Spectral Measurement using NI SCOPE

[Sahith24426]

https://youtu.be/xMWIATas_e0?si=xRIc8v5bshUqAgyy

This video is about doing spectral measurement using DAQ Assist. I need the same with NI Scope. I have this issue where fft peak is spiking at 0Hz irrespective of the signal I'm generating from function generator.

[NJKirchner]

Without seeing what each of your actual channel data is, there are only certain things that can be recommended.

First off, if you have any DC offset in your data (which really will be almost always due to never being 'PERFECT') you'll have some spectral peak at DC. But let's presume that's not the case.

The other thing, is that you're subtracting both waveforms, which is likely to cause issues, as you're subtracting each element of each cluster, which is going to cause TONS of issues. If you're trying to subtract the amplitudes, then fetch from SCOPE as a WDT and then run that through the subtract function (it'll be much smarter about it). 

Give that a go, share images of both time domain signals, their difference and the spectral output of all 3.
what you're looking to do is indeed simple, you've just likely mixed up the way to do it

[dsb@NI]

What Norm said.

Plus:
Set dt when you build the waveform. The default dt = 1 s
Explain mystery scaling (FFT / 900)*150000000*0.001
Attach your VI, preferably with some typical data saved (as default) in the waveform graphs.

[Sahith24426]

Thanks for the response NJKirchner.

plot0 is a linearly swept sinusoid and plot1 is slightly delayed version of it (Linear Freq. Sweep 100 to 1KHz). Followed by the subtracted part and FFT. Waveforms are moving continuously.
Additionally when I tried connecting the spectral analysis block with clean sinusoid of 1KHz, It still peaks at 0Hz.

[Sahith24426]

Hi dsb@NI, Thanks for the response.
The Idea is to obtain the fft of beat signal in time domain and get info on range measurement (Lidar). That mysterious scaling block is the formula setup and comes after achieving fft.
This gives another doubt, If it is a single peak I can just wire it out, what if fft has multiple peaks.

[GerdW]

Hi Sahith,

---

@Sahith24426 wrote:

Additionally when I tried connecting the spectral analysis block with clean sinusoid of 1KHz, It still peaks at 0Hz.

---

Did you understand the part about "at 0Hz is the DC part (aka offset) of the signal"?

How does the FFT look like when you switch to logarithmic scaling of the Y axis (as is common for spectra)?

When you follow the request of attaching a VI (with data included as default values!) then please downconvert for LV2019…

[jjohn1]

I might be missing something, but if you are trying to beat two signals together would not you multiply them together instead of subtracting them?   A receiver would mix, not sum.

[dsb@NI]

---

@jjohn1 wrote:

I might be missing something, but if you are trying to beat two signals together would not you multiply them together instead of subtracting them?   A receiver would mix, not sum.

---

beat frequency is evident in the total wave (sum of interfering waves)

[dsb@NI]

---

@Sahith24426 wrote:

Hi dsb@NI, Thanks for the response.
The Idea is to obtain the fft of beat signal in time domain and get info on range measurement (Lidar). That mysterious scaling block is the formula setup and comes after achieving fft.
This gives another doubt, If it is a single peak I can just wire it out, what if fft has multiple peaks.

---

Since none of that scaling affects the frequencies; let's ignore for now.

What are you going to do when FFT has multiple peaks? That sounds like some more research for you.

[dsb@NI]

---

@Sahith24426 wrote:

Thanks for the response NJKirchner.

plot0 is a linearly swept sinusoid and plot1 is slightly delayed version of it (Linear Freq. Sweep 100 to 1KHz). Followed by the subtracted part and FFT. Waveforms are moving continuously.
Additionally when I tried connecting the spectral analysis block with clean sinusoid of 1KHz, It still peaks at 0Hz.

---

In the FFT, if you want to distinguish signal peaks from DC (0 Hz), you must process longer blocks of time waveforms that include more cycles of the signal.

The following are all FFTs (window = None) of a 1 kHz sinewave: