Low power Guanella 1:1 balun with low Insertion VSWR using a Fair-rite 2843009902 binocular – design workup

The article Low power Guanella 1:1 balun with low Insertion VSWR using a pair of Jaycar LF1260 suppression sleeves describes a current balun with low Insertion VSWR for operation at modest power levels. The design was based on Jaycar LF1260 cores which are readily available in Australia.

This article presents the workup of a balun with similar design objectives using a low cost Fair-rite 2843009902 binocular core (BN43-7051).

Above, a pic of the core.


The design is a variation on (Duffy 2007) which used RG174 coax for the choke to give low Insertion VSWR.

For low Insertion VSWR, the choke uses 50Ω coax wound around a pair of ferrite tubes. The coax is a miniature PTFE insulated cable, RG316 with silver plated copper centre conductor (be careful, some RG316 uses silver plated steel and it is less suitable for HF).

Matched line loss in the 350mm length of coax is 1.2% @ 30MHz, 0.4% @ 3.5MHz, and could be higher or lower with standing waves.

PTFE coax is used for high voltage withstand and tolerance of high operating temperature.


Above, an insulation test of the RG316. It withstood 7kV peak (5kV RMS) from inner to outer, and the jacket also withstood 7kV peak at a knife edge. Voltage breakdown is more likely to occur somewhere else in the balun.

For this design, the cores need to be large enough to accommodate 4 passes of RG-316 coax, but no larger.

Above, the cores will accommodate four round conductors of diameter 2.6mm, so they will comfortable accommodate the four passes of the RG-316 coax (2.45mm each). (For the mathematically minded, the minimum enclosing circle diameter for four equal circles is 1+√2 times the diameter of the smaller circles.)

Al (10kHz) is about 9µH.

The main contribution to loss and heating will be the ferrite core losses, and they are dependent on common mode current.

Above is a first estimate of common mode impedance of 3.5t (4 in one hole, 3 in the other – an approximation) assuming an equivalent shunt capacitance of 2pF. The latter is an experienced guess, and will be adjusted upon measurement of a prototype.

Implementation will be described in a follow up article.