole transition amplitude as in the Cs experiments in Paris [2], or in Boulder [3], both using a CW excitation of the 6S-7S transition. Beyond an independent determination of the Cs 133 anapole moment, increasing the accuracy of the PV measurements is crucial in identifying possible additional vector bosons Z°', since they might escape high-energy experiments. Today Cs still represents the best compromise between the enhancement of the PV effects in heavy atoms (Z-cube law [1]), a calculable atomic structure, and feasibility of the experiments. While the previous experiments on the Cs 6S-7S transition have been using fluorescence detection, the new scheme implemented at ENS uses a pump-probe detection method. A pulsed laser beam of linear polarisation excites the atoms on the 6S-7S transition in a longitudinal electric field. The collinear probe beam, of linear polarisation, tuned to the transition 7S-6P3/2 is amplified. Parity violation results in a dependence of the probe gain on the handedness of the trihedron (E field, excitation polarization, probe polarization). The asymmetry is of the order of 10^-6 for a typical E field of 1.7kV/cm and is odd in the E field reversal. The resulting change in the probe polarisation is detected by balanced mode polarimetry. In the ideal experiment where the probe and excitation polarization are parallel at cell entrance, the PV asymmetry is measured at each shot. Lineshape-independent calibration of the measured PV signals is obtained by a controlled rotation of the excitation polarisation. This calibration procedure remains valid even if the probe gain is high, in which case the PV asymmetry can be amplified. Due to the order of magnitude of the asymmetry (and the final goal of a 1% precision on the PV measurement), a thorough study of all the error sources has to be conducted, even if one major source of systematics, the interference of the magnetic amplitude with the Stark induced amplitude, is absent due to the longitudinal field configuration. This contribution will detail the present status of this experiment, of the identification and reduction of the systematic errors, as well as reduction of the different noise sources. 1) Probe polarisation measurement. The use of two-channel (balanced-mode) polarimetry allows us to reject the fluctuations associated with pulsed lasers, and reach the shot noise limit on individual probe polarization measurements. The change in the polarisation signal associated with the 6S-7S excitation pulse is extracted by subtracting the polarimeter imbalance measured without excitation pulse. Spurious signals related to the geometry of the (amplified) probe beam, or to electromagnetic interference, are rejected using a half-wave plate inserted and removed in front of the analyser. Under such conditions the zero of the rotation of the probe polarisation acquires a real physical meaning to better than 1 micro-radian. 2) Linear dichroism vs optical rotation. One has to discriminate the expected probe linear dichroism against the optical rotation due to stray birefringences or Faraday effect. For this we use a half-wave plate inserted/removed before the atomic medium, to switch the probe polarization by 90°. Birefringences on the excitation and probe beam paths are compensated separately. 3) Rotational symmetry around the beam axis.. Additional insertable half-wave plates have been implemented to rotate the polarisations by 45° steps. The behaviour of the data under these rotations helps to identify the presence of transverse E and B fields in the vapour. 4) Stray fields observed in the glass vapour cells. More generally, a study of the magnetooptical effects in the vapour has been completed and gives access to all the components of stray E and B fields, together with their odd and even parts. A major contribution to transverse E fields has been associated to the presence of Cs ions due to photoionisation of Cs dimers. Transverse magnetic fields have been shown to originate from conduction currents flowing along the cesiated inner wall of the glass vapour cells. 5) Improvements obtained with sapphire/alumina cells. The use of sapphire cells allowed us to overheat the cesium vapour and obtain a considerable reduction of the dimer content and of the stray fields associated with dimer photoionization. Sapphire has been chosen because a preliminary study showed a very small surface conduction in the presence of Cs vapour. This allowed us to operate with sapphire cells without internal electrodes, much easier to fabricate and handle than the previous glass cells, and to benefit from the corresponding drop in the surface conduction currents. Sapphire windows proved to be highly resistant, while those of the glass cells were rapidly damaged by the electronic/ionic bombardment. Highly parallel sapphire windows were used to prevent reflection at each window by choosing the temperature for which the interference is destructive. Thus the energy available in the vapour is maximum, and the problems linked to the beam reflected at the output window are avoided. Radial E fields and orthoradial B fields were detected in the sapphire cells, indicating the presence of an electron charge flowing along the axis of the cell. This can be understood as due to electron emission from the window at the negative potential, followed by amplification of the charge by secondary emission due to collisions at grazing incidence on the inner wall of the cell. To prevent the grazing incidence collisions, a cell was made with grooves on the inner wall of the cell body. This proved to be efficient in reducing the stray fields. The current sensitivity of the PV measurement and prospects of the experiment is presented. REFERENCES [1] M.-A. Bouchiat & C. Bouchiat, "Parity violation in atoms", Rep. Prog. Phys. 60(1997)1351 [2] M.-A. Bouchiat, J. Guéna, L. Pottier & L. Hunter, "Atomic parity violation measurement in the highly forbidden 6S-7S caesium transition", J. Physique 47(1986)1709 [3] C.S. Wood et al., "Measurement of Parity nonconservation and an anapole moment in cesium", Science 275(1997)1759.