Forums

Network analysis using chirp signal.

Started by Dimitar Penev June 1, 2004
Hello All,

I am creating the simple network analyzer using the general purpose 
DAQ board.

Currently I am testing the system using linear chirp (tone with sweeping  
frequency) exitation signal x. Then I correlate the output of the system y with 
x and got estimation of the impulse system response h. FFT it to get frequeny 
response H

The procedure works because the autocorrelation of x is almost delta pulse.

I am testing the procedure for several systems and it appears that levels 
below -100 dB of the H are not estimated correctly. In other words I have abou 
100 dB dynamic range. 

I made compensation for the fact that autocorrelation of x is not ideal delta 
pulse but stil can not inprove the acuracy. 

I can use the white noise as stimul and will get much below -100 db but still 
think that the chirp should be better choise.

Opinions are well come.
penev
Hi Penev!

I'm also interresting in a similar problem. In my case I are analysing
quatisfication effects for different types of IIR filter structures. I have
also used a chirp signal to estimate the frequency response of a filter I
implement in lattice, WDF and other structures. When I applying the chirp
signal as an input to the filter and compute the frequency function for a
simple low pass filter via FFT  I got a moderated dc value. Even if I add a
separate constant to the chirp signal to compesate as you for non-ideal
delta function.

I have also notice the same results as you describe with the case of white
noise signal as a stimuli but I can't understand it really. So please let me
know if you or some else have a solution to this problem.

Best regards, Michael N.

"Dimitar Penev" <dpenev@yahoo.com> skrev i meddelandet
news:74727cb9.0406010548.68481a8b@posting.google.com...
> Hello All, > > I am creating the simple network analyzer using the general purpose > DAQ board. > > Currently I am testing the system using linear chirp (tone with sweeping > frequency) exitation signal x. Then I correlate the output of the system y
with
> x and got estimation of the impulse system response h. FFT it to get
frequeny
> response H > > The procedure works because the autocorrelation of x is almost delta
pulse.
> > I am testing the procedure for several systems and it appears that levels > below -100 dB of the H are not estimated correctly. In other words I have
abou
> 100 dB dynamic range. > > I made compensation for the fact that autocorrelation of x is not ideal
delta
> pulse but stil can not inprove the acuracy. > > I can use the white noise as stimul and will get much below -100 db but
still
> think that the chirp should be better choise. > > Opinions are well come. > penev
Hi Michael,

Yes it seems that the limited dynamic range in my case was due to my
data are time limited (ract window). I applied window to my chirp
signal and now I can estimate the notch in the noch filter up to -200
dB (using good window)

Hmm for low pass filter the DC value should pass troight... Probably
your filter is not LP?

BTW: what is quatisfication  effects? I have not heard about it.

Best Regards
Dimitar Penev 

"Michael Numminen" <michael.numminen@comhem.se> wrote in message news:<WO4vc.986$dx3.9865@newsb.telia.net>...
> Hi Penev! > > I'm also interresting in a similar problem. In my case I are analysing > quatisfication effects for different types of IIR filter structures. I have > also used a chirp signal to estimate the frequency response of a filter I > implement in lattice, WDF and other structures. When I applying the chirp > signal as an input to the filter and compute the frequency function for a > simple low pass filter via FFT I got a moderated dc value. Even if I add a > separate constant to the chirp signal to compesate as you for non-ideal > delta function. > > I have also notice the same results as you describe with the case of white > noise signal as a stimuli but I can't understand it really. So please let me > know if you or some else have a solution to this problem. > > Best regards, Michael N. > > "Dimitar Penev" <dpenev@yahoo.com> skrev i meddelandet > news:74727cb9.0406010548.68481a8b@posting.google.com... > > Hello All, > > > > I am creating the simple network analyzer using the general purpose > > DAQ board. > > > > Currently I am testing the system using linear chirp (tone with sweeping > > frequency) exitation signal x. Then I correlate the output of the system y > with > > x and got estimation of the impulse system response h. FFT it to get > frequeny > > response H > > > > The procedure works because the autocorrelation of x is almost delta > pulse. > > > > I am testing the procedure for several systems and it appears that levels > > below -100 dB of the H are not estimated correctly. In other words I have > abou > > 100 dB dynamic range. > > > > I made compensation for the fact that autocorrelation of x is not ideal > delta > > pulse but stil can not inprove the acuracy. > > > > I can use the white noise as stimul and will get much below -100 db but > still > > think that the chirp should be better choise. > > > > Opinions are well come. > > penev
Hi againg, 

As I have written in the previous message I put window on the input
chirp signal and the dynamic range become more than 200 dB. The
problem is that windowing the excitation chirp signal means reduction
the lowest and higest frequency components. It means more noise in the
network estimation at the low frequency and at the frequency close to
the Nyquist rate.

Some ideas?
 
dora

dpenev@yahoo.com (Dimitar Penev) wrote in message news:<74727cb9.0406012330.34bb4571@posting.google.com>...
> Hi Michael, > > Yes it seems that the limited dynamic range in my case was due to my > data are time limited (ract window). I applied window to my chirp > signal and now I can estimate the notch in the noch filter up to -200 > dB (using good window) > > Hmm for low pass filter the DC value should pass troight... Probably > your filter is not LP? > > BTW: what is quatisfication effects? I have not heard about it. > > Best Regards > Dimitar Penev > > "Michael Numminen" <michael.numminen@comhem.se> wrote in message news:<WO4vc.986$dx3.9865@newsb.telia.net>... > > Hi Penev! > > > > I'm also interresting in a similar problem. In my case I are analysing > > quatisfication effects for different types of IIR filter structures. I have > > also used a chirp signal to estimate the frequency response of a filter I > > implement in lattice, WDF and other structures. When I applying the chirp > > signal as an input to the filter and compute the frequency function for a > > simple low pass filter via FFT I got a moderated dc value. Even if I add a > > separate constant to the chirp signal to compesate as you for non-ideal > > delta function. > > > > I have also notice the same results as you describe with the case of white > > noise signal as a stimuli but I can't understand it really. So please let me > > know if you or some else have a solution to this problem. > > > > Best regards, Michael N. > > > > "Dimitar Penev" <dpenev@yahoo.com> skrev i meddelandet > > news:74727cb9.0406010548.68481a8b@posting.google.com... > > > Hello All, > > > > > > I am creating the simple network analyzer using the general purpose > > > DAQ board. > > > > > > Currently I am testing the system using linear chirp (tone with sweeping > > > frequency) exitation signal x. Then I correlate the output of the system y > with > > > x and got estimation of the impulse system response h. FFT it to get > frequeny > > > response H > > > > > > The procedure works because the autocorrelation of x is almost delta > pulse. > > > > > > I am testing the procedure for several systems and it appears that levels > > > below -100 dB of the H are not estimated correctly. In other words I have > abou > > > 100 dB dynamic range. > > > > > > I made compensation for the fact that autocorrelation of x is not ideal > delta > > > pulse but stil can not inprove the acuracy. > > > > > > I can use the white noise as stimul and will get much below -100 db but > still > > > think that the chirp should be better choise. > > > > > > Opinions are well come. > > > penev