Antennas · Volume 5
Transmission Lines & Feedlines
Coax types (RG-58 / RG-8X / LMR-240/400/600 / RG-213), ladder line and open-wire, loss per 100 ft per band, velocity factor, connectors (SMA / N / BNC / TNC / UHF / F / RP-SMA), adapters, the cost of bad coax
Contents
1. About this volume
The feedline is the silent killer of antenna performance. A 0.5 dB error in antenna gain is invisible to operators; a 6 dB loss in coax run from shack to roof is the difference between a strong link and a marginal one. The same operator who agonizes over 0.5 dB of dipole optimization regularly accepts 4-8 dB of unnecessary coax loss because “RG-58 is cheap.” This volume is the antidote.
The four-and-a-half things every operator needs from a feedline:
- Get the signal from the radio to the antenna with minimal loss.
- Don’t radiate from the feedline itself (common-mode currents = unintended antenna).
- Maintain a clean 50 Ω characteristic impedance so the SWR seen at the rig matches the SWR at the antenna.
- Survive the weather (UV, water ingress, ice, mechanical fatigue at terminations).
The “half” is connectors — they’re not feedline, but every feedline run terminates in them, and bad connectors cause 80% of feedline failures observed at any club station’s repair bench.
This volume cross-references with Vol 16 (BALUNs) for the choke-balun cure to common-mode currents, Vol 20 (Grounding) for lightning protection at the feedline entry, and Vol 22 (Weatherproofing) for connector sealing.
2. What a transmission line does — impedance, propagation, loss
A transmission line is a pair of conductors (or an equivalent guided-wave structure) that carries RF energy from one end to the other. The three characteristics that matter:
2.1 Characteristic impedance Z₀
The transmission line presents a particular impedance to RF flowing along it, set by its geometry and dielectric:
$$ Z_0 = \frac{60}{\sqrt{\varepsilon_r}} \cdot \ln!\left( \frac{D}{d} \right) \quad \text{(coax)} $$
$$ Z_0 = \frac{120}{\sqrt{\varepsilon_r}} \cdot \cosh^{-1}!\left( \frac{D}{2r} \right) \quad \text{(parallel-wire / twin-lead)} $$
For coax: D is the inner diameter of the shield, d is the diameter of the centre conductor, and ε_r is the dielectric constant of the insulator. For 50 Ω with solid polyethylene (ε_r ≈ 2.30), D/d ≈ 3.5; for 75 Ω with the same dielectric, D/d ≈ 6.5.
The “50 Ω” standard is a compromise — it’s near optimum for power handling, and close to (but not exactly) the impedance that minimizes voltage breakdown in air-core coax. The “75 Ω” standard (CATV, broadcast video) is closer to minimum loss for air-dielectric coax. Different optima, different communities, different standards that haven’t merged.
2.2 Propagation velocity (VF)
A wave on a transmission line travels at less than c by the velocity factor:
$$ v = \text{VF} \cdot c \quad \text{where VF} = \frac{1}{\sqrt{\varepsilon_r}} $$
For solid-polyethylene coax: VF ≈ 0.66. For foamed polyethylene: VF ≈ 0.78-0.85. For air-spaced hardline: VF ≈ 0.91-0.98. This was covered in Vol 2 §2.3; it shows up here because length-critical applications (matching stubs, phasing harnesses, beverage termination matching) require multiplying the electrical length by VF to get the physical length you cut.
2.3 Loss
Energy is lost along the line, dissipated as heat in two ways:
- Conductor loss: I²R heating in the inner conductor (mostly, because skin depth limits the current to a thin layer near the surface at high frequency). Scales as √f.
- Dielectric loss: the dielectric heats slightly each time the field oscillates. Scales as f (linear with frequency).
Total loss = conductor loss + dielectric loss. At low frequencies, conductor loss dominates; at high frequencies, dielectric loss dominates. The crossover is around 1-3 GHz for most plastic-dielectric coax; above this, foam-PE (lower dielectric loss) beats solid-PE.
Loss in coax is the fundamental number on every coax datasheet, reported as dB per 100 ft (or dB per 100 m) at a few specific frequencies. The full loss table is §5.
