FDTD, the need for FFT when computing the power spectrum

[roldao]

Source in FDTD

Hello People.

I have a simple question.

I´m using FDTD simulations to calculate the power of the signal on a scenario location.

I´m using a soft source that is a sinusoid modulated with a gaussian pulse with a determinate frequency.

I apply the source in the time domain and never do the fft after calculate the electric field and obviously never do the ifft to calculate the power.

Am I doing the things right? or I need to do a fft of the electric field and then when I calculate the power, I need to do a ifft before, that is what makes sense on a communication system but on this kind of simulation I dont know?

I´m confused on this matter.

Thanks for your precious help.

Joao Roldao

 

[rrumpf]

Re: Source in FDTD

There are many ways to do what I think you are saying, but the following may be the easiest to follow. Implement your source however you wish. Record your source as your incorporate it. Record your response wherever you wish. FFT both sets of data. Divide the response FFT by the source FFT to normalize it. This should give you the response you are looking for.

The reason you do this is because the power spectrum is not uniform. If you don't normalize in this way, fluctuations in the spectral response cannot be distinguished from dips and valleys in the source spectrum.

-Tip