Using Lossy Lines with Transfer Function

Hello,

I am working on a signal processing project where I am measuring the signal integrity of a signal. I essentially have a swept frequency cosine of the form:

t = 0:1/1e3:2;
y = chirp(t,0,1,250);

I want to model the effects of line loss at different frequencies. For this i am using a txlineDelayLossy object in this form where I can get the noise at a certain frequency:

dlossy = txlineDelayLossy("Z0",48,"LineLength",500e-3);
nf = noisefigure(dlossy,1e9);

The noisefigure element only returns an element at one frequency, 1e9, as specified below.

How would I be able to model this noisefigure at different frequencies through using a transfer function?

Best Regards

Answer 1

[Yash]

Hi Rohan,

As per my understanding, you're looking to estimate a transfer function in order to predict the noise figure at different frequencies.

To achieve this, I suggest you try out these steps:

1. Determine the noise figure at a few specific frequencies (as you have done for 1e9).

2. Calculate the noise power as follows:

      P_noise = P_signal / (10^(NF/10))

Here 'P_signal' is the signal power which can be calculated as:

      P_signal = rms(signal)^2

3. Set a sampling time 'Ts'. Create an 'iddata' object from your input output data using the 'iddata' function.

     data = iddata(output, input, Ts);

4. Open the System Identification Tool by typing 'systemIdentification' in the MATLAB command window.

5. Import your data, estimate the transfer function, and export it to the base workspace.

6. Now you can use the transfer function as needed. Note that this is the transfer function between P_signal and P_noise. Use them to calculate the 'nf'.

Another way to implement this(without transfer function) is to use interpolation. You can get various values of 'nf' for various frequencies and then interpolate it for the values you want.

To get more details about the interpolation, I suggest you refer to the following documentation here.

I hope this answer addresses your query.