QUESTION : FFT of a periodic pulsed RF from an oscilloscope

Hello all;

I have collected samples from a periodic pulsed RF signal with following properties:

RF : 50 MHz
PW : 1usec.
PRI : 3 usec.

Sampling frequency : 500 MHz
Collected Sample # : 1 Million

with a scope.

I have imported data to MATLAB to take FFT with the following code:

--
y=x50M500M1MSINDIS;
Fs = 500e6; % Sampling frequency
T = 1/Fs; % Sample time
L = length; % Length of signal
NFFT = 2^nextpow2(L);
Y = fft(y,NFFT)/L;
f = Fs/2*linspace(0,1,NFFT/2+1);
plot(f,2*abs(Y(1:NFFT/2+1)))
title('Single-Sided Amplitude Spectrum of y(t)')
xlabel('Frequency (Hz)')
ylabel('|Y(f)|')
--

I get the attached spectrum.



However, I expect to get some spectra around 50 MHz (the spectra of pulse with 3 usec period around 50 MHz to be exact).

What am I doing wrong?

Please help..

Hi,

Are you confident of your data?
What is y=x50M500M1MSINDIS; ?

I would plot y in order to check the data.
On the other hand, plotting |Y| in log vertical scale (dB) can reveal details that are not visible in linear scale.
Regards

Z

zorro said:

Hi,

Are you confident of your data?
What is y=x50M500M1MSINDIS; ?

I would plot y in order to check the data.
On the other hand, plotting |Y| in log vertical scale (dB) can reveal details that are not visible in linear scale.
Regards

Z

Click to expand...

Hello zorro,

y=x50M500M1MSINDIS means 50 MHz frequency sampled with 500 MHz for 1 Million samples with sin(x)-x reconstruction filter disabled. It is the name of MATLAB workspace file imported from a text file with same name. (MATLAB 2012)

Here are the matlab plots for y-time and 10logY-frequency. y-time seems correct to me..





What can you comment?

Please help..

Hi von Moltke

y should have samples both positive and negative. A pulse in y should have 50 cycles at 10 samples per cycle.
What we see seems to be a train of baseband pulses (not RF) plus noise. The obtained spectrum looks according to that.
Regards

Z

Hello zorro,

Thanks for your help. Your approach was correct. I manually generated a 50 MHz array and multiplied the square to get my intended signal.

I got the spectrum desired.

Thanks a lot again.