"Roy Lewallen" wrote in message
...
There are a lot of factors that can potentially affect measurements. The
first and most obvious is impedance matching, which has to be done to
extract the maximum possible power from the antennas (assuming that this
is what you'll be doing when actually using the antenna) and delivering
a known amount of power to each. Another is to make sure the feedlines
aren't radiating or picking up signals, by using proper baluns, which
you've mentioned. Move the feedlines around and change their lengths,
making sure the results don't change. If they do, your baluns aren't
adequate. If you want quantitative measurements, you'll have to
carefully calibrate your signal strength meter at the power level
involved. Don't make assumptions about its linearity. Better yet would
be to carefully match the antennas to 50 ohms, then insert a step
attenuator in series with the Yagi in the 50 ohm environment and
increase the attenuation until it reads the same as the dipole. That
way, the FS meter linearity is immaterial (although the attenuator
accuracy is important). Then there's the possibility of signals radiated
off the back of the dipole being reflected from near or distant objects
which would affect the dipole's measured field strength more than it
would the Yagi's. The different widths of the forward lobes can also
cause unequal reflections. Although reflections can affect the forward
gain up to several dB, their impact on nulls or front/back measurements
is likely to be greater.

I'd do a couple of things. One is to build an NBS reference Yagi which
is easy to construct and has a well known and documented gain. It's also
easy to model. Modern modeling programs do very well with full size
Yagis. Measure this on your range and verify that the measurements agree
with its known properties. Another check would be to rotate the dipole
and see if its pattern is what it should be. And rotate the NBS Yagi and
verify that its pattern matches modeled results. Only after doing those
tests would I have reasonable confidence in other measurements made with
similar types of antennas.

Roy Lewallen, W7EL


I thought that NIST prefers a dipole antenna up to 1 GHz, and then a
waveguide-fed horn beyond that. I would think that a pyramidal horn would be
more predictable that a Yagi, and not unreasonably large, even for 400 MHz.
Plus, the construction is simpler, and the design is more robust over time
(i.e., it's easy to bend a Yagi element slightly, not enough to notice it,
but enough to shift performance).

Ed
WB6WSN