Relative performance of transducers mounted in four different ways was evaluated in terms
of peak particle velocity (PPV), peak vector sum (PVS) and frequency. For the given
tolerance, the transducer freely placed on the surface recorded a few anomalous values of
PPV and PVS at corresponding acceleration levels lower than 0.20 g, which might be
suspected for poor coupling. Trace matching by superimposing one waveform on the other
did highlight the difference between the two waveforms. The greater the difference, the lower
was the correlation coefficient. This study indicates that transducers should never be placed
freely on the surface irrespective of anticipated vibration levels. As a few cases were also
suspected for decoupling with sandbagging or spiking, it would always be safer not to use
these methods.