Mechanical shock testing of crystal oscillators

The expected failure modes are loss of frequency stability as a result of shock and vibration. Environmental change which mechanically induce strains in the quartz cause a change in the frequency of operation, and mechanically induced strains in the mount will have an additional effect upon the frequency.
A tip-over test is a way to demonstrate the effect of gravity and acceleration on a quartz resonator. The tip-over test measures the change in frequency resulting from a 2G; i.e. the pull of gravity first from one side of the crystal then from the other. From this test, the fractional frequency change per G may be determined for various orientations of the crystal. The change in frequency resulting from larger, but constant, stress may be obtained by subjecting the crystal to the centrifugal force of a centrifuge. If the crystal resonator and its mounting system are rigid enough to remain free of mechanical resonance over the frequency range of the mechanical disturbance then the frequency deviation per G of disturbance, as determined from the tip-over or centrifuge test may be used to predict the expected frequency deviation for applied sinusoidal vibration up to the level where the elastic limits of the design are exceeded. Frequency changes of three general types are encountered in conducting vibration and shock tests.
First, there are those which dynamically follow the mechanical excitation and disappear completely and instantly upon cessation of the excitation.
Second, during the course of the vibration run, or during initial subjection to shock, the reference frequency may undergo a temporary change from that observe before the test, but returns within a few moments to its original value. The phenomena will repeat and is probably associated with the plasticity and flow-back nature of the solder or cement used to attach the resonator to its support.
Finally, there are permanent changes that may result from the chipping of the resonator, permanent deformation or fracture of the mounting system, or damage to the attachment points of the mount to the resonator.
The stiffer and more rigid the mounting system, the higher the natural resonance of the system but the lower the capability to withstand high impact shock. If the design is to undergo no permanent damage from high mechanical shock, either the deceleration time must be short enough for the natural compliance of the system to prevent the system from exceeding its elastic limits, or shock buffer must be designed into the system to limit the excursion of the resonator to displacement safely within the elastic limits of the regular mounting system. Designs incorporating shock buffers cannot be expected to provide frequency stability during shock, but if carefully designed, can provide minimal frequency change on a before-and-after exposure basis. The AT-cut crystal, with its relatively inactive peripheral mounting points provides opportunity for designs capable of withstanding much higher shock anti-vibration than designs with the resonator soldered to fine wire leads (low frequency cuts). The periphery of AT-cut crystals operating on overtone modes even more quiescent than those operating on the fundamental mode and such designs are generally preferred for maximum stability under shock and vibration.

Reference:
MIL-HDBK-978-B (NASA), 1 MARCH 1988, NASA PARTS APPLICATION HANDBOOK