3. Coaxial cable — geometry, materials, characteristic impedance
A coax cable is a four-layer concentric structure:
Coax cross-section (RG-8 / LMR-400 style)
╔═══════════════════════════════════╗
║ Outer jacket ║ PE / PVC / TPE / FEP
║ ╔═══════════════════════════════╗║
║ ║ Shield (braid + foil + braid)║║ bare or tinned copper
║ ║ ╔═══════════════════════════╗║║
║ ║ ║ Dielectric (insulator) ║║║ PE / foam-PE / PTFE
║ ║ ║ ╔═════════════════════╗║║║
║ ║ ║ ║ Centre conductor ║║║║ solid Cu / CCS / CCA
║ ║ ║ ╚═════════════════════╝║║║
║ ║ ╚═══════════════════════════╝║║
║ ╚═══════════════════════════════╝║
╚═══════════════════════════════════╝3.1 Centre conductor
Three common materials:
- Solid copper — the standard. Highest conductivity; corrosion-resistant if jacketed; expensive in larger sizes.
- Copper-clad steel (CCS) — copper plating over a steel core. Cheaper, mechanically stronger, RF performance equivalent to solid copper above ~10 MHz (skin effect confines the current to the copper layer). Used in RG-6 / RG-59 cable-TV coax and in some budget RG-58/RG-8 lines.
- Copper-clad aluminum (CCA) — aluminum core with copper plating. Even cheaper; avoid for transmit applications. The aluminum core oxidizes at any termination cut, increasing loss over months. Used only in budget cable-TV / consumer applications.
For 50 Ω antenna work: solid copper for any TX application above ~50 W; CCS acceptable for receive-only and lower-power TX. Never CCA for any TX.
3.2 Dielectric
Four common types:
- Solid polyethylene (PE): ε_r ≈ 2.30, VF ≈ 0.66. Standard in RG-58, RG-213, RG-8. Robust, predictable, ~3-7× more dielectric loss than foam.
- Foam polyethylene (foam-PE): ε_r ≈ 1.40-1.60, VF ≈ 0.78-0.85. Used in LMR-series (Times Microwave) and Belden 9914. Lower loss, lower weight, slightly more crush-sensitive than solid.
- Polytetrafluoroethylene (PTFE / Teflon): ε_r ≈ 2.05, VF ≈ 0.70. Used in mil-spec / high-temperature applications. Very stable across temperature; expensive.
- Air (with periodic dielectric spacers): ε_r ≈ 1.05-1.10, VF ≈ 0.95-0.98. Used in heliax / hardline / satellite cable. Lowest loss, lowest dielectric heating; physically large and inflexible.
For amateur work: solid-PE coax (RG-58, RG-213) for budget runs; foam-PE LMR-series for any serious work; hardline for fixed runs over 100 ft at VHF/UHF.
3.3 Shield
The shield serves two purposes: it carries the return current of the differential signal and it electromagnetically shields the inner conductor from outside interference. Shield quality is rated by:
- Coverage — the percentage of the dielectric surface covered by shield strands. Single-braid: ~85-90%. Foil + braid: ~95-99%. Tri-shield (foil + braid + foil + braid): >99.9%.
- Material — bare copper (best conductivity), tinned copper (slight corrosion resistance, ~3% conductivity loss), aluminum foil (lighter, lower conductivity).
For HF: single braid is adequate (skin depth large at HF; coverage isn’t critical). For VHF/UHF: foil + braid is the standard (LMR-400, Belden 9913, RG-8X foam). For microwave / EMI work: tri-shield or solid-jacket hardline.
3.4 Jacket
The outer protective covering:
- PVC (polyvinyl chloride): cheap, flexible, UV-degrades in 2-5 years outdoors. Indoor / temporary use only.
- PE (polyethylene): UV-stable, mechanically robust, used in outdoor-rated RG-8 and RG-213.
- TPE (thermoplastic elastomer): used in LMR-exterior series, weather-resistant and flexible.
- FEP (fluorinated ethylene propylene): high temp, high cost, in mil-spec coax.
