MATLAB FFT vs Oscilloscope FFT

Hello,

I have an issue with what the oscilloscope's FFT is showing and using the same time domain data and performing an fft in MATLAB.

I'm measuring a signal on a Keysight DSO-S oscilloscope. It's set to 20 MSaS and 30Mpts. There are two 1 MHz signals on channel 1 and 4. I'm performing an two FFTs on the scope on each channel and saving the data. The scope's mat file contains both time domain data for channel 1 and 4, and FFT function data on both.

Now when plotting the scope's FFT and doing a MATLAB FFT on the time domain data I get two different pictures. Why is that?

For MATLAB I'm using a standard code found in fft() help:

Fs = 1./Channel_1.XInc; % Sampling frequency (20e6 Samples per second)

f = Fs*(0:(L/2))/L; % Frequency vector

dataOut = TDS2; % Time domain data loaded previously

DS0 = fft(dataOut);

DS0 = abs(DS0./L);

DS = DS0(1:L/2+1);

DS(2:end-1) = 2.*DS(2:end-1);

DSF = 10.*log10(DS.^2/50e-3); % Voltage converted to dBm with a 50 ohm load

plot(f, DSF)

For the scopes functions I'm getting the following graphs:

For MATLAB plot I'm getting something different:

Does anyone know what I'm doing wrong in MATLAB and why I'm losing all the data around 1 MHz?

[Lucy]

Hello, I encountered a similar problem recently and would like to know if you solved it? Thank you!

[Mathieu NOE]

hello

we don't know what the scope does .

is the data splitted , windowed , averaged (method ? ) .... this must be known before you can expect the matlab code to produce a similar output

for example what is the frequency resolution in both situations ?