Techniques for Repeatable Pyroshock Testing on an Air Gun Shock Machine for Lean Satellites

摘要

An increase in the number of small satellites (especially CubeSats) launched on dedicated launchvehicles implies that small satellites must now be designed to withstand very high frequenciesand accelerations generated during satellite separation. These pyrotechnic events are difficult toreplicate during the verification phase in the laboratory. At least 6 shocks, 2 shocks per axis alongeach of the three mutually perpendicular axes (XYZ), are required during qualification testing;and 3 shocks (1 shock per axis) are required in the acceptance testing of the flight components asper NASA-STD-7003A. Generating such precise shocks on a representative test article has provendifficult in practice, requiring up to 25 trial-and-error tuning tests to determine the appropriateexperimental setup parameters that can produce the required SRS. Performing these tuning testsrepetitively subjects the test item to fatigue loads, which may lead to failure of components andjeopardize the mission. It is important to achieve the desired test levels in as few shocks aspossible since many Commercial-Off-The-Shelf (COTS) components used in lean satellites arenot usually designed to withstand severe environments. The air gun shock machine generatesshock by propelling a projectile onto a shock table using compressed air. The generated shock ismeasured in terms of a shock response spectrum (SRS).The shape of the SRS generated is adjusted by changing the pressure, or projectile material or theabsorber between the projectile and the shock table. The air gun shock machine can generate ashock on three axes simultaneously, reducing the number of shocks required to achieve the testrequirements. In this study, the initial tuning conditions for a pyro shock test (6 dB/Oct for 100 –2000 Hz, 1000 G at 2000 Hz) were derived from a database of previous shock tests using aweighted root mean square deviation algorithm. Using the derived conditions, shock tests weredone with a representative test article and the results investigated for repeatability. At anappropriate pressure, a specified SRS may be generated on a test article if the strain history andloading sequence of the absorbers used in the test is known. The SRS generated using prestrainedabsorbers were, in general, more repeatable than SRS generated with virgin absorbers. A newtuning method, that takes account of absorber preconditioning, reduces the number of tuning shocks to about 4 to 5 shocks, and minimizes over-testing in the high-frequency region of thespectrum is proposed.