Cascade filters and amplify

Hi

I don’t have much experience in MATLAB but I have a circuit that amplifies low frequency breathing it consists of four op-amps each amplify by 10 and each have a second order bandpass set to .1Hz to .4Hz.

I have now used one op-amp as above and a 24bit ADC sampling at 100Hz and my number of samples is 6000 (60 Seconds of data)

I have managed to bring the data in to MATLAB (please see attached) but I am struggling to implement the filters and gain could somebody please assist me.

Many thanks in anticipation

Kind Regards

David

Answer 1

[Mathieu NOE]

Here you are my friend

the demo is done on a second order bandpass Butterworth filter ; you ca simply increase the order to match was was done in your analog filters

hope it helps

all the best ,

clearvars

load('Raw_Data.mat');

Fs = sample_rate;
samples = number_of_samples;
x = raw_ad_data_sine;
dt = 1/Fs;
t = (0:samples-1)'*dt;

% IIr filter : bandpass butterworth
order = 2;
[numd,dend] = butter(order,2/Fs*[0.1 0.4]);

figure(1)
dbode(numd,dend,dt); % bode plot

y = filtfilt(numd,dend,x);  % Zero-phase forward and reverse digital IIR filtering.

figure(2) % time plot
plot(t,x,t,y);
legend('input signal','output signal');

[David Jones]

Hi Mathieu

I hope you are keeping well, thank for responding again, I have tried the code works fine exept if I try to change the order to 4 I get this error

Warning: Matrix is close to singular or badly scaled. Results may be inaccurate. RCOND = 4.047604e-17.

> In filtfilt>getCoeffsAndInitialConditions (line 246)

In filtfilt>efiltfilt (line 129)

In filtfilt (line 89)

In BPTest (line 12)

Can you help please

Kind Regards

David

[Mathieu NOE]

yes , that sometimes happen with higher order filter design when the sampling rate is fairly high vs the cut off frequencies and the request order is high

in your case , it will work better for higher order filters if yu could reduce Fs by downsampling your data by at least factor 2 (best is between 5 to 10)

here a demo , which uses

decimate

to reduce signal sampling rate

clearvars

load('Raw_Data.mat');

bits = 24; % ADC resolution
scale_adc = 1/2^(bits-1);
raw_ad_data_sine = detrend(raw_ad_data_sine.*scale_adc);

% signal decimation (sampling rate reduction)
R = 5; % decimation factor
x = decimate(raw_ad_data_sine,R);

Fs = sample_rate/R;
samples = length(x);

dt = 1/Fs;
t = (0:samples-1)'*dt;

% IIr filter : bandpass butterworth
order = 4;
[numd,dend] = butter(order,2/Fs*[0.1 0.4]);
%   If Wn is a two-element vector, Wn = [W1 W2], BUTTER returns an
%   order 2N bandpass filter with passband  W1 < W < W2.

figure(1)
dbode(numd,dend,dt); % bode plot

y = filtfilt(numd,dend,x);  % Zero-phase forward and reverse digital IIR filtering.
% y = filter(numd,dend,x);  % forward digital IIR filtering.

figure(2) % time plot
plot(t,x,t,y);
legend('input signal','output signal');

this is the plot it generates

[David Jones]

Thanks for all your great help Mathieu

[Mathieu NOE]

My pleasure