On Tue, 09 Mar 2004 10:40:10 +0100, Jeppe wrote:

But are you sure that 1 wavelenght is the optimum value for the other
dimension of the cross section? That is more than in the examples I
have seen on the net.

Is there some reason for choosing a 2 : 1 ratio of the two dimensions?

I may add that in the first horn I made, the dimensions were 0.36 x
0.72 wave. It worked, but probably not quite optimallly (I lack real
measuring equipment.) I then tried to put the antenne rod
perpendicular to the SMALLER side instead, but that did not work at
all (very weak signal).

73

Jeppe (former OZ3FV, hav'n't renewed my licence for about 30 years :-(

Hi Jeppe,

The physical/wavelength dimensioning in microwaves is one of the best
ways to visualize transmission concepts. The ratios you observe offer
the means to render conduction and isolation. The best treatment of
this subject will be found in Frederick Terman's work.

The placement issue of the rod is easily answered. In the picture you
have, it aligns with the creation of electrical fields that would be
naturally beneficial to propagation. When you experimented by placing
the probe on the other wall (90° off) you lost what is called mode
coupling. This is because what you have is a voltage probe that used
to couple to the voltage modes of the resonator. With this 90° twist,
your voltage probe was trying to excite the current modality. You
should have, instead, built a coupling loop to create the magnetic
fields that correspond to the cavity's mode. You would have observed
the same level of gain then.

Lest some of my treatment lead to confusion, voltage probes (the rod)
couple to electric fields; current loops (the alternative for that
experimental position) couple to magnetic fields. The electric fields
and the magnetic fields are at 90° angle to one another in what is
called a "mode." The mode is generally fixed in one orientation (we
can add complexities of many modalities, but it wouldn't add any
significant instruction). Thus with this fixed geometry, you shift
the position of the coupling to suit all considerations.
Additionally, you must take care to orient the loop correctly to work.

73's
Richard Clark, KB7QHC