Roy

When I was working with antennas, we considered the antenna's F/B ratio
used the max of the front compared to the max of the back. But, I get the
impression that the rules are different now.

Jerry


"Roy Lewallen" wrote in message
...
wrote:
. . .
I have seen computor results
that offer 50 db F/B based on NEC, Can I trust gain if F/B cannot
be trusted?

Absolutely! While you might get some very deep nulls at some particular
points in space, and fairly deep nulls in some particular
azimuth/elevation angle combinations, they're not likely to be exactly as
deep or in the directions the program reports. Gain, on the other hand,
can be strikingly accurate in many cases.

Put together any model you want with an extreme F/B ratio. Then fiddle the
model just slightly -- change the frequency, element length or diameter,
etc. Look at how much the gain changes, and how much the F/B changes.
Modify it more, and look again.

You'll see that the F/B is *much* more critical than gain. You can goof up
the model -- or real antenna -- a lot more without any appreciable change
in gain than you can before seeing major changes in F/B.

The reason is simple. To get a deep null and therefore good F/B ratio, you
have to add the fields from all parts of the antenna together to get zero
within a tiny, tiny fraction of a percent. If any one of the fields
changes just a tiny amount, they no longer sum precisely to zero. But
small change like that won't noticeably affect the gain. No model is good
enough to precisely predict extremely deep nulls -- there's always too
much difference between the model and reality.

I don't recall what Reg recently said, but I've gotten 50 dB and greater
F/B ratios from an array by adjusting the phasing network while listening
to a receiver placed in the null direction. But the null is that deep only
in that direction, at that height above ground. It's also noticeably
shallower a little ways away even in the same direction, because I've
compensated for re-radiation from nearby objects, too. Even coax shield
leakage becomes a very noticeable factor. So while I can tweak an array to
get a very deep null, there's no way I can expect that to hold when
anything changes, even just a little. I'd even expect it to change from
day to day as the ground moisture changes and the sap rises in the trees.

Obviously 50 db is hard to get but is it beyond the realms
of possibility?

For what, one particular azimuth/elevation combination at one single
frequency? You might be able to do it. But it would be only of academic
interest at best.

. . .

Roy Lewallen, W7EL