By clicking here you can get a data set for a five volt square wave signal.  (This is the actual data file saved from within Benchlink.)

• There are some things to note:
• The entire file was saved, including the header.
• The left column is the time in milliseconds.  (So, you need to multiply this by 1000 to get it in seconds.)

• Determine the frequency components in the signal.  In other words, specify the exact frequencies that you find in the signal.
• In order to analyze the signal, you will probably need to extract the numerical data and put it into a text file that you can load into Matlab or Mathcad.
• Here is a link to a Matlab m-file that will do the analysis.  You may have to change the name of the file - to whatever you name the data file in the preceding step.
• Here is a link to a Mathcad file that will do the analysis.
• Determine the a's and b's for the Fourier expansion.  Do that numerically.
• Determine the size of the frequency components.  (Square the a and b, add together, then take the square root.)

        Why do this problem?  Well, there is a confusion area.  When you try to figure out the frequency components from an FFT, it is easy to get confused between the fundamental frequency of the data set (in other words, the reciprocal of the time length of the data record) and the fundamental frequency of the signal in the data set.  (Here the signal is a square wave - and it has some fundamental frequency and many harmonics - although it should be missing the even harmonics!)  You should know something about the Fourier Series for this signal, and you should be able to use that knowledge to check your understanding when you do the problem.

        What it comes down to is this.  You can firm up your understanding of how to use FFT data with a problem that you know something about.  Then, you're ready to tackle the problem where you have to FFT the signal you get from a vibrating bar.