Learning Forum

[sara.bassil]

Hi, 

I'm trying to simulate a directional coupler in the EME solver 

For that, I defined my structure as followed :
But I can't seem to get anything at my ports when I do a sweep of the coupling length from 0 to 500µm:

 

So, I tried this configuration : 
where the upper waveguide is straight and the bottom one has 2 s-bends.
Thus, I get : 

 

Can someone help me please ? 
Is EME not the right solver for this ? 

 

Thank you very much!

Sara 

[Guilin Sun]

This is most likely due to lack of number of cells in the curved sections. EME depends on the mode matching at adjunct cells. If the waveguide shape changes fast, the mode power cannot effectively be transferred to the next cell.  Please try the second device with more number of cells and make sure the power coupling is correct.

You can check the mode overlao matrix at each cell and see if the coupling is lowere. Please refer to those articles:

Understanding EME error diagnostics

Convergence testing process for EME simulations

EME Convergence Testing - An Intuitive Approach

[Guilin Sun]

Maybe better to just test one curved waveguide and make sure its transmission is almost 100%.

[sepehr]

Check your port configuraitons. Use mode-selection as "user select" and then calculate modes and see if they makes sense. Its likely that your port geometry is not configured properly.