Finding width of FFT peak

[Bgmnt]

Hi,

i have a signal with single peak at a particular frequency.

Now i wish to filter out that peak using a bandstop filter.

However, for that I need to find the bandwidth of that peak which is observed. (in frequency domain obviously)

Is there any way I can do that?

[Norbert_B]

---

@Bgmnt wrote:

Hi,

[..]

Is there any way I can do that?

---

Sure. The question is what you expect to get.

The first thing is that your signal has to be free of alias frequencies. Otherwise, all analysis is non-sense.

The second thing is that the resolution of the frequency correlates to the resolution of the time. That means: the higher your sampling rate, the better your frequency resolution will be after the transformation (FFT). => "Bandwidth of the peak" makes only sense if the width is greater than the resolution of the frequency.

When using the Peak Detector.vi, you could evaluate the width input and 2nd derviate output if that includes data you are looking for.

Norbert

[Bgmnt]

I am using tone measurement block for measuring tone frequency.

Isnt peak detection used for time domain signals?

[Bgmnt]

Also width is an input in peak detector vi. I want to display the bandwidth as an output .

[Norbert_B]

Please read the help for Peak Detector.vi.

This function works on arrays arrays. It doesn't matter if the array reflects data of time or frequency domain.

Norbert

[Bgmnt]

Hi.

width is an input to peak detector vi

I want the width of the peak as a n OUTPUT. that is i want to findth the range of frequencies that it spawns so that i can filter that out using bandstop filter of the detected bandwidth

[Rizwan00]

use waveform analysis multi scale peak detection vi...

 There are  two ways to tell s thanks in forums : either give Kudos or  Marked Solutions

[Bgmnt]

also

"Peak detection is one of the most important time-domain functions performed in signal monitoring".

as mentioned by NI "http://www.ni.com/white-paper/3770/en/"

[Bgmnt]

Rizwan,

width is an input for this too.

[BowenM]

Use peak detect to find it, then just fit a Gaussian to the peak. You can then calculate the FWHM.