The Auto Spectral Density or Auto RMS spectrum analyses uses Fourier Transforms to process optionally overlapped sections of the input data. The result of each Fourier analysed section is called a periodogram. We then process all the resulting periodograms to produce a spectral result. Continue reading What is Auto Spectral Density?
Author: Dr Mike Donegan
Mike graduated from the University of Southampton in 1979 and then went on to complete a PhD in Seismic Refraction Studies in 1982. Mike joined Prosig as a special applications engineer. He now researches & develops new algorithms for Prosig's DATS software and assists customers with data analysis issues.
What sample rate should I use if I am measuring whole body vibration?
Strict ISO8041 compliance imposes a minimum sample rate of 900 samples per second. This ensures the amplitude and phase characteristics of the applied weightings meet the required criteria. To take account of the anti alias hardware filter we would recommend 1200 samples/second. If non-strict compliance is used the minimum sample rate is 300 samples/second.
ISO2631-1 recommends a minimum data duration of 227 seconds. However, for greater reliability, 600 seconds is now considered essential. That is, a 600 second or longer signal will be more representative and will provide better quality results.
A user has three signals captured using a triaxial accelerometer and asked “What is the simplest way to get the XYZ resultant from run-up file?” He had tried forming a resultant of the raw time histories, but didn’t fully understand the resultant time history.
Of course, the correct way of processing the data is to calculate the individual waterfalls from the x, y & z data and then calculate a resultant waterfall. Continue reading How Do I Get An XYZ Resultant Waterfall
A user recently asked…
I want to show the principle stress increments on the left side of the graph and the load scale on the right side of the graph. I can do this with some other programs but cannot find a way to do it in DATS for Windows.
Is this POSSIBLE?
Use the following sequence to get the Articulation Index (AI) vs rotational speed:
- Perform waterfall analysis on your noise signal (+ tacho). This gives you a standard waterfall display
- On the waterfall result perform the analysis
Analysis->Noise, Vibration and Harshness->Spectrum Input->Sound Quality Auto Extract
Number of Cylinders = x (set to appropriate value)
Write metrics to Input dataset = True
Set required metric to True
In this case select either $AI_ANSI or $AI_VEH.
These are different calculations with
AI_ANSI is calculated in the range 0 to 1
AI_VEH is calculated as percentage in range 0 to 100