How to choose the sampling frequency of a FFT based on my signal parameters ?

I have a signal coming from an expermient, and I want to extract the frequencies from it. I made a simple method that extract the peaks and divide the number of the peaks by the duration of the whole signal, but I want to use the FFT method also.

I made the code and I have my FFT, but I don't know how to define the samplig frequency. I noticed that 100 Hz seems to be a good value but I don't want to choose it randomly, so my question is : is there any trick to determine the sampling frequency based on the signal parameters ?

I know there is the Nyquist theorem that says sampling frequency = 2x (max frequency-min frequency), but in this signal I don't really know what would be the min and the max.

For information here are the parameters of the signal I know :

• Duration
• Numer of points = 10 points/second

Here is a graph of the FFT with 100 Hz of sampling rate with the frequency dound with the simple peak method :

If I change the sampling frequency I see a shift betwen the two methods.

Here is my code if needed :

%FFT method

Fs = 100;             % Sampling frequency                    
T = 1/Fs;             % Sampling period       
L = Length/10;        % Length of signal in s
t = (0:L-1)*T;        % Time vector

% x frequency
FFTx=fft(x,L); 
P2 = abs(FFTx/L);
P1 = P2(1:L/2+1);
P1(2:end-1) = 2*P1(2:end-1);
f = Fs*(0:(L/2))/L;

figure('Name','Frequency spectrums');
tiledlayout(2,1);    
nexttile;
hold on
plot(f,P1);
line([Peak_fx Peak_fx], [0 max(P1)],'Color','red');
hold off
legend('FFT method','Peak method');

The variable "Length" is the duration in 10th seconds, the variable "x" is the data from my signal and the "Peak_fx" variable is the frequency coming from the peak method.

[Mathieu NOE]

hello

sorry but as far as I can see, you know the time length of your signal (60 s ) and you know how many samples are in your data; so your sampling frequency is determined by Fs = 1/dt with dt = (time(end)-time(1)/(samples -1);

[Hugo Fotia]

I already tried this but it gives me sometimes wrong frequencies as far as I can judge with my noisy signal, but I think I will keep this equation because it seems the most accurate for now. Thanks for your answer !

[Mathieu NOE]

this sounds a bit strange to me ... would you accept to share your data file ?