Hello Owen, excuses for the delay, because I was caught by a very bad
cold...

Your design is for an o/c stub. Haven't you modelled a s/c stub by
connecting the transmission line to W998?

No, at least it was not meant to be. When TL cards are specfivied
without
additional real- and imaginary shunt admittance values they are to be
considered as an open end, although a dummy wire/segment is need
for connecting the TL end.

I have checked your proposed design on Winsmith, and it gives a 50
ohms input impedance.

Aha, nice to know. It seems the Cebik program is not in error.

Nevertheless, it is hypothetical as it is not
practical to connect a 300 ohm stub to 50 ohm line, 50 ohm line will
be coax (in practical cases) and it is not practical to make a 300 ohm
coax. If you were to mix coax and open wire line, you have a serious
balance issue that you have not dealt with in your model.

Yes, I know.

Who told you that the 50 ohm line was not symmetrical... (hi, hi, joke)

The other thing about the design is that two of the transmission line
elements are very close to 90 deg in length, which means that
operation is very sensitive to frequency change.

That's an interesting aspect of the situation.

If you really want to make a 40m inverted v work well on 80m, you
might have to match at the feedpoint. It is hard to get an unloaded
dipole of length 25%wl to work well without matching at the feedpoint.

The specified problem was only hypotetical, and does not exist in real
life (at least overhere).

Thank you, and the other posters, for the feedback,
Arie.