Time to Acquire data for power spectrum analysis

[Ken Kearney]

Hi,

I am trying to do a frequency analysis on a signal and how I do it at the moment is that I do power spectrum ptbypt and this works fine up to a point.

What happens is that if I try to analyse up to 20Hz, I find that I'm having to take around 30 to 40 seconds of data as a sample to analyse before it will return the result and I want to bring this down. Is it reasonable to expect that I should only have to sample a number of periods of the signal to be able to analyse it? i.e. I should be able to analyse a 10Hz signal after around 1 second - with 10 periods of the signal.

Can anybody help me in doing this?

Ken

[johnsold]

Ken,

Ten cycles is about the minimum for good results using FFT based methods, which includes most of the LV spectral ananlysis VIs. So you should be able to get something faster than your 30-40 seconds at 20 Hz.

Does your frequency vary significantly during this time? How good is the signal to noise (or interference) ratio?

Can you post a copy of your VI with some typical data? Acquire a block of data, display it on an indicator, Save Current Value as Default. Then remove the Data Acquisition parts of the program and change the indicator to a control which feeds the typical data into the analysis section. Post this simplified VI.

Lynn