High-performance small transmitting loop antennas for restricted spaces — physics, construction, RF management, and automation for 400W+ operation.
Compelling performance from a small footprint — with caveats that demand careful engineering at QRO power levels.
Magnetic loop antennas (MLAs), also called small transmitting loops (STLs), offer a compelling option for amateurs operating from restricted sites — flats, apartments, HOA-restricted properties, or where a discrete antenna profile is required. Unlike electrically small dipoles that suffer severe efficiency penalties, a properly constructed MLA achieves respectable radiation efficiency through an extremely high-Q resonant circuit that concentrates energy in the magnetic field component.
The key trade-off is bandwidth. Q factors of 300–1500 are common, giving 2:1 VSWR bandwidths of 3–15 kHz on 40m. This demands precise, motorised remote tuning. In return, the same narrow bandwidth provides exceptional noise rejection — a significant receive advantage in urban RF environments, where band noise can be 20–30 dB worse than at a rural site.
This guide specifically addresses construction and operation at 400W and above. This places hard demands on every component, particularly the tuning capacitor. Most commercial MLAs top out at 25–150W; QRO operation requires thoughtful engineering decisions throughout the design.
At 100W on 7 MHz with a typical Q of 500, capacitor peak voltages exceed 6,800V. At 400W this approaches 14 kV peak. Vacuum variable capacitors are mandatory for 400W+ operation — air variable types will arc and be destroyed. Treat all QRO MLA construction as high-voltage electrical work. Never make adjustments to the capacitor or coupling loop with power applied.
Understanding the underlying physics is essential for making sensible engineering choices in loop construction.
The radiation resistance of a small loop scales with the fourth power of frequency and the square of loop area:
This steep frequency dependence is why MLA performance degrades sharply on lower bands. A 1m diameter loop is marginal on 80m and essentially unusable on 160m for transmit. For 40m operation, a minimum of 1.0m diameter is recommended — larger is always better within the λ/10 constraint.
Loop efficiency is simply the fraction of total resistance that is radiation resistance. The rest is loss:
Even 15% efficiency is only −8.2 dB below 100% — roughly 1.4 S-units. In an urban environment where ambient noise is 20 dB above the rural baseline, a loop that is 10 dB down on signal but 15 dB down on noise results in a net signal-to-noise improvement vs a full-size antenna. This is the practical case for urban MLA operation.
The impedance bandwidth (VSWR ≤ 2) is the retuning interval for transmit. The noise receive bandwidth is approximately twice as wide. When back-calculating efficiency from measured bandwidth, you must use the correct value for the formula. Using the wrong one gives efficiency figures that are 6 dB (one S-unit) over-optimistic — one of DK3BA's core findings.
Larger diameter tube always outperforms smaller diameter, but with rapidly diminishing returns. Going from 15mm to 22mm copper tube saves ~1.5 dB; from 22mm to 54mm saves only ~0.6 dB further. 22mm (¾") or 28mm (1") copper plumbing tube is the optimal practical compromise for a fixed indoor/rooftop installation. Silver-plate all connections — contact resistance is a dominant loss mechanism when RR is only 0.1–0.5 Ω.
Manfred Mornhinweg DK3BA published a thorough analysis showing that widely-cited MLA efficiencies are systematically overstated by approximately one S-unit. His key points:
Capacitor selection rule: Choose a vacuum variable capacitor rated for ≥ 1.5× the calculated peak voltage. For 400W on 40m, minimum 20 kV peak rating. Most ex-Soviet military КПВ types are rated 10–25 kV — check the datasheet for your specific unit.
Practical efficiency measurement: Connect a NanoVNA, tune for lowest SWR, then measure the −3 dB bandwidth. Enter the measured Q (= f/BW) and loop dimensions into Owen Duffy's calculator to derive real-world efficiency. This method accounts for all real losses and environmental effects.
Loop dimensions, band coverage, conductor choices, and online tools.
To maintain magnetic loop behaviour (and avoid the antenna acting as a resonant full-wave loop), the perimeter must remain at or below λ/10:
A 1.0–1.3m diameter loop is the practical optimum for 40m fixed installations, balancing efficiency, physical size, and structural feasibility. Larger is always better within the constraint.
