Basic question Time Domain vs Frequency Domain

[Mica188]

frequency sinus fourier basic

Hi all,

I'm confuse between Time Domain and Frequency Domain solvers. Well in my understanding, Time Domain is useful when you have an excitation voltage or current which is not a simple sinus. Because, if you do Fast Fourier Transform on a delta impulse ,let's take for example, it will give you frequencies from - ∞ to + ∞ , so, this is not praticle if you have a Frequency Domain solver . Please correct me if I'm wrong. In other words, the frequency domain solver is limited in excitation pulse shape (input voltage or current)? Please, can someone provide more info on it.

Thanks,

Mica

 

[eirp]

The key point comes from Fourier transform theory

For harmonic signals following equation is valid:

\[\frac{d}{dt}\rightarrow j\omega\]

 

[radiohead]

What you said is correct; only you forgot one important thing. There is no such thing as delta impulses in real design. The best pulse you can get is a few ps width, depending in your technology. So your frequency domain representation for this IS limited.

Another thing: suppose you could make a delta impulse. Any transistor in the signal chain would turn it into a real pulse, like I discussed before.

The thing to keep in mind is that both simulations are important, but have different application domains.

Time domain simulation = baseband processing (eg optical receiver design)
Frequency domain simulation = passband processing (eg modulated carrier processing, mixing, oscillator)

Both are complementary and needed!

 

[nicleo]

In other words, the frequency domain solver is limited in excitation pulse shape (input voltage or current)?

Not really. In switching power supply design, we might wish to know the harmonics or THD of the input or output current of the power supply. To get the harmonics or THD, we need to perform Fourier Analysis (in frequency domain) to find the harmonics of the current waveform, which in some cases, are pulses, discontinuous, or non-sinewave.

 

[Mica188]

Thanks for the useful info.

Let's say I have a simple dipole :

1) If I have a pluse in time domain and I did a FFT , and gives me two freqencies. Then, if I have a frequency domain solver, I need to simulate these two frequencies for my antenna anaylsis ? And assuming each of the frequency is a sinus distribution?

2) The problem in 1) but now I have a Time Domain solver, which I can directly put the pulse in the sofware and simulate? So, I don't have to do a FFT ?

But the results should be the same in 1) and 2) ?

Regards,

Mica

 

[radiohead]

You should revise your information about the FFT of the pulse. A waveform with only two harmonics is repetitive and will hardly resemble a pulse! It's a sinusoid with a second or third harmonic mixed in it!