power spectrum

[Dost]

I am usin gpower spectrum Expres sVI The thing wanted to ask is from DAQ board we get voltage (from pressure transducer) than we apply a linear relation to convert this voltage in to pressure. i use formula P=P0+ gain(V-v0). i want to know if i apply direct signal from the AI read to the Power Spectrum Express VI than will it be fine as than it will be just voltage not the pressure. Secondly iwanted to know more about the out of Power spectrum Express Vi .after reading some articles on NI developer zone I think power spectrum is measure of frequency with amplitude. the power spectrum gives this energy in RMS or peakvalues.if  frequency ranges from 0 to 500 in power spectrum graph and the RMS values shows us  say a value of -180 than what does this mean how would  we interprate it .KIndly suggest me some good article to know  more about it.

[RobSims]

Hi Dost,

The power spectrum of a signal is essentially a mathematical operation that allows you to examine that signal in the frequecy domain.

You will be familiar with interpreting signals plotted as amplitude against time, e.g. a simple 1Hz sine wave. If we were to apply a Fourier Transform to this signal, we will recieve back a power spectrum of this signal. A power spectrum is basically, as you said, a plot of amplitude against frequency. Ideally, in the case of a 1Hz sine wave, the power spectrum will look like a spike centred at 1Hz on the x-axis. This is shown in the top two graphs on the attached image.

If, however, you have a signal with a mixture of frequencies, for example a mix of two sine waves, one of 1Hz and another of 5Hz, then the power spectrum will contain two spikes centred on these frequencies. This is depicted in the lower two graphs on the attached image.

You can find lots of information on Fourier Transforms and power spectra on the web, or in physics/mathematical/engineering/science textbooks. A good website to refer to for in-depth information is http://scienceworld.wolfram.com/ .

Now, onto your questions:

Firstly, the conversion of your signal to the frequency domain is most reliant on the timing information in your signal. So by carrying out the Fourier transform to the raw voltage, as opposed to your scaled pressure values, you will not get any false information regarding the frequency components returned in your power spectrum. However, as the power spectrum plots frequency against amplitude you may wish to apply the transfromation to the pressure values if the amplitude infromation is important to you. You know what you need from your program, so its down to you as to which route you take.

Secondly, the options of RMS or peak values refer to the averaging you can employ on your data. The RMS option tells the function to carry out RMS averaging on the amplitude of the signal, however 'peak values' will not employ averaging.

The negative values returned to you are decibel (dB) values. This is a measurement of amplitude, commonly used when measuring sound pressure. 0dB relates to the loudest (highest amplitude) signal component, and quieter (lower amplitude) components are increasingly negative. Each decrease of 10dB realates to a halving of loudness/amplitude. Again, you will find more info on the ScienceWorld website.

I hope this helps. Thanks,

[RobSims]

Image attached as promised.

[Lammers]

I don't suppose you could display the block diagram for this front panel?  I've wired an auto power spectrum vi, but the waveform graph that I wired to it displays amplitude versus time.  I have a "build waveform" function with the power spectrum output wired to the Y input, and df output wired to the dt input.

[Lammers]

Attached is an image of my wiring diagram.  Still not sure why my waveform graph on the front panel shows "amplitude vs. time" rather than "power vs. frequency."  Any suggestions?  Thanks in advance!  This board has been incredibly helpful.

[Dennis_Knutson]

Amplitude and time or power and frequency are just labels with amplitude and time being the default. Use the text tool to change the label to anything you want or right click on the graph and choose Properties>Appearance>Scales. Select the axis and change the name.