The EFHW Capacitor: It’s Just a Shunt Capacitor — Nothing More

The primary-side EFHW capacitor improves the match at the radio, but it does not increase the transformer’s intrinsic efficiency. It modifies what the radio sees, not what happens inside the transformer’s copper and core.

What the shunt capacitor actually does

Referred to the 50 Ω side, the EFHW input looks like:

• Rₑff ≈ 50 Ω (when the antenna is ~2450 Ω)
• jXσ = leakage inductance
• Lₘ and Rcore = magnetizing inductance and core loss

On the upper HF bands, leakage inductance makes the input look inductive. The shunt capacitor adds opposite susceptance at the input, pulling the impedance toward 50 Ω resistive and lowering VSWR.

But because it is placed at the port, it cannot cancel the leakage inside the transformer.

Why this does not increase transformer efficiency

Deliver fixed real power P into the referred 50 Ω load:

• Series current = √(P ÷ 50)
• Load-node voltage = √(P × 50)

Both quantities depend only on P and the real load, not on the shunt capacitor.

• Copper loss (I²R in the primary) does not improve
• Core loss (voltage-driven flux cycling) does not improve

Why builders still use the capacitor

• Feedline loss drops because SWR drops
• PA foldback is avoided because the radio sees a clean load
• Usability improves, especially on 20–10 m

These are system-level benefits. The transformer’s own insertion loss is unchanged.

The tuner analogy

A tuner in front of something lossy does not make that thing efficient — it just makes the interface look better. The EFHW shunt capacitor is exactly that: a fixed, one-component tuner at the input.

What actually improves transformer efficiency

• Core mix, size, stacking
• Tight magnetic coupling and minimized leakage
• Appropriate conductor diameter
• Avoiding core saturation on low bands
• Controlling stray capacitances
• How the coil is wound (single, double, triple, or multi-conductor)

A simple numeric check

Assume primary-referred impedance:

• Series current = 1.414 A
• Load-node voltage = 70.7 V

Adding 100–120 pF does not change these values. Only the input SWR changes.

The Hidden Capacitor Trap

The capacitor across the 50 Ω side resonates with the transformer’s magnetizing inductance Lₚ. At that resonance, the Lₚ ∥ C branch becomes purely resistive because their reactances cancel.

If the core is lossy at that frequency, this branch behaves like a real shunt resistor that absorbs power — while still presenting a near-50 Ω input.

SWR looks excellent; efficiency does not.

How large can the hidden loss be?

At resonance, the inductor’s minimum equivalent parallel resistance is:

Worst-case power going into that branch (with a 50 Ω source):

Example: FT240-43, 2-turn primary

• Lₚ ≈ 4.3 µH
• Resonance with 100–120 pF ≈ 7–7.7 MHz
• ω Lₚ ≈ 207 Ω
• Rₚ,min ≈ 414 Ω
• Hidden draw ≈ 50 ÷ (50 + 414) ≈ 11 % (≈0.5 dB loss)

This is real transformer heating — even while your SWR meter still reads “1.0”.

Where the shunt capacitor resonates

Practical meaning

• #43 with 2 turns: resonance ≈7–7.7 MHz → possible 40 m heating
• #52 with 2 turns: resonance ≈12.6–13.9 MHz → watch 20/17 m behavior
• Stacked cores: resonance shifts downward → less impact on upper HF

How to see it (simple checks you can do)

You do not need exotic equipment to spot the shunt-resonance trap. A basic VNA and—optionally—a thermometer are enough to confirm whether the capacitor is creating a benign match correction or a hidden loss branch.

Sweep R and X, not just SWR (with a dummy high-Z load)

• • Put a 2.4–3.3 kΩ non-inductive resistor on the high-Z port.
  • Sweep 3–30 MHz and plot R and X (or use a Smith chart).
  • A broad rise in the real part (phase ≈ 0°) near the computed f₀ = 1 ÷ (2π √(Lₚ · C)) is the signature of the Lₚ ∥ C shunt branch.
  • Repeat the sweep with the capacitor removed. The hump that does not move with the load is the Lₚ ∥ C branch — not the antenna.

Move the load to prove it is not a “real match”

• • Repeat the sweep with 1.5 kΩ and 5.6 kΩ on the high-Z side.
  • A true broadband match follows the load impedance.
  • A dissipative shunt branch (the Lₚ ∥ C resonance) stays near the same frequency, barely shifting.

When the capacitor can be detrimental

• If f₀ sits inside an operating band → increased loss and heat.
• If you size it correctly, it will usually look neutral or even helpful on 20–10 m from a matching perspective — but it also hides the unavoidable efficiency loss caused by leakage, especially on 15 m, 12 m and 10 m.

Interested in more technical content? Subscribe here.

Questions or experiences? Contact RF.Guru.