4. The coax catalogue — RG-58, RG-8X, LMR-240, LMR-400, LMR-600, RG-213, hardline
A summary table of the workhorse coax types in this series:
| Coax type | OD (mm) | Z₀ (Ω) | VF | Dielectric | Use case |
|---|---|---|---|---|---|
| RG-58 (A/U) | 4.95 | 50 | 0.66 | Solid PE | Patch cables, short jumpers — high loss, budget |
| RG-58 foam | 4.95 | 50 | 0.78 | Foam PE | Improved RG-58 alternative |
| RG-8X (Mini-8) | 6.0 | 50 | 0.75-0.84 | Foam PE | Mid-range portable, HF feedlines under 50 ft |
| RG-213/U | 10.3 | 50 | 0.66 | Solid PE | Heavy-duty HF, budget VHF runs |
| Belden 9913F7 | 10.3 | 50 | 0.84 | Solid PE / air | Better RG-213 alternative |
| LMR-240 | 6.1 | 50 | 0.84 | Foam PE | High-quality jumper, UHF short runs |
| LMR-400 | 10.3 | 50 | 0.85 | Foam PE | The standard high-quality outdoor coax through 2.4 GHz |
| LMR-600 | 14.9 | 50 | 0.87 | Foam PE | Long microwave runs, contest stations |
| LMR-900 / LMR-1200 | 22 / 30 | 50 | 0.87 | Foam PE | Very long runs, broadcasting, very high power |
| Heliax 1/2” | 13.8 | 50 | 0.88-0.91 | Air | Long microwave runs; very low loss; rigid |
| Heliax 7/8” | 22.2 | 50 | 0.88-0.91 | Air | Broadcasting, microwave links; very low loss |
| RG-59 | 6.15 | 75 | 0.66 | Solid PE | CATV; not for ham (75 Ω) |
| RG-6 | 7.0 | 75 | 0.83 | Foam PE | CATV / satellite; not for ham |
4.1 The “good enough” decision tree
For most amateur work:
- RG-58: only for jumpers <2 m. Long runs lose embarrassing amounts of signal.
- RG-8X: HF feedline runs under 15 m. Worse than LMR-240 by a few dB at VHF/UHF; tolerable at HF.
- LMR-240: high-quality jumper, NMO mobile, indoor patch up to 5 m at UHF.
- LMR-400: the workhorse. Pick this for almost any outdoor run from 5-50 m at any frequency through 2.4 GHz.
- LMR-600: long runs (>30 m) at VHF/UHF, or 2.4 GHz+ over 20 m. The diameter starts to limit which connectors you can use (UHF / N are fine; SMA is mechanically awkward).
- Hardline (1/2” Heliax): long fixed runs at HF (>50 m), where the bend radius doesn’t matter and you’ll terminate with proper hardline connectors (DIN-7/16 or N).
4.2 What to absolutely avoid
- Copper-clad aluminum (CCA) coax in any TX application.
- Unjacketed shield-only “wire” sold as RG-58 — there’s no inner-conductor insulation, just a thin layer at the connection points.
- CATV RG-6 / RG-59 for ham work — it’s 75 Ω; the impedance mismatch creates SWR even into a perfect antenna.
- Solid-shield “BNC video” coax at HF — usually too thin a centre conductor, high loss.
- Twin-lead masquerading as coax — flat parallel-wire cable with shielding around both conductors. It exists; it’s not 50 Ω.
