Have you ever “heard something” on your slow-speed machine, but not seen evidence of it in the vibration data? Conventional time waveforms and spectra can sometimes fall short for machines with shaft speeds under 200 RPM. Short-duration (high frequency), repetitive events are usually stripped out by anti-aliasing filters, leaving no way to identify these faults or gauge their severity. Impact demodulation captures and preserves these high frequency events by sampling the acceleration signal at the maximum sample rate of the data acquisition hardware. This presentation will explain the signal processing used to capture the impacting events, show how to interpret the resulting demod waveform, and provide some real-world examples where impact demod identified faults that would have otherwise been missed.

Michael S. Johnson P. E. is a former Naval Officer (Nuclear Propulsion) and is currently a Senior Vibration Engineer at AzimaDLI. Michael joined Azima DLI in 1992, is responsible for maintaining the Expert Automated Diagnostic System (EADS) rule base, and teaches the advanced course in setting up and using the Expert System. He is a licensed Mechanical Engineer in Washington State.