NEC-LIST: Modeling a bondwire

Got stuck with the seemingly very simple problem: bondwire, that is a
single short curved wire above the perfect ground fed at one end and
terminated to 50 Ohms at opposite end. Wire is about 1 mm long, maximum
height above ground is about 0.25 mm, frequency 1..100 GHz. This is
actually an inhomogeneous open transmission line and the stumbling stone
is that I cannot figure out how to attach generator/load to the wire
ends without adding any additional wires that provide return path to the
ground. I guess these additional wires will distort the actual current
distribution.
Actually, the terminated end have a sloped segment that can be grounded
but what about generator end that must float in air? I tried NT and TL
to provide "invisible" connection to small vertical grounded segment but
failed.
BTW, even if I let wire start and end to join the ground, attach
generator somewhere at half of total height and load the last segment,
results do not look very encouraging.
IMHO, resonance frequency is too low (61 GHz).

Have two questions:
Isn't this "low profile wire" problem too hard for NEC to get reliable
results?
Can anybody offer a workaround to the problem of "floating and grounded"
wire?

Thanks,

Samuel

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