5. Loss per 100 ft per band — the canonical table
Loss is the most important number for any feedline decision. Below is the canonical table for the seven workhorse types in §4, compiled from Times Microwave and Belden datasheets:
| Coax type | 1.8 MHz | 7 MHz | 14 MHz | 28 MHz | 50 MHz | 144 MHz | 432 MHz | 1296 MHz | 2400 MHz | 5800 MHz |
|---|---|---|---|---|---|---|---|---|---|---|
| RG-58 (solid) | 0.4 | 1.0 | 1.5 | 2.0 | 2.5 | 4.5 | 8.9 | 16 | 22 | 38 |
| RG-58 foam | 0.4 | 0.9 | 1.3 | 1.8 | 2.3 | 3.9 | 7.5 | 14 | 19 | 33 |
| RG-8X | 0.3 | 0.5 | 0.8 | 1.1 | 1.7 | 3.6 | 6.5 | 11 | 14 | 26 |
| RG-213 | 0.2 | 0.3 | 0.5 | 0.8 | 1.5 | 2.6 | 4.8 | 8.3 | 12 | 21 |
| Belden 9913 | 0.1 | 0.2 | 0.3 | 0.5 | 1.0 | 1.8 | 3.2 | 5.6 | 7.8 | 14 |
| LMR-240 | 0.2 | 0.3 | 0.5 | 0.8 | 1.1 | 1.9 | 3.4 | 6.0 | 8.4 | 14 |
| LMR-400 | 0.1 | 0.2 | 0.4 | 0.5 | 0.8 | 1.3 | 2.2 | 3.9 | 5.4 | 8.9 |
| LMR-600 | 0.07 | 0.15 | 0.2 | 0.3 | 0.5 | 0.8 | 1.4 | 2.5 | 3.5 | 5.9 |
| Heliax 1/2” | 0.04 | 0.07 | 0.10 | 0.15 | 0.20 | 0.4 | 0.7 | 1.2 | 1.7 | 2.6 |
| Heliax 7/8” | 0.02 | 0.05 | 0.07 | 0.10 | 0.15 | 0.25 | 0.45 | 0.80 | 1.15 | 1.85 |
All values: dB per 100 ft. For dB per metre, divide by 30.48; per 100 m, multiply by 3.28.
5.1 Loss interpolation
For frequencies not in the table, the linear-in-√f rule works to within 5% across the range:
$$ L(f_2) \approx L(f_1) \cdot \sqrt{\frac{f_2}{f_1}} $$
This is the conductor-loss approximation, which dominates below ~3 GHz for plastic-dielectric coax. Above 3 GHz, dielectric loss matters more and the actual loss climbs slightly faster than √f.
5.2 The loss-vs-frequency intuition
Two important patterns visible in the table:
- Loss at 7 MHz is one-fifth to one-eighth of loss at 144 MHz for the same coax. HF work tolerates much longer runs of the same coax type than VHF work.
- At 2.4 GHz, even “good” coax (LMR-400) loses 5.4 dB per 100 ft. A 30 m (100 ft) cable run dissipates 70% of input power; only 30% reaches the antenna. This is why Wi-Fi installations use mast-mounted access points instead of long feedlines.
For any link budget, feedline loss enters twice on a one-way link (TX feedline + RX feedline) and four times on a round-trip transceiver path. A 50 ft (15 m) LMR-400 run at 2.4 GHz is 2.7 dB one-way, 5.4 dB round-trip — the difference between a usable Wi-Fi link and a marginal one.
5.3 The feedline-as-fuse insight
A bit of trivia: lossy coax masks high SWR at the rig end. A 6 dB lossy run between rig and antenna shows a quarter of the actual mismatch at the rig (since reflected power passes through the same loss again going back). The rig “sees” SWR 1.5:1 when the antenna actually has SWR 5:1 because 4 dB of round-trip loss has flattened it.
This is exactly the wrong outcome for actual performance — radiated power is reduced by the round-trip loss — but it does provide a soft “feedline-as-fuse” against transmitter damage from antenna failures (open / short connector at the antenna becomes a 6 dB lossy match instead of a 0 dB short at the rig).
6. Velocity factor and length-critical applications
Three common amateur applications require physical coax lengths that are specific fractions of an electrical wavelength. In each case, you must correct for the velocity factor:
6.1 Quarter-wave matching stubs
A coax stub λ/4 long and Z₀_stub between the rig and a load Z_L transforms the impedance:
$$ Z_{\text{in}} = \frac{Z_0^2}{Z_L} $$
Used to transform 25 Ω (a Yagi driven element with hairpin match) to 50 Ω via a λ/4 of 35 Ω coax (= 50 × √(25/100)). The 35 Ω coax doesn’t exist as a stock product; in practice you use two parallel runs of 70 Ω coax, which gives 35 Ω effective.
Lengths for stubs at the canonical bands:
| Band | f (MHz) | λ/4 air (m) | λ/4 in RG-58 VF 0.66 (m) | λ/4 in LMR-400 VF 0.85 (m) |
|---|---|---|---|---|
| 2 m | 146 | 0.514 | 0.339 | 0.437 |
| 70 cm | 432 | 0.174 | 0.115 | 0.148 |
| 23 cm | 1296 | 0.058 | 0.038 | 0.049 |
Get the VF correction wrong and the stub doesn’t transform — the match fails.
