Ohmsums

Antenna tables modelled in NEC2

Resonant length and feed-point impedance of a half-wave dipole and a quarter-wave vertical with 4 radials for every HF band and 6 m, calculated with the NEC2 antenna simulation engine instead of a rule of thumb, with the 468/f rule beside it.

Computer model, not a measurement. Cut the wire a little long and trim it.

Antenna length calculator A4 or Letter, portrait.

Model: bare copper wire 2.0 mm thick (about AWG 12), straight and in the clear. Resonant means the feed-point reactance is zero at the band centre; R is the feed-point resistance there. Average ground: relative permittivity 13, conductivity 0.005 S/m. Lengths are tip to tip.

Half-wave dipole: resonant length and impedance

BandCentre MHzFree space10 m over average ground
mft-inR Ωmft-inR ΩSWR 50 Ωheight in λ
160 m1.90576.76251′ 10¼″74.476.12249′ 9″38.31.300.06
80 m3.65040.01131′ 3″73.739.46129′ 5¾″52.81.060.12
60 m5.35927.2289′ 3½″73.426.8388′ 0½″68.11.360.18
40 m7.10020.5367′ 4¼″73.220.3166′ 7¾″80.61.610.24
30 m10.12514.3847′ 2″73.014.3947′ 2½″87.91.760.34
20 m14.17510.2633′ 7¾″72.810.3333′ 10¾″73.31.470.47
17 m18.1188.01626′ 3½″72.78.00826′ 3¼″63.81.280.60
15 m21.2256.83822′ 5¼″72.76.80222′ 3¾″70.41.410.71
12 m24.9405.81519′ 1″72.65.81019′ 0¾″78.71.570.83
10 m28.8505.02316′ 5¾″72.55.04416′ 6½″74.41.490.96
6 m52.0002.7789′ 1¼″72.42.7719′ 1″71.91.441.73

Total length of a horizontal dipole fed in the centre; each leg is half of it (the centre insulator is not modelled). SWR is at the feed point on 50 Ω coax through a 1:1 balun, without line loss. At 10 m a 160 m or 80 m dipole is very low in wavelengths, so much of its resistance is loss in the ground, not radiation.

The 468/f rule against the model

BandCentre MHz468 / f(MHz)NEC2 length is longer byA 468/f wire resonates at, MHzBand edges MHz
ftm (142.65 / f)free space10 m highfree space10 m high
160 m1.905245.774.88+2.5 %+1.7 %1.9531.9361.810–2.000
80 m3.650128.239.08+2.4 %+1.0 %3.7363.6853.500–3.800
60 m5.35987.3326.62+2.3 %+0.8 %5.480*5.402*5.3515–5.3665
40 m7.10065.9220.09+2.2 %+1.1 %7.254*7.1807.000–7.200
30 m10.12546.2214.09+2.0 %+2.1 %10.332*10.348*10.100–10.150
20 m14.17533.0210.06+1.9 %+2.7 %14.446*14.549*14.000–14.350
17 m18.11825.837.873+1.8 %+1.7 %18.446*18.414*18.068–18.168
15 m21.22522.056.721+1.7 %+1.2 %21.594*21.480*21.000–21.450
12 m24.94018.775.720+1.7 %+1.6 %25.354*25.351*24.890–24.990
10 m28.85016.224.944+1.6 %+2.0 %29.30829.42728.000–29.700
6 m52.0009.002.743+1.3 %+1.0 %52.65052.52550.000–54.000

* Outside the IARU Region 1 band edges.

Why they differ. 468 / f(MHz) feet (= 142.65 / f metres) is 0.952 × λ/2. In the model, a bare 2 mm wire in free space resonates at 0.964–0.976 × λ/2, so the modelled length is 1.3–2.5 % longer than the rule, and a wire cut to the rule resonates about that much higher (on 20 m: 14.446 MHz in free space instead of 14.175 MHz). How much shorter than λ/2 a dipole is depends on how thick the conductor is compared with the wavelength (table below). Height matters too: in the model, at 10 m the dipole is shorter than in free space on 160 m, 80 m, 60 m, 40 m, 17 m, 15 m, 12 m, 6 m and longer on 30 m, 20 m, 10 m; the effect of the ground swings with the height in wavelengths. Real antennas add things this model leaves out: insulation on the wire, end insulators and the wire looped through them, sag or an inverted-V shape, and nearby objects. These usually make the resonant length shorter, towards 468/f. NEC2 cannot model wire insulation.

