A Low-Loss EFHW Transformer

 The End Fed Half Wave (EFHW) is a convenient antenna. The core component is a wide band impedance matching transformer. As with all antenna systems, efficiency is a key metric. Special selection of the transformer core and winding style can allow for an acceptably efficient EFHW antenna.

Quick Facts

Other authors have already documented the design in far greater detail, so I will provide only light detail.

The toroidal core is part number 2643251002 from Fair-Rite. It is a 43 mix, outer diameter of 39mm, and a length of 22mm. It has a similar weight to the classic FT240 size core, however quite different dimensions.

The winding is made as per autotransformer style, with 14 total turns and a tap at 2 turns. This provides a turns ratio of 14:2 and so an impedance transform of 49:1.

The compensation capacitor is a 100pF silver mica capacitor, for low loss and high stability.

Well Proven

There are a number of examples of this specific design from others.

• Owen Duffy constructed and validated the design. Check out the multi-part blog posts.
• Evil Lair Electronics posted a YouTube video on the construction and validation of this design.
• MM0OPX also posted a YouTube video on the construction and validation of this and a plethora of other designs. The table of results is available for viewing.

The loss for all examples is less than 0.8dB across the entire HF band, with an average of around 0.5dB of loss. For reference, -0.5dB is approximately 0.89 as a linear figure - this is acceptably efficient.

My Own Instance

Motivated by the dB loss figures (or lack thereof) I set out to construct my own.

A Vector Network Analyzer (VNA) is used to measure the loss. Most VNAs have ports with an impedance of 50 ohms. This creates a roadblock since this transformer deliberately transforms the impedance away from 50 ohms.

The way around this roadblock is to use two impedance transformers and connect them back-to-back. This means the VNA measures the characteristics of two transformers. Assuming the transformers are identical, then the VNA measurements can be halved to get per-core figures.

It's important to keep stray inductance and capacitance to a minimum. See photos of my measurement below. Note the worst |S21| figure of -0.88dB. Halving this figure results in a per-core loss of 0.44dB which is in the ballpark of the measurements by the other authors.

The |S11| measurement is a good sanity check to validate the transformers are near identical.