Accelerometer response

[tikina33]

In what form is the accelerometer responds during random vibration? Or in other words, how does the response plot looks like? Is it in g vs Hz or g^2/Hz vs Hz or g vs time? A known PSD profile is input and the area under the plot is G rms.
Thanks.

 

[IRstuff]

Whatever are you talking about?

An accelerometer measures acceleration, hence, its output is acceleration, usually in mV/g.

PSDs are processed from the raw data.

TTFN

FAQ731-376

 

[tikina33]

Thanks a bunch. I know that accelerometer gives data in volts/g. But I would like to know how the PSD is derived from the raw data. Do I have to know the resonant frequency of the system? And do I have to know the transmissibility? Please explain.

 

[IRstuff]

Is this for school?

TTFN

FAQ731-376

 

[tikina33]

No. I am trying to understand the process because I have never done a vibration test analysis before. Please help.

 

[IRstuff]

You are supposedly an engineer, so you should have some inkling of how a time series with time in the horizontal axis gets transformed into something with a horizontal axis in frequency.

This is not the place to get tutorials. You need to do some reading, like here:

http://www.plant-maintenance.com/maintenance_articles_vibration.shtml

TTFN

FAQ731-376

 

[floattuber]

The response plot will look like whatever you inputed, so if you used g^2/Hz, then that's what you will get out. If you happen to have a resonance at 1000Hz, then there will be a spike at that frequency in g^2/Hz.

 

[btrueblood]

To be a bit more specific than IRstuff, you need to review Fourier transforms, and then discrete Fourier transforms. These transform time-based data functions (or series of discrete time-based data points) into frequency space.

 

[IRstuff]

If he couldn't figure out "time" "transformed" "frequency" he needs more help than is possible here.

TTFN

FAQ731-376

 

[BobM3]

A PSD producing program takes in acceleration (vs. time) in g's and produces a plot with units of g^2/hz on the vertical axis and frequency on the horizontal axis. That plot tells you where the energy of the signal is. If there is a peak near 10 hertz then you might have a resonance near that frequency (or just a lot of activity near 10 hertz for a reason other than resonance). The sharpness of the peak gives an indication of the damping at that frequency.

 

[btrueblood]

I don't disagree, IR - but somebody else may stumble on these posts and wonder...