Look at a polar antenna plot (either in the vertical (elevation) or
horizontal plane, though horiz would be best for this exercise). The greater
the gain *straight* forward the less the angle that the gain is within
say -3dB of the peak. F/B ratio is a measure of how much more sensitive it
is dead ahead compared to dead behind, and ignores the sensitivity at any
other angle, generally (but not always) there will be one large lobe
pointed forward and only little bumps in sensitivity to the sides.

Another way to think of it (a thought experiment if you will) is as a garden
hose w/ a fine stream out the front (very high gain and very high F/B ratio
(very little water shoots straight back at you, and only a small spot that
you aim at gets wet, vs. a hose w/ the end cut off and flattened to form a
flat fanlike spray -- still good F/B (you stay dry) but the water goes less
far and covers a bigger horiz angle. Last, imagine you unflatten the open
hose end, point it straight up and hold a plate over it so water sprays in
every direction horizontally -- this is an omnidirectional pattern (like a
dipole, where RF leaves perpendicular to the wire or element) and goes even
less distance (gain) but goes in every horizontal direction. In all cases
the same amount of water comes out, but as you change the nozzle you send
more in one direction at the expense of all the others. The antenna is the
same when transmitting or receiving (collecting if you will) RF energy.

If your hose has some small pinholes in it you may get "side lobes" but they
will usually be narrow angled or weak (low gain) -- differant cause than the
antenna, but same effect for the purposed of the thought experiment.

Last, if you had a hollow spherical ball much bigger than the hose diameter
that fastened on the hose's end and this ball had little holes evenly spaced
all over it (and a good imagination) water would go in EVERY direction
EQUALLY -- the equvalent of the *theoretical* isotropic antenna.

The analogy is not exact, but it should give a feel of the thing.

Note too some water or RF goes out at an angle from horiz (this is what you
see in the vertical plane polar charts) and if you hold th ehose or antenna
too close to the ground none can go down (it either soaks into the ground or
bounces off and goes upward a bit) giving a slight "takeoff angle" to the
beam (or stream).

That help, or only confuse?

"Fractenna" wrote in message
...
Then how come it doesn't give us 20db of forward gain?

It does, if you (incorrectly ) define forward gain as being relative to
the
backside.

As I said, the signal (that) now no longer goes out the back gets its
power---redistributed--and that redistribution is like pushing the air
forward.
It doesn't -- all-- get concentrated in one spot up front, thus the
forward
gain isn't 20 dB.

73,
Chip N1IR