4:1 current balun – review and fix

This article reports tests on two 4:1 current balun configurations – a collaboration between Bruce, VK4MQ, and myself.

Purported current balun on a single magnetic core

Above is an attempt at a 4:1 current balun on a single core. Note that this is NOT wired in the insane series opposed connection of the WIA 4:1 current balun. Note also that this is NOT a Guanella 4:1 current balun (see below).

Lets measure the Insertion VSWR by placing a good 200+j0Ω load on the output terminals and measuring input VSWR over the range 1-30MHz. This load is what we will call an Isolated Load meaning it has only two terminals, and the current that flows into one terminal must flow out of the other terminal… in other words, the current MUST be balanced (ie equal magnitude but opposite phase currents in the two terminals)… we will come back to the Isolated Load later.

Above, measured InsertionVSWR. It is not too good, but not very bad either. Broadly the balun gives an almost reasonable 4:1 impedance transformation from load to input.

Now if the balun was a good current balun (Definition: Current Balun, Voltage Balun) we would expect the currents to be nearly equal magnitude and opposite phase if the load also provided a path the ground (being the input connector shield for this discussion), and we would expect the InsertionVSWR to remain very close to that of the Isolated Load.

So, let’s connect a short direct wire from the coax connector shield to one output terminal.

Above, the InsertionVSWR has gone off scale, this balun is hopeless at transforming this asymmetric load of 200+j0Ω with one end connected to the shield connection.

So, let’s connect a short direct wire from the coax connector shield to the other output terminal.

Above, the InsertionVSWR is a little better but not good, this balun is very poor at transforming this asymmetric load of 200+j0Ω with the other end connected to the shield connection.

Other tests were performed grounding the load for conditions 50+150Ω, 100+100Ω (ie perfectly symmetric and 150+50Ω. The configuration that gave the best InsertionVSWR curve was the 150+50Ω configuration, see A review of the Guanella 4:1 balun on a shared magnetic circuit for an explanation.

These measurements don’t really quantify the problem, but they do sound a warning that this is far from a good 4:1 current balun for general purpose use.

Notwithstanding that, they are very popular. The popularity might derive from (Sevick 2001) though he does state it works only on Isolated Loads, but Isolated Loads MUST have perfect current balance, they do not require a balun. Most HF wire antennas are not well represented as a Isolated Load.

Nice case, can this be fixed?

Guanella 4:1 balun (on two independent magnetic cores)

(Guanella 1944) described a 4:1 current balun in his 1944 article, he did not show the winding pairs coupled by a magnetic core as shown above.

Guanella421

Above is Guanella’s circuit, and he does not show coupling between the two winding pairs.

So, can we fit two cores in the box and wire them as a Guanella 4:1 balun?

The Guanella 4:1 balun comprises two 1:1 Guanella baluns, one end of each connected in parallel to the coax socket and the other ends in series. The transmission line in each of the component baluns should be of uniform Zo and the same Zo and length for each. That constraint presents a little challenge in accommodating the two component baluns, and the solution involves a longer that desirable transmission line tail on each of the component baluns.

As with any Guanella 1:1 balun, the design is a compromise between using the shortest transmission line section and having sufficient turns on the particular core to deliver fairly high common mode impedance which is important to good current balun behavior. Further, that for a 200:50Ω balun, Zo of the transmission line sections should be close to 100Ω for best InsertionVSWR.

A good choice for his enclosure is the locally available LO1238 core from Jaycar (2 for $5) with ΣA/l=0.0009756/m. The LO1238 is a toroid of size 35x21x13 mm, and medium µ (L15 material).

We will use 9 turns on each core to give a Zcm profile similar to the predictive model above,  and it should have sufficient turns for 3.5-30MHz.

Note the twisted pair transmission line is maintained as uniform as reasonably possible throughout, and the transmission line length for each core and its tails is the same.

Lets measure the Insertion VSWR by placing a good 200+j0Ω (Isolated) load on the output terminals and measuring input VSWR over the range 1-30MHz

Above, measured InsertionVSWR. It is much better than the single core balun above, and gives good 4:1 impedance transformation from load to input.

Now if the balun was a good current balun (Definition: Current Balun, Voltage Balun) we would expect the currents to be nearly equal magnitude and opposite phase if the load also provided a path the ground (being the input connector shield for this discussion), and we would expect the InsertionVSWR to remain very close to that of the Isolated Load.

So, let’s connect a short direct wire from the coax connector shield to one output terminal.

Remember the number of turns is chosen for 3.5-30MHz. Too few turns causes the Insertion VSWR to rise at the low frequency end and too long a transmission line causes a steeper slope at the higher frequencies.

So, from 3-30MHz, the worst InsertionVSWR is 1.2 @ 30MHz. This balun is very good at transforming this asymmetric load of 200+j0Ω with one end connected to the shield connection.

So, let’s connect a short direct wire from the coax connector shield to the other output terminal.

Above, the InsertionVSWR is a little better, near perfect a the lower frequencies (ie down to 3.5MHz) and rises to just 1.16 @ 30MHz, this balun is very good at transforming this asymmetric load of 200+j0Ω with the other end connected to the shield connection.

These measurements don’t really quantify the behavior, but they do give encouragement that this is an effective 4:1 current balun over the range 3.5-30MHz for general purpose use.

This is the fix for the single core 4:1 current balun described earlier… and Guanella discovered this 76 years ago.

Do you really want / need a 4:1 balun?

It might seem a bit late to visit this question, but better late than never as they say, see Is a 4:1 balun a good choice for use with an ATU on HF?

Wire type

Readers of On use of enameled wire in transmitting baluns will know that I would prefer PTFE insulated silver plated wire of around 0.5MM^2 size, it has superior voltage handling. The enameled wire was on hand and it allowed measurement of the proposed design.

Conclusions

The single core balun is far from a good general purpose current balun, it has high InsertionVSWR for all but Isolated Loads and one specific case 150+50Ω asymmetric loads.

You do not need a balun for Isolated Loads, they MUST have perfect current balance.

Real world HF wire antennas are unlikely to be well represented as an Isolated Load (Equivalent circuit of an antenna system).

The Guanella balun described and measured uses low cost cores readily available in Australia and performs quite well.

References