The manganese bath technique is the reference method for the calibration of neutron source emission rates. It is used to calibrate neutron sources using radionuclides in terms of neutron emission rate under 4p sr. The neutron source to be measured is immersed in a manganese sulphate solution where the emitted neutrons interact with the elements of the bath. In a typical configuration, approximately half of the neutrons lead to the creation of 56Mn via the 55Mn(n, γ) capture reaction. This radionuclide has a half-life of approximately 2.6 h and the bath reaches saturation when the number of nuclei decaying is equal to the number of nuclei created per unit time. The neutron emission rate from the source can then be deduced from the 56Mn activity at saturation, assuming proper modelling of the nuclear reactions occurring within the bath. The manganese bath facility has been recently refurbished in order to comply with appropriate safety and radioprotection regulations. This has lead to the upgrading of both the measurement methodology and the modelling of the bath, and a study on the development of a new detector for the on-line measurement of the manganese activity was developed. This new detector uses the 4πß-γ coincidence measurement method. The beta channel consists of photomultipliers tubes for the detection of Cerenkov light, and the γ channel uses a scintillation detector. The advantage of this measurement method is that it allows the determination of the bath activity without any prior calibration. A detector has been built and the results obtained have been compared to those from a primary measurement method already in use at LNE-LNHB. Furthermore, a comparison of the Monte-Carlo simulation results using GEANT4, MCNPX and FLUKA has been undertaken. This comparison revealed some discrepancies between the codes and uncertainty factors, such as the modeling of the neutron emission and the choice of the cross-section library.