different jsurface-results between MuPAD- and Octave-Evaluation :(

[Leo Baumann]

Hello,

with MuPAD Computer Aided Mathematic I have correct evaluated this 
(watch the figure):

www.leobaumann.de/newsgroups/horDipol.pdf

I am learning Octave since Sunday and tried to adapt that problem to 
Octave. The results are different, the Octave-result is wrong :(

www.leobaumann.de/newsgroups/Dipol.png

I used this Octave-script:

www.leobaumann.de/newsgroups/horDipol_Test.pdf

I have absolute no explanation for the different results. Did I make a 
fatal mistake? -

Does anyone have a hint?

thanks - regards - Leo