As per my understanding if we choose the default excitation signal in CST to excite an antenna/structure, the power stimulated spectrum is the normalized one at .5W through out the band. Default signal is 10-100MHz, gaussian, in this case.
Now when we choose a non-default signal :10-100MHz, gaussian, the power stimulated spectrum is the non-normalized one having peak around 4.5e-17W. According to me this value is absurd. If I take the fourier tranform of this non-default signal and then take the power of that spectrum, the output peak would be of the order of (e+3-4). This value seems unreasonably low.
I am attaching default/non-default excitation input signal and its power-stimulated spectrum in default and non-default case.
Kindly help me to explain this. What am I missing ?
Not an easy question to answer - my first try was wrong - see below.
I don't think it is far out. The time pulse is about 1 watt for 10 ns so is about 10 nJ
In the frequency domain the power is 4.5e-17W over about 40 MHz, which is about 180e-11 J or about 1.8 nJ. There does seem as if there may be a factor of 5 somewhere in there which may be my error, but the orders of magnitude are not absurd.
I think my answer above is wrong.
The time profile can have instantaneous power, and the integral of this over time gives the total energy in the system, so my top calculation is roughly right.
however, I think that in the spectrum, the vertical axis should not be power. It needs to be Joules per Hz, as would be measured by a narrow-band filter hit by a pulse. The integral of Joules per Hz over the frequency range is then Joules. In order to compare it with the time domain Energy you need to know what the vertical axis is. If I was measuring it using an analogue spectrum analyser I'd need to know the pulse desensitisation factor for the filters used and the pulse shape. I'm not sure how that carries through into the CST built-in maths.
After more thought:
The energy in the time domain is about 10nJ
If this is treated as periodic, with say 1 sec repeat, then the mean power is 10nW
If it is periodic with a period of 1 sec, then there will be a line spectrum with lines at 1 Hz separation, and each line having power. The sum of all the lines is the total power. With 1 sec repeat, the power is about 2 nJ. I think the two powers can be made the same by adjusting the fictional repeat time.
If the repeat is 2 sec then the power in the impulse train is 5 nJ and the power in the spectrum is about 4 nJ.
As the height of the spectrum is measured in power, I think that a periodicity has to have been assumed somewhere in the calculation. If not, the height would need to be measured in energy.
I'm not sure if a factor of 2pi might come in somewhere, which is close to the difference between 2 nJ and 10 nJ! 1s default period would be a natural value to have happened.
- Badges
- Science topic