In Chapter 2,
Theory of Operation, one of the premises of the tape technique was linear
behavior of the vibration system with respect to frequency and amplitude. This
idealization is not often realized in actual test operations. In the initial
study program conducted a 2.5g (peak) sine sweep yielded more accurate
transfer characteristics, (for the required 6.0 Grms random spectrum), than
did the 1.09 (peak) sine sweep. The difference was only significant at low
frequency resonances and anti-resonances where friction damping decreased in
effectiveness as amplitude increased.
This situation does not follow non-linear vibration theory and tends to be
unique to black box attachment hardware. This hardware, such as standard
"Mil-spec" swing bolts and high-torque cam locking nuts, tend to respond in a
non-linear fashion when the attachment load is momentarily exceeded.
Therefore, since it is not considered desirable to increase the amplitude of
the initial sine sweep to 2.5g (peak), i.e., to minimize the risk of
accidental over test the following is recommended. Where it can be determined
that a non-linear resonant condition exists, due to mounting hardware,
decrease the synthesized random spectrum, in those filter bands affected, to a
maximum of -3 dB.