Thanks for this. It is not clear to me why a 4:1 balun should be 'more
tolerant' of a mismatch but clearly impedance will be different at
different frequencies. Below is a quote from
http://www.packetradio.com/windom.htm, one of several descriptions of
multiband dipoles I have seen all agreeing that 4:1 works in practice.


QUOTE
"I'm feeding my dipole with 600-ohm line. At the station end I need
a balun to convert to 50-ohm coax. I need a 12:1 balun, right?"
Wrong! A 4:1 balun would be better.

Why is that? If your dipole is up, let's say, 35 feet then on 80
meters it will probably have a resistance at resonance of about 40
ohms. The actual resistance depends on the height above ground in
wavelengths.

If the dipole is 40 Ohms then what do you see at the transmitter end of
your 600 ohm line? If the line is a half-wave long (120 ft on 80
Meters) you'll see 40 ohms. Remember, a half-wave line repeats what
it sees at the other end. But if it is a quarter-wave long you'll see
8500 Ohms! At other line lengths you'll see impedances somewhere
between these two extremes.

So you are not going to see 600 ohms at the end of your 600-ohm line.
That only happens if you have a 600-ohm antenna hooked onto it. With
such a variation in impedance at the trans=ADmitter end of the line
there is no one balun transformer that will match it. Most of the time
the impedance will be above the 50 Ohms of your coax so a high
impedance balun would be desirable. Unfortunately high imped=ADance
baluns don't work well when not matched.

Experience has shown that 4:1 baluns work best in this service. They
are more rugged and will take bad mismatches especially if they are
wound on an iron powder core. So stop searching for that 12:1 balun.
Use a 4:l BALUN and your system will work great. =20

ENDQUOTE