| Band | Freq (MHz) | λ (m) | Max Perimeter (m) | Max Diameter (m) | Practical Choice |
|---|---|---|---|---|---|
| 160m | 1.85 | 162 | 16.2 | 5.2 | Impractical |
| 80m | 3.6 | 83 | 8.3 | 2.6 | 2.0m (very low η) |
| 60m | 5.3 | 56.6 | 5.66 | 1.8 | 1.5–1.8m |
| 40m | 7.1 | 42.2 | 4.22 | 1.34 | 1.0–1.3m ✓ |
| 30m | 10.1 | 29.7 | 2.97 | 0.94 | 0.9–1.0m ✓ |
| 20m | 14.2 | 21.1 | 2.11 | 0.67 | 0.6–0.9m ✓ |
| 17m | 18.1 | 16.6 | 1.66 | 0.53 | 0.5–0.9m ✓ |
| 15m | 21.2 | 14.1 | 1.41 | 0.45 | 0.5–0.9m ✓ |
| 10m | 28.5 | 10.5 | 1.05 | 0.33 | 0.5–0.9m ✓ |
Multi-band coverage tip: A 1.0m diameter loop with a 15–1000 pF vacuum variable capacitor covers approximately 40m through 10m. To add 80m, a separate 2.0m diameter loop is more practical than enlarging a single loop — the efficiency difference is significant.
Best practical choice. 22mm (¾") or 28mm (1") plumbing tube is widely available, easily bent, excellent conductivity. Use 1.2mm+ wall thickness for structural rigidity. Silver-plate all joints and capacitor terminals. Avoid flux residue — clean thoroughly after soldering.
50–100mm wide copper strap offers maximum surface area for a given weight. More flexible than tube. Popular for portable designs. High contact resistance at joins unless carefully clamped or silver-soldered. More prone to RF current crowding at edges.
~60% the conductivity of copper but significantly lighter. Useful where weight matters. Cannot be soldered — requires compression fittings or stainless clamps. Remove anodising at all RF contact points. Efficiency ~0.5–1 dB worse than copper for same dimensions.
The outer braid carries RF current. Inner conductor shorted at one end. Low cost, very flexible. Efficiency roughly 2–3 dB worse than copper tube but adequate for 20m–10m. Limited to QRP/low power — coax outer braid contact resistance rises rapidly at QRO.
Comprehensive — inductance, required capacitance, Q, efficiency, and bandwidth. The most widely used tool for HF loop design.
Classic Steve Yates calculator. Radiation resistance, efficiency, bandwidth, and coupling loop sizing. Good for verifying VK3CPU results.
Detailed small transmitting loop output including voltage across capacitor — critical sanity-check for QRO power levels.
Enter your measured Q (from NanoVNA) and loop dimensions to back-calculate real-world efficiency. The correct way to characterise a built antenna.
Padding capacitor combinations to reduce voltage across an air variable — useful for understanding air-cap limits and the physics of why vacuum wins at QRO.
The single most important design decision for any loop rated above 150W.
MLA schematic — Faraday coupling loop (blue, ≈D/5 diameter) inside main loop (green). Vacuum variable capacitor at bottom. RF choke toroid at coax connection point.
| Property | Air Variable Capacitor | Vacuum Variable Capacitor |
|---|---|---|
| Power handling (40m) | 20–50W before arcing | 500W–2 kW (type dependent) |
| Peak voltage rating | 1.5–3 kV typical | 5–25 kV standard |
| ESR / contact loss | Moderate — plate contacts | Extremely low — 0.005–0.02 Ω |
| Capacitance ratio (max/min) | ~5:1 to 10:1 (butterfly) | 50:1 to 100:1 common |
| Multi-band coverage | Limited — narrow ratio | Excellent — 40m–10m from one unit |
| Sensitivity to humidity | High — arcing risk outdoors | None — hermetically sealed vacuum |
| Motor torque required | Low (light plates) | Higher — needs gear reduction |
| Cost | £20–£100 new | £50–£300 (eBay ex-Soviet surplus) |
| Physical size | Large for high V/C ratings | Compact cylindrical form |
| Dielectric stability | Air: humidity affects tuning | Vacuum: perfectly stable |
Air variable capacitors fail at QRO for three compounding reasons:
The conclusion reached independently by every serious builder (including KI5AIF's detailed analysis): Start with the vacuum variable capacitor. Design everything else around it.
American-made with well-documented specifications. CVDD-1000 = 15–1000 pF at 10 kV peak — covers 40–10m for a 1.0m diameter loop. CVDC series reaches 25 kV for 1 kW+ operation. Available new from RF Parts; used units on eBay.
Abundant USSR military surplus on eBay, typically £40–£100. Common types: 10–1000 pF / 15 kV. Russian datasheets available online. Inspect for physical damage before purchase. Commission at low power first — if the vacuum has degraded you will hear arcing at a few watts.
Commercial broadcast and RF engineering grade. Very low ESR, precisely characterised capacitance vs shaft-turn curves, defined end-stops. 15–20 kV versions readily available. Premium cost justified for a permanent high-use installation. Voltronics available new from Digi-Key.