Trimming. Resonance moves about inversely with length: 1 % shorter raises the resonant frequency by about 1 % (on 40 m, 1 % of 20.31 m is 20 cm, or 10 cm per leg, for about 71 kHz). Start a little long, measure where the SWR is lowest, and shorten both legs equally.

Quarter-wave vertical with 4 radials (ground plane)

BandCentre MHzFree spaceBase 3 m over average ground234 / f(MHz) ft rule
mft-inR Ωmft-inR ΩSWR 50 Ωm (71.32 / f)NEC2 (3 m) longer by
160 m1.90539.20128′ 7¼″23.639.08128′ 2½″38.21.3137.44+4.4 %
80 m3.65020.4667′ 1½″23.320.4867′ 2¼″35.61.4119.54+4.8 %
60 m5.35913.9345′ 8½″23.213.9745′ 10″33.01.5113.31+5.0 %
40 m7.10010.5234′ 6″23.110.5634′ 7½″30.71.6310.05+5.1 %
30 m10.1257.37224′ 2¼″23.17.40624′ 3½″27.61.817.044+5.1 %
20 m14.1755.26517′ 3¼″23.15.28817′ 4¼″24.92.005.032+5.1 %
17 m18.1184.11813′ 6¼″23.14.13313′ 6¾″23.22.153.937+5.0 %
15 m21.2253.51511′ 6½″23.13.52511′ 6¾″22.32.243.360+4.9 %
12 m24.9402.9919′ 9¾″23.12.9969′ 10″21.72.312.860+4.8 %
10 m28.8502.5858′ 5¾″23.12.5878′ 5¾″21.52.332.472+4.6 %
6 m52.0001.4334′ 8½″23.31.4324′ 8¼″23.62.111.372+4.4 %

The length is the vertical radiator and each of the 4 radials (all the same length), fed at the base between the radiator and the radials. The radials are straight and horizontal at 90° to each other. With horizontal radials the feed-point resistance is low (about 23.1 Ω in free space), so on 50 Ω coax the SWR is about 2:1 without a matching network (see the SWR column). On the low bands the nearby lossy ground raises R; part of that is loss, not radiation. Radials lying on or buried in the ground behave differently and are not covered here: NEC2 cannot model wires on or in the ground. 234 / f(MHz) feet (= 71.32 / f metres) is half of 468/f, 0.952 × λ/4; with 4 horizontal radials the model resonates close to a full λ/4, so the rule comes out about 5 % short here.

Wire thickness: 20 m dipole in free space at 14.175 MHz

Diameter mm≈ gaugeλ / dLength mft-ink = length / (λ/2)R Ω
0.5AWG 2442 29910.2933′ 9¼″0.97375.3
1.0AWG 1821 14910.2833′ 8¾″0.97273.7
1.6AWG 1413 21810.2633′ 8″0.97173.0
2.0AWG 1210 57510.2633′ 7¾″0.97072.8
3.0AWG 97 05010.2433′ 7¼″0.96872.5
6.0AWG 33 52510.2033′ 5¾″0.96572.2
12.0tube1 76210.1633′ 3¾″0.96072.1
25.0tube84610.0933′ 1¼″0.95472.0

A thicker conductor resonates shorter. The 2.0 mm wire used in the tables above gives k = 0.970 on 20 m; the 468/f rule (k = 0.952) matches only a conductor about 25 mm thick (k = 0.954) on this band. The effect depends on λ / d, so a given wire is relatively thicker, and shorter, on the higher bands. Gauge: nearest AWG from the wire gauge calculator.

Method and assumptions

Sources (checked 2026-11-04):

Ohmsums. Computer model (NEC2), not a measurement; trim to resonance.