6.2 Phasing harnesses
For phased verticals or stacked Yagis, antenna elements need feedlines that are specific electrical fractions of a wavelength different in length. The classic 4-square vertical phasing system uses 84° and 71° electrical lengths between elements; physical lengths must be VF-corrected.
For an 80 m four-square (3.75 MHz):
- 84° electrical = 84/360 × λ = 0.233 × 80 m = 18.66 m electrical
- Physical: 18.66 × 0.66 (RG-213) = 12.32 m
Cutting the physical length to 12.32 m gives the correct phasing. A 6 cm error (a typical measurement tolerance) gives a 0.5% phase error, fine for amateur work; a 30 cm error gives 2.4% phase error, which moves the radiation pattern noticeably.
6.3 Beverage termination feedlines
A beverage’s far-end terminating resistor (typically 470 Ω) needs a clean transmission line to its mounting point. Often a small section of 75 Ω TV cable is used because it’s a closer match to the 470 Ω termination through a 9:1 UNUN. The physical length isn’t critical (since the beverage is non-resonant) but the impedance match at the termination matters.
6.4 The “I cut this too short” recovery
If a length-critical run is too short:
- For matching stubs: insert a
λ/2(electrical) section of coax to extend the stub by exactlyλ/2without changing its transformation properties. Theλ/2insert is a “neutral” extension. - For phasing harnesses: re-cut. Phasing tolerance is a few degrees; you can’t fudge.
- For beverage terminations: re-cut; the termination needs to be at the physical end of the beverage to absorb the back-wave correctly.
7. Ladder line, open-wire, twin-lead — when balanced feedline wins
Coax is the universal modern feedline, but balanced feedline (twin-lead, ladder line, open-wire) has one massive advantage: dramatically lower loss at HF.
7.1 The loss math
Open-wire / ladder line is mostly air between the conductors — air has essentially zero dielectric loss. The conductor losses are the only meaningful loss term, and at HF they’re small.
A representative loss table for balanced feedline:
| Feedline type | Z₀ | VF | 1.8 MHz | 7 MHz | 14 MHz | 28 MHz | 50 MHz |
|---|---|---|---|---|---|---|---|
| 600 Ω open-wire | 600 | 0.96 | 0.02 | 0.04 | 0.06 | 0.10 | 0.14 |
| 450 Ω windowed ladder | 450 | 0.91 | 0.04 | 0.08 | 0.12 | 0.18 | 0.27 |
| 300 Ω twin-lead (Ham-O) | 300 | 0.82 | 0.10 | 0.20 | 0.30 | 0.45 | 0.70 |
(values: dB/100 ft)
Order of magnitude lower than coax at HF. Where LMR-400 at 14 MHz is 0.4 dB/100 ft, 600 Ω open-wire is 0.06 dB/100 ft — 7× lower loss for the same length.
7.2 Where balanced feedline wins
Three specific cases:
- Long HF runs from a tuner-fed doublet (cross-link to Vol 7 §6). A 100 ft run of 450 Ω ladder line costs 0.4 dB at 14 MHz; the same run in LMR-400 costs 0.4 dB — but the ladder is cheaper and doesn’t need a BALUN at the antenna feedpoint.
- High-SWR feedlines — a balanced feedline carrying a 5:1 SWR loses far less power to the mismatch than coax carrying the same SWR. Where coax round-trip loss compounds the SWR penalty, balanced feedline mostly just carries the SWR onward.
- Multi-band antennas with tuners — the doublet + ladder + balanced tuner (Vol 17 §10) is the most efficient multi-band HF antenna architecture.
7.3 Where balanced feedline loses
Three reasons most operators use coax instead:
- Balanced lines radiate if asymmetry develops. A nearby metal object, a kinked routing, even an asymmetric guy point can unbalance the feedline and turn it into an unintended antenna.
- Balanced lines can’t enter the house cleanly. Coax has a shield that’s grounded; balanced lines need clear space and a 4:1 BALUN at the rig end to convert to coax for the last few feet.
- Connectors don’t exist in any standardized form for balanced feedline at amateur prices. Connection is mechanical (terminal posts or eyelets), not a quick connector.
