Signal Processing Converts Vibration Signals

Acquisition of vibration signals for such tasks as machine condition monitoring, manufacturing test, process control, and mechanical design requires selection of appropriate sensors. At first pass, you might think that sensor choice would depend on whether you are interested in acceleration, vibration, or displacement. After all, specific sensors exist for acquisition of each of these three. For several reasons though, other factors such as temperature, bandwidth, signal conditioning, and other factors are more important. For one, vibration sensor types including accelerometers, velocity probes, and displacement probes all have strengths and weaknesses based on these factors. This article discusses a second reason: you can apply some straightforward signal processing to find velocity and displacement from your accelerometer signal.
Vibration Shakedown A fundamental idea behind most vibration signal analysis is that you can break vibration down into components associated with your mach…

Take Control of Noise with Spectral Averaging

Most engineers have seen the moment-to-moment fluctuations that are common with instantaneous measurements of a supposedly steady spectrum. You can see these fluctuations in magnitude and phase for each frequency bin of your spectrogram. Although major variations are certainly reason for concern, recall that we don’t live in an ideal, noise-free world. After verifying the integrity of your measurement setup by checking connections, sensors, wiring, and the like, you might conclude that the variation you see is the result of noise.
Realizing that you cannot eliminate noise, you next set about trying to reduce its influence. If you are building a custom measurement system around a PC, you have a number of software options for the job. This article discusses how to reduce spectral noise with software-based digital signal processing (DSP) techniques.
Averaging is DSP Most modern engineering software for measurement and analysis ships with ready-to-use DSP routines, for noise reduction a…

How Precise is my Measurement?

Some might argue that measurement is a blend of skepticism and faith. While time constraints might make you lean toward faith, some healthy engineering skepticism should bring you back to statistics. This article reviews some practical statistics that can help you satisfy one common question posed by skeptical engineers: “How precise is my measurement?” As we’ll see, by understanding how to answer it, you gain a degree of control over your measurement time.
An accurate, precise definition To answer this question, let’s first settle on a definition for “precision.” Doing so is useful, because you might hear the term precision interchanged with accuracy or measurement error. Precision and accuracy are types of measurement error (Ref 1), where measurement error is the difference between a measured value and the actual (or true) value. Accuracy (sometimes called bias) is a number that quantifies repeatable or systematic measurement errors such as those that result from resolution limitati…