Chirp spectrum

[rashp8]

I am facing certain problems regarding the chirp signal. To start with, why the power spectral density of inbuilt chirp signal has time vs. amplitude plot and not the frequency vs. amplitude.

[ConnerP]

Computers only do what they are told. Labview graphs aren't smart enough to update their axes automatically. Double click the axis label and edit it yourself, or do it through the properties window.

[rashp8]

Thanks for replying so soon. Even when I change the legend on X-axis from time to frequency the graph doesn't change. It is still showing a single frequency peak which is not true in case of chirp signals. A sinusoidal is a single frequency signal not the chirp. The spectrum of chirp must be different. 

[rashp8]

I am attaching the graph image that I am getting.

[jamiva]

---

@rashp8 wrote:

I am attaching the graph image that I am getting.

---

It works fine for me. So unless you post your VI (not a picture of your results), there's not much I can do for you.

[ConnerP]

My guess is you are seeing the results of a windowing function applied before the PSD which is a Hanning by default if you are using the waveform version of the PSD.

[rashp8]

Yes. It was windowed. Now I got it. Thanks a lot. 

[rashp8]

Now I also got the result. Thanks a lot. Also can you tell me why chirp spectrum behaves like this. I mean why starting and ending frequencies have somewhat more amplitude. Isn't it supposed to increase from start to end.

[ConnerP · Accepted solution]

---

@rashp8 wrote:

Now I also got the result. Thanks a lot. Also can you tell me why chirp spectrum behaves like this. I mean why starting and ending frequencies have somewhat more amplitude. Isn't it supposed to increase from start to end.

---

It is because the signal is finite and does not repeat. There are discontinuities on both sides of the time signal which causes rippling in the frequency domain. It's not just the ends that are effect, the whole range has some small ripple. The only was around is to window the time domain function to make it continuous, but then the steps in the frequency domain won't be as steep. You have to pick your battles. Wikipedia explains it farely well.

[rashp8]

Also I want to add high frequency noise(MHz range) to my chirp signal (with frequency 270-330KHz) and then filter it out by using band pass filter. Can you suggest me how to add noise to any signal in labview?