Vibration resistance testing for technical objects is currently carried out using electromagnetic vibration exciters. The samples under test are mounted to dedicated fixtures, which in turn are attached to headexpanders or slip tables. While fixture fastening to tables is standardized, the design of the fixture is tailored to each application. Designing the fixture requires ensuring as little distortion as possible in the transfer of vibrations from the exciter head to the sample through the fixture and its mounting to both the sample and the slip table. The sample fixture in vibration testing should allow for the mounting of the object as in operational conditions and be characterized by low mass and cost, high stiffness, and a transmissibility as close to 1.0 as possible across the entire working bandwidth. This means that vibrations should be transferred from the exciter head to the sample without distortions. In practice, achieving unitary transmissibility across the entire vibration frequency range is impossible due to distortions caused by mechanical resonance, which generates undesirable amplitude reinforcement and phase shifts. Research conducted by the authors on actual fixtures mounted to headexpanders revealed the presence of additional components besides distortions from resonance. These components stand out for their significant share in the power of the signal representing accelerations transmitted from the exciter to the sample, while also being characterized by their harmonic nature. It was hypothesized that they arise due to the cyclic detachment of free areas of the fixture located between the screw connectors used to tighten it to the exciter head. To observe and quantify the hypothesized phenomenon, a research model of the fixture in the form of a flexible beam mounted at the ends was prepared. The system was subjected to vibrations in the vertical direction. Then, a computational model using the Finite Element Method (FEM) was prepared, and real experiment conditions were recreated. Calculations were performed using a solver for nonlinear calculations by directly integrating the motion equations. A satisfying convergence of simulation results compared to experimental model studies was obtained. Based on the proven convergence of simulation and experimental study results, it was concluded that selecting constructional features of fixtures for vibration testing is only possible based on nonlinear analyses. The currently common approach to designing fixtures based on solvers that use a simplification involving calculations in modal coordinates prevents the accurate reproduction of the behaviour of the fixture screwed to the headexpander of the vibration exciter.

Keywords: vibration testing, fixture design, fixture transfer function, nonlinear FEM