Tuner Baluns on Input

There is a common perception that placing a balun on the input of a tuner causes the balun to work better. The thought is the balun operates with a matched impedance and that reduces balun losses. It also is thought that moving the balun improves balance.

Anyone who tells you moving a current balun to the input of an unbalanced tuner helps balance or helps the balun is mistaken.

Let's take a look at how this concept actually works.

Spice Models

Ideal Tuner and Balun with Balun on Input

The following model has a near perfect balun on the input of a perfect unbalanced network. In the perfect case below ill-effects caused by floating the network, such as the unbalanced capacitance to ground from the network we are sure to have, are ignored. This is a better tuner than anyone, no matter how careful, can actually build.

The load is perfectly balanced with 1000 ohms total impedance. In this example we will assume it is purely dissipative. 50K ohms would be a VERY good balun, one nearly impossible to obtain.

Point A or R1's power dissipation is indicated by the green trace. Point B or R2's power dissipation is represented by the yellow trace. All voltages and power levels are instantaneous peaks. The capacitors and inductors are assumed lossless. Applied "power" can be considered to be about 30 watts.

Note: Resistor R3 represents represents the common mode impedance of the balun. The windings in T2 have unity mutual coupling, similar to a real bifilar balun. This is a simple but reasonably accurate way to represent the common mode losses of a 1:1 current balun.

In the above case we see the load is very well balanced. The green and yellow peaks are nearly identical. The load receives well balanced power. Balun heating is 1% of R2's power.

String of Beads Balun on Tuner Input

Let's see what happens when a more typical "string of beads balun" is used.

In this case power in the upper side of the balanced load, represented by R1, increases. There is also a shortfall of power in the lower side of the load. Power input slightly increases (I didn't readjust the network for the changed impedance) to 32.5 watts

R1 = 18 watts R2 = 11.5 watts We have 6.5 watts of load unbalance. The balun dissipates about 3 watts.

Balun on Tuner Output

Now let's compare the system above with a system having the same string of beads balun on the tuner output.

Notice all power levels are unchanged!

Unbalance in the load is the same
Balun heating is the same

We lost absolutely nothing by moving the balun to the output!

With the balun on the input we have the following peak common mode voltage.

We have about 75 volts peak across the 2K ohms of balun common mode impedance, or about 3 watts dissipation. This is about 10% of applied power. With 1500 watts the balun would be dissipating 150 watts. The beads would be too hot to touch in a matter of seconds.

Now let's examine heating caused by common mode voltage with the balun on the output.

Notice we have exactly the same balun heating power. Nothing changed. The balun core is under exactly the same electrical stress. Nothing was gained by relocating the balun to the input. Nothing was lost by moving it to the output! The same core size is required, balance is the same, and heating is the same.

Real Component Test

Some people might say the model is flawed. Let's look at real components and see how they work. First let's measure the common mode impedance of the bead balun I'm going to use.

This is the tuner configuration. In this case it was setup for a floating-L network with balun on the input:

The ground side of the floating L measured 10mV at the load:

The "hot side" of the floating L measured 185mV:

This is worse than the model because the beads do not have the 2000 ohms impedance the model assumed. They only have a few hundred ohms impedance.

Clearly have a string of beads on the input of a floating unbalanced network is a waste of time. It models this way, and it also it measures this way.

When does moving the balun help?

Moving the balun helps performance only under very specific conditions.

1.) When the network is a double balanced FLOATING network that is perfectly symmetrical

In this case, absent any common mode current fed back from the antenna, the move will reduce excitation of the balun core almost completely. In this case the balun actually has very little use at all! It can be eliminated in most cases with very little effect on system balance. In this case the tuner becomes a balanced voltage source.

2.) When the network is a double balanced network that is perfectly symmetrical and grounded in the center

In this case a modest balun is required, but the tuner becomes a very stiff split voltage source.

It's actually just as difficult to build the best kind of source, a balanced current source with high common mode isolation, by using a double network. If we spent only a fraction of the additional component expense required to build a "balanced tuner" on a good current balun, we could have a better balanced tuner in much smaller space! Instead of working towards a good current balun, designers have thrown money into a system that creates as many new problems as it cured old problems.

Balanced tuners are great when the load is closely balanced. They are not so good feeding Windoms, Zepps, or other antenna in the nether world between perfectly balanced and perfectly unbalanced. They also, for the same dollar amount invested, will not outperform a conventional network with a good balun on the output.