For most modern stations, coax wins on convenience. Balanced feedline is the choice for the operator who has an outdoor multi-band antenna farm, who’s willing to take the time to route the ladder line cleanly, and who wants the lowest possible feedline loss on HF.
7.4 The 4:1 BALUN at the rig end
A doublet fed by ladder line into the shack needs conversion to 50 Ω coax at the rig end. The 4:1 current BALUN (Vol 16 §6) is the standard:
- Antenna → 100-200 ft of 450 Ω ladder line → 4:1 BALUN at the shack wall → 5 ft of 50 Ω coax → tuner → rig.
- The 4:1 BALUN converts ~200 Ω balanced to 50 Ω unbalanced; the tuner handles the residual mismatch over the bands of interest.
8. Connectors — SMA, N, BNC, TNC, UHF, F, RP-SMA, the maximum frequency of each
Connectors are the most underrated component in any feedline system. They’re cheap, they fail in subtle ways, and the wrong choice imposes losses or impedance discontinuities that ruin an otherwise well-engineered link.
8.1 The standard connector catalogue
| Connector | Threaded | Max f (real) | Impedance | Mating cycles | Use case |
|---|---|---|---|---|---|
| UHF (PL-259/SO-239) | Yes | ~300 MHz | not constant | ~500 | HF-only; mechanically robust, RF-poor |
| BNC | Bayonet | ~4 GHz | 50 Ω | 500-1000 | Test gear, ham VHF, scope probes |
| TNC | Yes | ~11 GHz | 50 Ω | 500-1000 | Outdoor mil/aero — BNC but threaded |
| N | Yes | ~11 GHz | 50 Ω | 500-1000 | Outdoor durable, microwave standard |
| SMA | Yes | ~18 GHz (real ~6 GHz) | 50 Ω | ~500 | Test gear, SDRs, NanoVNA |
| RP-SMA (reverse) | Yes | ~18 GHz | 50 Ω | ~500 | Wi-Fi / 802.11 mandated for consumer gear |
| F | Yes | ~2 GHz | 75 Ω | ~250 | Cable TV, SAT — not for 50 Ω ham |
| MCX | Push-on | ~6 GHz | 50 Ω | ~500 | Subminiature, some SDR dongles |
| MMCX | Push-on | ~6 GHz | 50 Ω | ~500 | Subminiature, RTL-SDR Blog v3 / v4 |
| 7/16 DIN | Yes | ~7 GHz | 50 Ω | >1000 | Broadcast, high-power |
| 4.3-10 (Mini-DIN) | Yes | ~10 GHz | 50 Ω | >1000 | Cellular base stations, low-PIM |
8.2 The PL-259 / SO-239 (UHF) gotcha
The PL-259 plug + SO-239 jack (collectively “UHF connectors” — the name is a 1930s artifact, before UHF was even a band designation) are the standard ham connectors at HF and 2 m. They’re cheap, robust, easy to install, and not a constant-impedance design.
Specifically: the PL-259’s geometry doesn’t maintain 50 Ω through the connector body. Above ~300 MHz, the impedance discontinuity creates SWR of 1.2-1.4:1 per mated pair, and insertion loss climbs to 0.5-1.0 dB per pair at 432 MHz.
The right connectors for VHF/UHF/microwave amateur work are N (outdoor) or BNC (indoor). PL-259/SO-239 belongs on HF rigs and below.
8.3 RP-SMA and the Wi-Fi mandate
Reverse-Polarity SMA (RP-SMA) reverses the male/female assignment of the centre pin: the plug side has a socket; the jack side has a pin. Electrically identical to SMA otherwise.
Why it exists: the US FCC (in the 2.4 / 5 GHz consumer rules) mandates that consumer Wi-Fi equipment must not have standard antenna connectors, specifically to discourage users from replacing internal antennas with high-gain external antennas. RP-SMA is the workaround the industry adopted — equipment ships with RP-SMA jacks, antennas come with RP-SMA plugs. The technical effect is zero (the connector is fully mateable with adapters); the legal effect is that the FCC-required certification is preserved.
Practical implication: keep an SMA ↔ RP-SMA adapter in the Wi-Fi tool bag. You will need it.
