[2D FDTD TFSF problem] got error in result but don't know why

HI all,

I did wrote codes for 2D FDTD TFSF TMZ mode using yee algorithm. But the output I got was not as expected. The curve was bended, and major leaking after some time when approaching ilast, jlast. Can anyone help me to analyze the output graph eg, why the leaking happened, and why the curve was bending. I didn't put any ABC,PEC or PML for this time being, and 100% referred to taflove.

t = 68sec
http://i820.photobucket.com/albums/z...ni/data068.jpg


t=79sec
http://i820.photobucket.com/albums/z...ni/data079.jpg

t=94sec
http://i820.photobucket.com/albums/z...ni/data094.jpg

t=104sec
http://i820.photobucket.com/albums/z...ni/data104.jpg

please advice and thank you in advanced.



image for my 2d fdtd tfsf

can you describe "TFSF"? From your gif, there is obviously reflections at outer most boundaries and you also introduced a inner boundary that reflect field as well. You said you did not put ABC, PEC, PML, but obviously you have two boundaries at outside and middle. You are contradicting your own statement.

For a 2D FDTD, I always start with pure 2D problem, one cell along z direction with top and bottom set as PEC. Then the XY walls set as ABC. In such case, you should see the wave excited from the center should propagate out as a pure circle, and get some reflections at the XY corners due to the 1st order ABC's poor performance. If you can see this, it means your 2D FDTD should be in principle working.