As powertrain noise is better and better controlled, road inputs become more important. The trend to mount 6 cylinder engines in smaller cars also emphasizes the importance of road induced noise.
A method to qualify and quantify the different contributions is presented and illustrated. This methodology is based on a novel combination of existing technology: transferpath analysis, traditionally used for ranking of powertrain inputs on one hand and principal component analysis, traditionally used for visualisation of operating shapes in a multiple uncorrelated input environment.
As suspension inputs represent multiple incoherent sources, the classical vector summation used in noise path analysis is not applicable. On the other hand, root mean square summation of all contributions does not keep track of phase relations between suspension-body connections which are important in the understanding of the global picture.
Therefore the methodology uses as a first step principal component analysis to split all vibration and acoustic signals up into several independent phenomena. In a second step a classical transferpath analysis is carried out for each phenomenon separately. Finally results are either summed in vector form or rms form.
The details of the methodology are discussed. The sensitivity of the procedure to measurement noise is critically studied and mathematical procedures to reduce noise sensitivity are discussed based on vehicle data. A discussion of the results shows that the method gives accurate results for full vehicle testing.