I mentioned in my article WIA 4:1 current balun that the use of a single toroidal core in the above graphic compromises the balun. This article gives a simple, but more detailed explanation for the technically minded of why the shared magnetic circuit ruins the thing.
I have marked two windings with blue and magenta dots.
If the windings were on separate toroidal cores, the inductance of the blue and magenta windings each (given the construction details) measures 10µH, and both would appear in series across the input terminals to give an equivalent inductance of 20µH in shunt with the input terminals. The impedance will include not only the reactance of the inductors, but also some relatively small resistance due to wire and core losses.
In the case where they are on the same toroid, the magnetising force due to current in one coil opposes the magnetising force due to current in the other coil, so for a given current, less flux is developed (they don’t fully cancel due to 20% flux leakage) and so the pair of windings in series opposed connection on one core has much lower inductance, 4µH in this case, or about one fifth of the case of using two separate cores.
It is that 4µH equivalent shunt inductance that is the main cause of the VSWR rising below 20MHz in this test with a 200+j0Ω load.
The simple fix is that if you don’t want the current in the blue coil causing flux in the magenta coil (and vice versa), and you don’t, put them on separate toroids as intended by Guanella.
If this balun was built on two toroids, that SWR should be much better from 4 to 20MHz due to the much higher inductance (20µH).
Note though that simply using two cores won’t address all the design defects of this balun, the material is not a good choice for a HF general purpose balun.
More at WIA 4:1 current balun – further measurements.
References
- Duffy, O. 2008. A review of the Guanella 4:1 balun on a shared magnetic circuit. VK1OD.net.