Given the experimental difficulties, most of the available flame speed database is for relatively reduced thermodynamic conditions and for non-simultaneous variations of pressure and temperature. This limitation may be overpassed by using spherically expanding flames with the constant volume method. This methodology, introduced in the 30 s by Lewis and von Elbe, requires the knowledge of the pressure evolution in the combustion chamber. It has been penalized for a long time because of the underlying assumptions and problems in flame instability detection. This method has been greatly renewed recently by Egolfopoulos following a coupled experimental/numerical approach integrating the effects of radiation and dissociation while maintaining moderate computing costs. In parallel with this study, we have worked on an alternative method providing a maximum of information for each test minimizing uncertainties. The current study uses a new experimental device allowing simultaneous recording of pressure and flame radius inside the chamber during the full combustion process. The direct use of these data over the whole flame propagation allows testing kinetic schemes over large pressure and temperature domains with good accuracy. These new experimental targets allowed the identification of key reactions needing improvements.