Skip to content

Three ledgers of efficiency: modeling the POTA PERformer

Greg Mihran KJ6ER’s PERformer is one of the most popular published portable antennas: an elevated quarter-wave vertical for 40M–6M — a 17’ telescoping stainless whip with the feedpoint 52” up on a tripod or spike, and two elevated tuned radials sloping down to stakes. The free plans include 4NEC2 model results and a headline claim: over 90% efficient, versus “only 37%” for a typical ground-mounted quarter-wave with four ground-coupled radials.

verticals.pota_performer models it faithfully — the plans’ 15M reference geometry, per-band variants straight from the whip/radial tables, the 90° directional and 180° omni radial spans, and VA3KOT’s single-radial simplification. This page is about what happens when you check the numbers: the claims duplicate, and reading the efficiency number in full teaches the most useful lesson in portable antenna modeling.

Four models from three modelers — spanning two genuinely independent solver cores, the NEC-2 family and momwire — on the 15M configuration, two radials, average ground:

ModelPeak gainTakeoffF/B (90° span)El. beamwidth
KJ6ER, 4NEC2 (published)+0.31 dBi24°3.37 dB46°
VA3KOT, EZNEC (independent)+1.19 dBi25°3.34 dB47°
antennaknobs, momwire+1.06 dBi24°2.91 dB44°
antennaknobs, PyNEC+1.02 dBi23°~3 dB44°

The beam shape is unanimous: takeoff 23–25°, a mild ~3 dB front-to-back toward the radial span, mid-40s elevation beamwidth. The directional-vs-omni delta duplicates too (KJ6ER +0.98 dB, we measure +0.76 dB), and so do VA3KOT’s single-radial trends (more F/B, wider elevation, narrower azimuth). Absolute peak gain spreads about 0.9 dB across engines, with KJ6ER’s own figure the lowest of the four — ordinary cross-model variation in ground constants and whip idealization.

The SWR story holds up as well: at the plans’ exact 15M lengths the directional feedpoint solves to 45 + 1j Ω — SWR 1.10, matching the “better than 1.1:1” field measurements (and the omni span solves lower, just as the plans say elevating radials should).

“Over 90% efficient” is a structural efficiency: input power that isn’t burned in conductors and components. We confirm it. Solving the same geometry with lossless wire and taking the ratio: the stainless whip plus radials cost only ~2–3%, structural efficiency ≈ 98%. KJ6ER’s 90.8% is a few points lower — plausibly the real whip’s thin stainless tip sections, which our uniform mid-taper radius flatters (his −0.12 dB choke is itemized separately in his table, not inside the 90.8%). And the physics behind his 90%-vs-37% comparison is real: elevated tuned radials remove the ground-coupled loss resistance that sits in series with a ground-mounted vertical’s feed and eats over half its power in-circuit.

But there is a second ledger, and it’s already visible in the table above. Gain and efficiency are the same measurement. Gain is power density per input watt; average linear gain over the sphere is exactly Pradiated/Pinput (radiated_fraction in the library computes it from any pattern). Equivalently: this beam shape has a directivity of ~6.5 dBi — that’s what the peak would read if every input watt were radiated. The peaks actually read about +1 dBi:

ModelPeak gainImplied radiated fraction
KJ6ER, 4NEC2+0.31 dBi~24%
VA3KOT, EZNEC+1.19 dBi~30%
momwire (integrated directly)+1.06 dBi29%
PyNEC (integrated directly)+1.02 dBi34%

So for 100 W in: ~2 W warms the stainless, ~68 W warms the ground within a few wavelengths, ~30 W leaves as sky wave. The missing 4.5–6 dB between “+5.5 dBi if lossless” and “+1 dBi as modeled” is the Fresnel-zone ground absorption, and it appears in nobody’s “efficiency” figure — including the 90.8% — because structural efficiency by definition stops at the antenna’s terminals-and-metal. KJ6ER’s own published +0.31 dBi peak is a ~25% radiated fraction, stated in dB.

Ground absorption at these heights is nearly identical for every portable vertical. For calibration, the same integration on a 7 m-high inverted vee gives 72% radiated on 20m, 72% on 15m, 69% on 10m — the classic reason horizontal wire beats a vertical over real ground when you have the supports, and the vertical’s counter-argument is the low takeoff angle (23° here vs the vee’s 51° on 20m), not efficiency. You cannot pack better dirt. That’s precisely why the convention in antenna write-ups is to quote structural efficiency: it’s the part the builder controls, and on that score the PERformer is genuinely excellent — the elevated-radial design earns its numbers, and the relative claims (beats ground-mounted, 3 dB of steerable F/B from a 90° radial span, resonant with no tuner) all survive independent modeling.

The lesson is about reading claims, not doubting this antenna: when a spec sheet says “90% efficient” and the same page’s pattern plot peaks near 0 dBi, both numbers are correct — they are different ledgers. The gain plot is the one the ionosphere sees.

The rest of the trio: Challenger and Dominator

Section titled “The rest of the trio: Challenger and Dominator”

KJ6ER publishes two more antennas in the same family, and both are now in the tree: verticals.challenger (off-center-fed halfwave vertical: 25’ whip is ~77% of the halfwave, a short ~10% λ counterpoise completes it through a 4:1 unun) and verticals.dominator (a true vertical EFHW: the whip is the whole halfwave, fed through a 49:1 with a long ~33% λ counterpoise). Same duplication exercise, same result — the claims hold:

Claim (4NEC2, 15M)Publishedantennaknobs
Challenger peak / takeoff / el BW−0.32 dBi / 20° / 33°+0.14 dBi / 21° / 34°
Dominator peak / takeoff / el BW+0.60 dBi / 18° / 27°+0.34 dBi / 17° / 26°
Takeoff ordering, trio18° < 21° < 24°17° < 21° < 23°

The transformers are where these two get interesting, because KJ6ER measures their insertion losses and itemizes them honestly — and our Transformer branch puts the same numbers in the power budget: the stock 49:1 burns −0.96 dB (~20% of input power) — the same black box interrogated in the end-fed question — the Challenger’s 4:1 only −0.34 dB, with the “plus” upgrades at −0.40/−0.24 dB (plus variants; the magnetizing branch is calibrated to the measured loss at 15M, not derived from core datasheets). His own comparison table shows the interesting tension — the Dominator has the trio’s highest “structural efficiency” (99.5% — it’s just one aluminum tube) and the trio’s lossiest component — and to his credit both rows sit side by side on his page 14. Our power budget simply adds them, plus the dirt’s share, into one delivered-watts number.

And the third ledger says: Challenger radiates 25%, Dominator 25% (transformer included), PERformer 29%. Nobody escapes the dirt. What actually separates these three antennas — exactly as KJ6ER’s “primary reach” table frames it — is the takeoff angle, and that claim reproduces perfectly.

Open verticals.pota_performer in the simulator. Flip radial_span_deg between 90 and 180 and watch the front-to-back appear and vanish; drop n_radials to 1 for VA3KOT’s fast-deploy config; walk the band variants (band20band6) and watch the droop angle steepen as the radials shorten against the fixed stake height — exactly the table in the plans.

The third ledger is on screen while you do it: with a finite ground selected, the solve readout’s radiated (incl. ground) row (and the radiated % readout on the pattern views) fills in whenever the knobs settle — ~29% for this antenna, the number this page is about. See the workbench reference for how it’s computed.

Sources: KJ6ER’s PERformer plans (rev 2025-02); VA3KOT, “Testing and modifying the POTA PERformer antenna” (2025-05).