Difficulty in Calculating Signal Frequency with an Offset

So I wrote a Matlab program that calculates and plots some key values of a signal. The signal is stored in a .csv file.

The problem is that the calculation of the frequency only works for signals that don't have an offset.

I would appreciate any kind of help since I can't seem to figure it out on my own.

Matlab program:

% Reading .csv

data = readmatrix('Sinus.csv');

% Extract time and voltage columns

time = data(:, 1);

voltage = data(:, 2);

% Calculate peak value

peak_value = max(abs(voltage));

% Calculate arithmetic mean in mV

mean_value = mean(voltage) * 1000; % Conversion from V to mV

% Calculate rectified value

rectified_value = mean(abs(voltage));

% Calculate RMS (Root Mean Square) value

rms_value = rms(voltage);

% Calculate form factor

form_factor = rms_value / rectified_value;

% Calculate crest factor

crest_factor = peak_value / rms_value;

% Calculate sampling frequency (Assumption: uniformly sampled signal)

sampling_frequency = 1 / (time(2) - time(1));

% Calculate FFT (Fast Fourier Transform) of the signal

N = length(voltage);

Y = fft(voltage);

f = sampling_frequency*(0:(N/2))/N;

% Find the index of the maximum in the frequency domain

[~, max_index] = max(abs(Y(1:N/2+1)));

% Extract frequency at the maximum

frequency = f(max_index);

% Calculate period duration in ms

period_duration = 1/frequency * 1000; % Conversion from s to ms

Thank you for your time!

Answer 1

[Paul]

Hi Raphael,

The first point in the FFT is always the sum of the elements of the input. When the signal includes the offset, the abs of the sum of the elements is larger than the peak of the FFT at the signal frequency.

% Reading .csv
data = readmatrix('SinusOff.csv');
% Extract time and voltage columns
time = data(:, 1);
voltage = data(:, 2);
figure
plot(time,voltage),grid

% Calculate peak value
peak_value = max(abs(voltage));
% Calculate arithmetic mean in mV
mean_value = mean(voltage) * 1000; % Conversion from V to mV
% Calculate rectified value
rectified_value = mean(abs(voltage));
% Calculate RMS (Root Mean Square) value
rms_value = rms(voltage);
% Calculate form factor
form_factor = rms_value / rectified_value;
% Calculate crest factor
crest_factor = peak_value / rms_value;
% Calculate sampling frequency (Assumption: uniformly sampled signal)
sampling_frequency = 1 / (time(2) - time(1));
% Calculate FFT (Fast Fourier Transform) of the signal
N = length(voltage);
Y = fft(voltage);
f = sampling_frequency*(0:(N/2))/N;
figure
plot(f,abs(Y(1:numel(f))),'-o')
xlim([0 1e4])

% Find the index of the maximum in the frequency domain
[~, max_index] = max(abs(Y(1:N/2+1)));
% Extract frequency at the maximum
frequency = f(max_index);
% Calculate period duration in ms
period_duration = 1/frequency * 1000 % Conversion from s to ms

period_duration = Inf

You can ignore the first point in the FFT or you can subtract the mean voltage before taking the fft

Y = fft(voltage-mean(voltage));

[Raphael Brugger]

Thank you for taking the time and answering to my problem. I tried to substract the mean(voltage) from the voltage, and the values fit. But if I do that the plot doesn't show a signal at all. I tried to work around with temp variables but I can't quite figure it out. An example of what I tried is shown below.

% Reading .csv

data = readmatrix('SinusOff.csv');

% Extract time and voltage columns

time = data(:, 1);

voltage = data(:, 2);

% Calculate peak value

peak_value = max(abs(voltage));

% Calculate arithmetic mean in mV

mean_value = mean(voltage) * 1000; % Conversion from V to mV

% Calculate rectified value

rectified_value = mean(abs(voltage));

% Calculate RMS (Root Mean Square) value

rms_value = rms(voltage);

% Calculate form factor

form_factor = rms_value / rectified_value;

% Calculate crest factor

crest_factor = peak_value / rms_value;

% Calculate sampling frequency (Assumption: uniformly sampled signal)

sampling_frequency = 1 / (time(2) - time(1));

% Calculate FFT (Fast Fourier Transform) of the signal

N = length(voltage-mean(voltage));

Y = fft(voltage-mean(voltage));

f = sampling_frequency*(0:(N/2))/N;

% Find the index of the maximum in the frequency domain

[~, max_index] = max(abs(Y(1:N/2+1)));

% Extract frequency at the maximum

frequency = f(max_index);

% Calculate period duration in ms

period_duration = 1/frequency * 1000; % Conversion from s to ms

% Output the calculated values

fprintf('Peak value: %.4f V\n', peak_value);

fprintf('Arithmetic mean: %.4f mV\n', mean_value);

fprintf('Rectified value: %.4f V\n', rectified_value);

fprintf('RMS value: %.4f V\n', rms_value);

fprintf('Form factor: %.4f\n', form_factor);

fprintf('Crest factor: %.4f\n', crest_factor);

fprintf('Period duration: %.4f ms\n', period_duration);

fprintf('Frequency: %.4f Hz\n', frequency);

% Create the plot of the input signal

figure;

plot(time, voltage, 'DisplayName', 'Signal'); % Added DisplayName for legend

hold on;

% Plot the calculated values

plot([time(1), time(end)], [peak_value, peak_value], '--', 'DisplayName', 'Peak value');

plot([time(1), time(end)], [mean_value, mean_value], '--', 'DisplayName', 'Arithmetic mean');

plot([time(1), time(end)], [rectified_value, rectified_value], '--', 'DisplayName', 'Rectified value');

plot([time(1), time(end)], [rms_value, rms_value], '--', 'DisplayName', 'RMS value');

legend('show');

title('Signal with Calculated Values');

xlabel('Time (s)');

ylabel('Voltage (V)');

grid on;

[Paul]

The mean is in mV, but the rest of the information on the plot is in V.

[Raphael Brugger]

Oh. Thank you!