8.4 The mil-spec hierarchy
For high-reliability or long-life installations:
- TNC is a threaded BNC. Performs the same as BNC up to 11 GHz, more reliable mechanically.
- N connector is the workhorse outdoor 50 Ω connector. Excellent SWR through 11 GHz; weather-resistant.
- 7/16 DIN (also called L29) is the broadcast / high-power standard. Used on commercial transmitters at 500 W+; physically large.
- 4.3-10 is the modern cellular-base-station connector. Replaces 7/16 DIN in new installations; smaller footprint, similar performance, lower passive intermodulation (PIM) — important for cellular cell sites where co-located TX and RX must not generate intermod products.
8.5 Practical connector hygiene
A few tactical rules:
- Always use the right tool to install crimp connectors. The crimp tooling sets the geometry; eyeballing it produces a connector that works mechanically but has 0.5-2 dB extra loss.
- Solder connections at the centre pin should reach the centre pin’s full base. A cold solder joint at the centre pin is a high-Z point that creates SWR.
- Tighten connectors to the manufacturer’s torque spec (typically 0.6-1.0 N·m for SMA, 1.5-3 N·m for N). Over-tightening damages the threading; under-tightening lets the connector loosen with vibration.
- Don’t mix connector types inside a single run (don’t go SMA → UHF → N → SMA). Every transition is a discontinuity.
9. Adapters — the loss budget of a “no-loss” adapter
Adapters between connector types — SMA-to-N, N-to-UHF, BNC-to-SMA — are convenient and ubiquitous. They are also lossy.
9.1 Typical adapter loss
A high-quality connectorized adapter (manufactured by reputable vendors: Amphenol, Pasternack, Mini-Circuits, Amphenol RF):
| Adapter type | At 1 GHz | At 2.4 GHz | At 5.8 GHz |
|---|---|---|---|
| SMA-male to N-female | 0.05 dB | 0.10 dB | 0.20 dB |
| SMA to SMA (gender adapter) | 0.05 | 0.10 | 0.15 |
| N to UHF | 0.20 | 0.50 | not useful above 1 GHz |
| BNC to N | 0.10 | 0.25 | 0.50 |
A cheap eBay adapter (no published specs):
| Adapter type | At 1 GHz | At 2.4 GHz | At 5.8 GHz |
|---|---|---|---|
| SMA-male to N-female | 0.3-0.6 | 0.5-1.0 | 1.0-2.0 |
| SMA to SMA | 0.1-0.3 | 0.3-0.7 | 0.5-1.5 |
| N to UHF | 0.4-1.0 | 0.8-2.0 | not useful above 1 GHz |
The premium-vs-budget gap widens at higher frequencies. At HF, a cheap adapter is fine. At 5 GHz, the difference between a premium and a budget adapter is several dB.
9.2 Cumulative loss in long chains
In a measurement setup with multiple adapters:
- HackRF → SMA cable → SMA-to-N adapter → 1 m LMR-240 → N-to-N → 5 m LMR-400 → N-to-SMA → antenna SMA → ANT
- At 2.4 GHz: 0.5 (cable) + 0.1 + 0.55 (cable) + 0.1 + 2.7 (cable) + 0.1 + 0.5 (cable) = 4.55 dB of loss.
Three adapters at 0.1 dB each = 0.3 dB. Three cheap adapters at 0.5-1.0 dB each = 1.5-3.0 dB. The cumulative adapter loss is often comparable to the cable loss.
9.3 The right approach to connectors and cables
Terminate cables in the connectors you need. The N-female on the antenna side, SMA-male on the HackRF side, no adapters. Each cable is purpose-built; no field-improvisation budgets are blown by adapter loss.
This is more work upfront (you need to know what connector each end requires before cutting the cable) but pays off in every measurement and every link-budget calculation. The labour cost of terminating a custom-length LMR-400 cable with N-female on one end and SMA-male on the other is 15-30 minutes; the cumulative loss savings over months of measurement and operation are large.
10. Common-mode currents, feedline radiation, and the case for a choke balun
The single most common feedline problem isn’t loss — it’s common-mode currents: RF flowing on the outside of the coax shield, making the coax itself part of the antenna.
