Discussion Forum

In even simpler terms,

I am having trouble with reflections at the air-glass interface in COMSOL...!

cheers

I had similar experience. I sent a plane wave from a rare medium to dense medium and vice versa to compute the reflection phase jump. In case of the former, phase should be pi while for the latter it has to be 0. However, my computed values were around 3.02 and 0.55 radians respectively.

Moreover, modulation patters were clearly visible in the second medium indicating reflection from scattering boundary which is supposed to be perfectly non-reflecting for plane waves.

Anyhow, fortunately, the specific problem I solve involve overlap integrals between orthogonal modes and therefore spurious reflections are reduced in effect down to tolerable limits.

I would suggest you to take up this case directly with comsol support. In case you find out something useful, please share with everyone here :)

Shakeeb,

Thanks for your post,

Did you find that the reflection and transmission coefficients did not match as well?

Also did you turn of the sources in the other mediums? I have always been doing this but now I seem unsure!!

Something quite unusual happenned today, I have two boxes of free space connected, the source in one of them turned on and no source in the other, A one volt wave is sent in due to the source but when I plot the absolute value, the amplitude was only 0.5Volts!

The abrupt ending of the source at the boundary of the two mediums (even though they are both the same index) somehow makes this happen but I am unsure how else to specify a source!

I am waiting on COMSOL support about my initial issue, however it would be great to hear your suggestions on this, so let me know what you think - Maybe I am making a really simple error somewhere - it'd seem odd if something like reflection didn't work in COMSOL!!

Regards,
Prateek

Prateek,

In my own calculations, reflection coefficient has been coming out to be quite fine. Just that supposedly non-reflecting boundaries turned out to give away a little reflected signal in my case.

I use 2D axial symmetry mode with occasional shift to 3D cartesian coordinates only to verify that I am not messing up something in 2D axial model. And so far, results are giving perfect correspondence. Just to cross-check my model, I have also computed reflection coefficient for some cases presented in journal papers and that too worked out pretty nicely.

If you wish, send me up your model here. I can take a look.

Hi Shakeeb,

Thanks for taking out your time for this, Attached is my model, it is extremely simple, just a free space box and a glass box.

The free space box is the source set up as myHz = exp( i*k0_rfweh*(-y)). Maybe the way I setup my sources is incorrect!

I am using the Scattered Hybrid Mode Waves as the Field Type.

Thanks,
Prateek

Prateek,

I am very sorry for coming up so late because of being impossibly preoccupied at work. Hope you have the answers already from Comsol's support and hence my response is irrelevant already.

Nevertheless, I took a look at your model and I am afraid I cannot say much from experience as I have never worked with the scattering mode in Comsol. I always use Harmonic propagation module and the boundary condition definitions or rather the whole operating mechanism seem to be different in the two cases. Though I see that differences might be well understood by someone experienced in either of the two by taking a quick look at the manuals but honestly, my present work is too demanding already.

For example, in Harmonic propagation, I declare matched boundary and can also specify if I want wave excitation or not in addition to providing the value of propagation vector. Here in your case I didn't find any such possibility and cannot understand what is the default behavior then. Similarly, I don't understand why you have set the incident field zero. May be it was motivated by some of the examples given in model library but I cannot go through it.

In the end, my humble apologies once again for keeping you wait in vain. I wish you all the best.

Shakeeb,

That's fine, don't worry about it, I did notice that the behavior of the RF module varies quite a bit when changing from the Scattered mode to the non-scattered mode, I set the sources to zero because in my mind it implies the fact that previously there was no field present and the field has only resulted due to the presence of extraneous sources. Ofcourse, what this does is create complications at the boundary, but if I don't do this, I don't get the results I want - there might be an alternative way of doing this for sure.

Anyway, thanks for trying,

Regards,
Prateek