@Mary Claire — The problem with your data is that they are really noisy (that appears to be thermal or 
noise) that probably originates in the instrumentation. There does not appear to be any specific trend in the time-domain data, so detrending or highpass or bandpass filtering (to remove baseline drift, and optionally high-frequency noise) will not improve the result. The noise is broadband, and cannot be removed with a frequency-selective filter. A Savitzky-Golay filter occasionally works to eliminate broadband noise, however it does not with these data, at least when I have tried it.
noise) that probably originates in the instrumentation. There does not appear to be any specific trend in the time-domain data, so detrending or highpass or bandpass filtering (to remove baseline drift, and optionally high-frequency noise) will not improve the result. The noise is broadband, and cannot be removed with a frequency-selective filter. A Savitzky-Golay filter occasionally works to eliminate broadband noise, however it does not with these data, at least when I have tried it. T1 = readtable('RF68CENE4.6M05305mins.csv');
T1.Var2 = str2double(T1.Var2)
figure
plot(T1.Var1, T1.Var2)
grid
figure
plot(T1.Var1, T1.Var2)
grid
xlim([T1.Var1(1) T1.Var1(500)])
% dt = [0; diff(T1.Var1)];
% dt.Format = 'hh:mm:ss.SSSSSS'
% dt_stats = [mean(dt);std(dt);mode(dt)]
Ts = 0.01;
Fs = 1/Ts;
Fn = Fs/2;
L = size(T1,1);
NFFT = 2^nextpow2(L)
FTVar2 = fft((T1.Var2-mean(T1.Var2)).*hann(L), NFFT)/L;
Fv = linspace(0, 1, NFFT/2+1)*Fn;
Iv = 1:numel(Fv);
envu = envelope(abs(FTVar2(Iv))*2, 100, 'peak');
figure
plot(Fv, abs(FTVar2(Iv))*2, 'DisplayName','Fourier Transform')
hold on
plot(Fv, envu, '-r', 'DisplayName','Upper Envelope')
hold off
grid
xlabel('Frequency (Hz)')
ylabel('Magnitude (Units)')
legend('Location','best')
You need to examine the instrtumentation to determine where the noise is originating, and then do what you can to minimise it.
.
