[TowerTalk] Baluns/tutorial/notes.

Jim Brown jim at audiosystemsgroup.com
Sat May 22 11:56:07 PDT 2010

On Sat, 22 May 2010 18:51:14 +0100, Steve Hunt wrote:


>Thanks for the clarification. As you say: "He who never made a mistake 
>never made anything".

>I remain puzzled by what you just said about the flux in the cores of 
>N:1 Current Baluns. Any analysis I've done shows that the flux in one or 
>more of the cores *must* be determined directly by the Differential Mode 
>signal. Only in a 1:1 current balun is the flux solely dependent on the 
>Common-Mode signal.

>A simplistic way of looking at it is to take a 4:1 current balun 
>comprising two 1:1 baluns connected in parallel at the input and in 
>series at the output. For a differential voltage applied across the 
>input of V, the voltage across the load will be 2V. The extra "V" has to 
>be developed somewhere, and it is by "voltage transformer" action in one 
>or more of the cores.

I think that's what I said. If we recognize the Guanella balun as two 
common mode chokes, the differential voltage is across the common mode 
impedance of each core. Doing the math, for 1500 watts keydown, that's 275 
volts across 5,000 ohms, which is 15 watts. Applying a reasonable duty 
cycle for CW or SSB, that's something like 5 watts when you're 

>I sketched the voltages here for a floating load:

>Excuse the rough drawing. That example is the most benign situation, 
>where the differential-mode signal is split equally between the cores. 
>There's V/2 across each core, and that directly drives the core flux.

>If you try another load-balance configuration - for example ground the 
>"top" of the load - you'll find a much more extreme situation. In that 
>example you'll get V across one core and 2V across the other - that's a 
>worse condition than a 4:1 voltage balun! Ground the "bottom" of the 
>load and you will get V across the top core and 0 across the bottom.

"Ground" is probably not the right word/concept here unless something is 
badly broken. Say you're using a Guanella balun to feed a balanced line and 
one side of the line gets shorted to SOMETHING at some point on the line 
displaced from the balun. Even if that something is the earth, or a tower, 
the transmission line and the rest of the return path add a lot of 
impedance in that short circuit. In other words, "ground" is not a single 

Now, you COULD mis-use a Guanella balun by grounding one high Z side of it 
to the shield of the low Z side, and that WOULD place higher voltage across 
one of the cores. :) 

>No matter how the load is balanced or unbalanced wrt ground, there is 
>always that "extra V" being developed by voltage transformer action on 
>one or more of the cores .......

Again, the word/concept "ground" is a poor concept. Transmitters develop 
power between the terminals of their output connector. A connection to 
earth (or even the chassis) is not necessary. In general, antennas need no 
connection the earth to work well. They do, of course, need some 
arrangement of conductors to provide a return for antenna current and 
fields (i.e. a dipole or a counterpoise for a single-ended antenna), but 
again, neither earth or the chassis needs to be part of it.  

>.... I believe ;)

Look at it this way. 

There is no transformer action in a Guanella balun if each choke is wound 
its own core. Rather, the circuit is a sort of bridge, where the voltage at 
the high Z side is the IZ drop across each choke added together. We ARE 
transforming impedances, but the power is not going through the cores to 
get from one side of the balun to the other. Indeed, if the chokes are on 
individual cores, there is virtually no magnetic coupling between them! 
W8JI has also written about this. 


Jim Brown K9YC

More information about the TowerTalk mailing list