SUBTLE ERRORS IN DIGITAL EXPERIMENTAL DATA PROCESSING

Abstract

Summary Seemingly minor instrumentation and data analysis errors are shown to invalidate damping measurement test results. Both a hardware error and a software error introduced time shifts between data channels that became quite significant. These time shifts in the applied force and resulting displacement data were relatively small, yet they masked the true phase shift due to material damping. These experiences illustrate the problems of moving from continuous conceptual models to experimental determination using digital data processing. In particular, a systematic bias error in the velocity calculation was found to be particularly detrimental in this case where a small quantity was calculated from a series of large numbers. This example illustrates the need to be vigilant in using data acquisition equipment since even well functioning hardware can contain the seeds of improper functioning under conditions unforeseen by its designer. As users, we must be alert to the functions that the equipment is performing.