05-23-2010 08:32 PM
03-14-2011 10:00 AM - edited 03-14-2011 10:01 AM
On a very related topic, is there a way for you to put frequency labels on the x-axis of the power spectrum of the Express VI for wavelet analysis, and also provide the code or strategy on how to determine the endpoint frequency for the X-axis for the power spectrum.
Don
03-15-2011 03:26 AM - edited 03-15-2011 03:32 AM
Hi Don,
Thanks for the feedback. Yes there must be ways to put the frequency labels on the x-axis of the power spectrum. We will put your suggestion into our feature request database and it will definitely be considered for ASPT 2012 release.
Mac
03-15-2011 06:15 AM - edited 03-15-2011 06:17 AM
Unfortunately, I am not sure I can wait that long !
If you can just tell me how to get the frequency endpoint value for your power spectrum display (e.g., is it the same as the end point from the FFT of the time-domain signal?), that would be of great value.
Thanks,
Don
03-16-2011 10:15 PM - edited 03-16-2011 10:16 PM
Hi Don,
The frequency endpoint value for the power spectrum display can be decided in the following way. Since we display only the single side spectrum, if the signal length is even, the frequency endpoint will be exactly the Nyquist frequency.
If the signal length is odd, the frequency endpoint will be a little bit smaller than the Nyquist frequency.
Mac
03-18-2011 08:01 AM
Ok - so let's do an example to make sure we have it right.
Nyquist frequency = fs/2 where fs = sampling frequency.
fs = 1/(delta t) where delta t is the time spacing between samples in the time-domain signal.
so
Endpoint (signal length is even) = Nyquist frequency = 1/(2* delta t).
Specific example:
Mumber of points = 11010
delta t (sec) = 4.541670e-09
Endpoint = 110091662 Hz = ~ 110 MHz.
Does this look right to you?
Thanks,
Don
03-20-2011 09:26 PM
Hi Don,
That's exactly right.
Mac