Discussion Forum

Hello Rishad,
I noticed that the "Attach" button seems to be misbehaving today on Firefox, but works on IE and Chrome. If you tried on Firefox earlier, can you try using one of those browsers to upload your file?
Sorry about this, I have reported the problem to our web team.
Best,
Jeff

Yes. I have been using Firefox, that's why. Thanks Jeff ! I am trying to state the my problem again and also uploading my file.

I have attached my image file here. I would like to find the change of the e-field with respect to time along that certain path (red line). Please kindly let me how can I do that.

Thanks,
Rishad

Rishad,

again the question about a temporal plot in a frequency domain study. But maybe I found out what you want.
You can make a line plot along this red line indicated in your picture of the E-field component you are interested in. This will give you a field profile E = E(x), right?
Now divide the x-coordinate by the speed of light. This will result in a profile E = E(t). The physical meaning of this is then the temporal profile of a certain phase of the wavefront along this path. Of course this depends on the phase of the solution you are looking at. You don't get an unambiguous result.
So far this is the only way I can think of, that refers a spatial path to a function of time in a frequency domain study.

Cheers
Edgar




--
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com

Thanks a lot Sir for your reply. I was looking forward for someone to response to my problem.

Well, If I do 1D plot>line plot along this red line, comsol generates the e-field distribution with respect to Arc length (I assume that is what you were saying E=E(x), right ? ). I have also attached the E(x) plot that I have obtained.

But if I now divide x- coordinate (arc length) by model frequency, then how it becomes E(t)? I am sorry but I didn't understand you point here. How the time is defined here in this case ?

Please kindly let me know.

Thanks,
Rishad

Sorry, my mistake. You have to divide by the speed of light, not frequency. I edited the other post too.

--
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com

Thank you Sir for you concern. I will try to simulate as you told and will try to see how it works. I will post the result here.

Thanks,
Rishad

Hm, you don't need a new simulation run for that. It is all in the post processing. You only have to replace x by x/c in the line plot.
This is really trivial, so I am still not sure if it is what you have in mind.

--
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com

Sir, yes , I did the post processing in my simulation as you said but I was expecting something different than the attached figure (22).

I was expecting a time varying e-field which is bidirectional in nature. Time very e-field means in this case- e-field in time domain. However, I have e-field distribution in f domain. How can I plot e-field from in t domain from that ?

I have also attached kind of similar plot which I was expecting. (attached 'expected', please ignore the scale in that case.) Please let me know what you think of it.

Regards,
Rishad

I want to insert the time step size and the final time for simulation. Like for example, from t=0 to t=100 s, how the e-field is varying with time along that path (red line).

However, by dividing the x coordinate /c_const doesn't make much difference in the plot. It's almost similar plot that I obtained for E-field Vs arc length, right? Moreover, there is also negative time in the fig when I divide it by c_const, how that can be explain , Sir ?

Sure, as I said it is trival. It is only a rescaling of the x-axis. And nothing else is possible in a frequency domain study. There is no explicit time in frequency domain. Period!


--
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com

Hello again Sir , let me discuss the real scenario here in more detail.

1. The E(x) plot that I have obtained is in frequency domain, which means at a one single frequency E(x) is a phasor or sinusoidal signal at every specific point along x with different amplitudes.

2. However, I want to see E(x) in time domain, which means it will show me how E(x) changing with x at a specific time t=0s, 5s, 10s and so on. This is essentially a 1D plot of E(x) Vs x at different time. I hope, I made my question clear to you.

Would you please suggest me how can I simulate this time domain analysis ? Is there any useful approach ? Please kindly, let me know. I am enlisting your expertise.

Thanks,
Rishad

Hi Rishad,

What is the size of your antenna?
At which frequency do you analyse it?
If your frequency e.g 100MHz than you should analyse it in the ns range and not 100s.

As far as I know if you want to transfer a signal from the frequency to the time domain you need a inverse Fourier transformation for it. I didn't do it for 2D signals and so I can't tell you how.

If you need the change of the E-filed during 2*pi you can do following. The E-filed has a phase change along the red line in respect to the frequency. E_max * cos(2*PI*l/LAMDA+Phi_0)
E_max: Maximum E-filed at this point
l: distance from source along the red line
LAMDA: Wavelength c/f
Phi_0: undefined phase

Because your antenna has a size compare to the wavelength the phase will change with the distance from the source with 2*PI*l/LAMDA.
Phi_0 can be any value and if you change it from 0 to 2*PI you will see how the E-filed will move along the structure.

To calculate it you extract the complex value of E-filed along the red line. Add Phi_0 from 0 to 2*PI with the numbers of step you like and plot the real part of the E-filed (2D-plot).

I hope it is correct. I did it long time ago and I don't have any more any documentation.

Regards,
Stefan

Hi,

My antenna dimension is in micro range and I am dealing with THz. In that case, I think the time step should be Femto sec, right ? Correct me, if necessary.

Yes. E-field has a phase change along the red line. Can you please suggest me from where I can change the phase of E-field so that I can observe the change of E-field with respect to time? Let me know Stefan.


Thanks,
Rishad