10.1 Why common-mode currents happen
A coax cable carries RF as a differential signal: equal-and-opposite currents on the centre conductor and the inner surface of the shield. The outer surface of the shield, in principle, carries no current.
In reality, current on the outer shield surface can be induced by:
- Antenna asymmetry: feeding a balanced antenna (dipole) directly from coax means the dipole’s two legs see different impedances (one to the coax centre, one to the coax shield), and the imbalance forces current onto the shield’s exterior.
- Coax routing too close to the antenna: an antenna’s near-field couples into the coax shield, inducing current.
- Ground proximity: a vertical antenna’s ground plane (radials) drives current down the coax shield’s exterior.
The result: the coax becomes part of the antenna. The antenna pattern is distorted, the rig sees an unexpected impedance, and the feedline radiates into the shack.
10.2 Symptoms of common-mode current
The classic symptoms:
- “RF in the shack”: touching the rig changes the SWR reading. A laptop near the rig crashes when the operator transmits. Strange buzzing on the headphones.
- Pattern distortion: a dipole that’s supposed to be omnidirectional (vertical orientation) has nulls in specific directions because the coax shield is contributing to the pattern.
- Mysterious SWR drift: SWR changes as the operator moves the radio across the desk because the operator’s body capacitance shifts the coax-shield resonance.
- Bursts of RFI to neighbors: the coax shield radiates a few mW unintentionally, enough to disrupt TV, audio, baby monitors, etc. in adjacent buildings.
10.3 The cure: choke balun at the feedpoint
A choke balun is a high-impedance device installed on the coax shield that blocks common-mode current while passing the differential RF signal. Two common implementations:
Beaded choke (the W2DU classic): 50-100 ferrite beads (Fair-Rite 2643-002701, Mix 43) slid over the coax. The beads’ magnetic permeability creates a high inductance on the shield’s exterior. Common-mode impedance: 500 Ω - 2 kΩ across HF.
Wound coil choke: 10-12 turns of the coax around a large ferrite core (FT240-31 for HF). Higher common-mode impedance than the beaded choke; more compact; mechanically tighter.
Both are detailed in Vol 16 §17. The beaded choke is the operator-friendly choice: install in 10 minutes with no tools, transparent operation, $30 of beads protects 30 ft of coax.
10.4 Where common-mode current matters most
- Coax-fed dipoles without a center BALUN — the classic symptom case. The fix: 1:1 current BALUN at the dipole feedpoint (Vol 16 §5).
- EFHW antennas — the high-Z feedpoint at the end of the wire couples strongly to the coax shield. Common-mode choke 0.05λ from the UNUN is essential (Vol 10 §7).
- Verticals fed from the base — ground-mounted verticals’ ground systems can drive significant current onto the coax shield. Common-mode choke at the rig-end of the coax cures this.
- Wi-Fi installations — usually not a problem because of short cable runs and high frequencies, but Mini-Circuits offers SMA-to-SMA chokes for the rare case.
10.5 The “I have a perfect SWR but the antenna doesn’t work” diagnostic
If your SWR is perfect but the antenna doesn’t perform as predicted:
- Disconnect the coax from the rig at the shack end.
- Wave a hand near the disconnected coax. If the rig’s S-meter twitches when you bring your hand close to the loose coax shield, the coax is radiating — common-mode currents are happening.
- Install a choke balun (W2DU style) and re-test.
A choke balun cures more antenna problems than any other single intervention.
11. Resources
- Times Microwave datasheets (LMR series): https://www.timesmicrowave.com
- Belden RG-series datasheets: https://www.belden.com
- Andrew (CommScope) Heliax datasheets: https://www.commscope.com
- ARRL Antenna Book Ch. 24 (transmission lines)
- Sevick, Transmission Line Transformers (the W2FMI classic, 4th ed.) — the canonical text on coax-and-ferrite matching transformers
- W2DU beaded-choke design — QST, 1973 and following; archived at the ARRL website
- Pozar, Microwave Engineering (4th ed.), Ch. 2 (transmission line theory)
- Cebik, W4RNL papers on common-mode currents (archived at antenneX)
- Maxim Integrated app note 4081 — “Coaxial Cable RF Transmission Lines”
- QST magazine “Coax Test Reports” series — measured loss data for many coax types