I have a problem at hand that requires urgent attention. Thing is I’ve never used Simulink before and right now its required for me to FFT a continuous signal output from my vehicle model.

Thing is it is a continuous signal and I am facing problems trying to even implement the FFT. Even when there is no errors there would not be any output. So yes i’m desperate for help.

thanks.
XY

I think it would be useful to have a matlab function that directly computes time domain response from s-domain system transfer function available only as numerical data:
G(s=jw). While I realize there are subtleties in the FFT and iFFT scaling I’d rather not be bothered by these issues, since my focus is on dyanmics rather than signal procressing techniques and algorithms.

Amit.

Thanks for the naswer regarding Saurabh’s question. I’am having the same problem. I don’t have the Signal Processing Blockset install in my Matlab. I wanted to write a script to generate it form my output signal. As a newbie with Matlab I don’t really know how to introce the time and the different part of the script. This is what I have writen but it doesn’t work.
dt = 1/100;
T = 100;
t = (0:dt:T-dt);

x1 = voltage2.signals.values(:,1);
y = fft(x1)/length(x1)*2;
m = abs(y);
m(1)= m(1)/2;
f = (0:length(y)-1)*100/length(y);
plot(f,20*log10(m)),

hold on
ylabel(‘Abs. Magnitude dB’),

x2 = voltage3.signals.values(:,1);
y = fft(x2)/length(x2)*2;
m = abs(y);
m(1)= m(1)/2;
f = (0:length(y)-1)*100/length(y);
plot(f,20*log10(m),’-.’),
axis ([0 max(f)/2 -60 10])
hold off
h = legend(‘Uu’,'Uu0′);

It’s the ouput voltage of a 2 level 4 quadrant IGBT circuit.
X1 is the line to virtual neutral of the “Y” connected out put load voltage and X2 is the line to ground load voltage.
Please could u help me.
Many thanks.

Regards,
Chai

k=1;
p=-1;
dt=1e-2;
tstop=30.01;

%Time period
t=0:dt:tstop;

%Apply  Freq Response
freq=0:1/tstop:1/dt;
jw=j*2*pi*freq;
w=2*pi*freq;
for d=1:length(freq)
    g(d)=k/(jw(d)-p);
end
% %double sided response
n=length(g)
m=fliplr(conj(g(2:((n)/2+1))));
mm=[g(1:(n)/2) m]*n;
 x=real(ifft(mm))/tstop;
plot(t,x)
hold on
 %Known Time domain result from Laplace
 for d=1:length(t)
   h(d)=exp(p*t(d));
 end
plot(t,h,'r')

How would I handle the maximum frequency point with an even number of points? Is my scaling of my result correct? I hope that you can help me clear the issues behind ifft/fft. Thanks once again,

Dragan

http://www.mathworks.com/access/helpdesk/help/toolbox/dspblks/ug/f13-54010.html#f13-23126

Thanks again.

This example is quite good to know.