Re: NEC-LIST: Basis functions on short segments.

Jerry,

I had already read pp9-10 in the NEC4 Theory, and found tyhe
wording unclear. On its face, it would seem to mean that
the coefficients in the NEC2 expansion will lose precision
as kd->small. However this isn't supported by the equations
for the coefficients in the two manuals, which are very
similar. In fact, the equations for the A coefficient for
the tail functions in NEC4, (3020) and (3-23), do have a
small-difference-of-large-numbers problem as kd->small. I
had a hard time getting an approximation for them, and wound
up using successive numerical evaluations to tell me what
the good approximation should be.

On the other hand, maybe the bottom of page 9 was referring
to the evaluation of the fields due to each subcomponent
function. A reading of the section on field equations
appears to show that the field from a cosine current is
being considered as in NEC2, not the field due to a
cos(k(s-s_j))-1 current. Without digging into the code, it
appears to me that NEC4 uses A-C to find the field for
constant current elements, and C to find the field for
cosine current elements.

It is true that charges at the segment ends are handled
differently, but this is not related to the organization of
the basis function subcomponents.

Regards,
Doug

P.S. Sorry for sending this to you twice. I forgot to
change the entry in the "To.." box after I did the reply.

-- 
The NEC-List mailing list
NEC-List_at_robomod.net
http://www.robomod.net/mailman/listinfo/nec-list