Three methods to transform the high loop impedance (thousands of ohms at resonance) to 50Ω for the feedline.
A small secondary loop placed inside the main loop acts as the primary winding of a loosely coupled transformer. Most common coupling method — forgiving, adjustable, and DC-isolated.
A tap point on the main loop conductor, connected via a series capacitor to the coax centre. Essentially a tapped transformer — very low loss but requires physical connection to the high-voltage main loop.
A short shorted stub (hairpin) connected across the capacitor forms an L-network with the feedline. Simple — just a bent piece of wire or copper tube — but narrow-band and requires reconfiguring when changing loop size.
For QRO fixed installations: Use the Faraday coupling loop. It maintains complete DC isolation between feedline and main loop, significantly reducing common-mode RF problems, and separates the coupling function from the high-voltage resonant circuit.
At 400W, unsuppressed common-mode RF on the coax outer and RFI from the stepper driver are significant interference sources that must be actively managed.
A magnetic loop with a Faraday coupling loop is inherently unbalanced when connected to coaxial cable. The intense near-field of the resonating loop induces significant common-mode current on the coax outer conductor. At 400W, circulating loop currents can reach 80–160A — the induced common-mode is proportionally high.
Symptoms of unsuppressed common-mode RF include: SWR meter errors; RF burns from metal chassis; interference on other station equipment; corrupted stepper motor position data; and potentially misleading VSWR readings that mask real antenna problems.
An RF choke installed at the shack end of the feedline is largely ineffective. The coax outer has already picked up common-mode RF over its entire run. The choke must be within 30cm of the coupling loop connection — this is where the common-mode current originates. A weatherproof toroid choke at the antenna is mandatory, not optional.
Selecting the wrong ferrite mix for your frequency range produces a choke that appears visually correct but provides 10–20 dB less impedance than expected. The key parameter is resistive (lossy) impedance — not inductive impedance, which is reversible and can store/re-radiate RF.
The preferred choice for general HF use. High resistive impedance across the entire HF spectrum. If you only buy one type, buy Mix 31. Fair-Rite material #31. Available as FT240-31 toroids and snap-on cores.
Peaks around 25–150 MHz. Excellent for 12m–10m. Significantly less effective than Mix 31 at 3.5–7 MHz. Good if you primarily operate the upper HF bands. Widely stocked at component suppliers.
High permeability — excellent on 160m and 80m. Peak impedance around 1–10 MHz. Less commonly stocked. Type 78 is a useful alternative for 80m/40m focus operations.
Low loss, high Q material — excellent for inductors and transformers but poor for RF suppression chokes. Low resistive component means RF can pass through. Often sold as a generic "ferrite" — verify mix number.
The stepper motor cable runs from the controller in the shack to the motor at the antenna — potentially 5–20m of wire acting as an antenna in the near-field of the loop. At 400W, this cable will pick up significant RF and conduct it back into the controller and transceiver unless specifically suppressed.
A ferrite choke at only one end of the motor cable is insufficient. The cable segment between a single choke and the unprotected end continues to act as an RF antenna. Chokes are mandatory at both the controller output connector AND immediately adjacent to the motor. Measured improvement: ~20 dB noise reduction with proper dual-end chokes vs none.
DRV8825 and A4988 stepper drivers generate fast current-switching transients on motor winding outputs. These radiate from the motor leads and conduct back into the power supply. At QRO power levels this causes detectable noise on the received signal:
Reference: K9YC's "A Ham's Guide to RFI, Ferrites, Baluns, and Audio Interfacing" is the definitive ham radio ferrite reference. Available free at k9yc.com/RFI-Ham.pdf
Motorised remote tuning is not optional at QRO — the 4–8 kHz VSWR bandwidth on 40m demands sub-kHz precision.
200 steps/revolution (1.8°/step). Holding torque 40–60 Ncm. Supported by all common drivers (DRV8825, A4988, TMC2209). With 1/16 microstepping and 10:1 gear reduction = ~32,000 steps per capacitor revolution. More than adequate resolution.
Adds rotary encoder on motor shaft for closed-loop position control. The controller can detect motor stall (ice, mechanical jam) and take corrective action or raise an alarm. More complex but essential for unattended outdoor operation.
Vacuum variable capacitors require precise, high-torque shaft rotation — often 10–20+ full turns to traverse the full capacitance range. A direct-drive stepper provides insufficient torque on some VVC types and insufficient resolution on all:
Typical configuration: NEMA 17 + 10:1 planetary or 3D-printed spur gear reduction + 1/16 microstepping = ~32,000 effective steps per VVC shaft revolution. At 10 turns total VVC travel = 320,000 steps total. At 7 MHz, frequency resolution per step ≈ 0.02 kHz — excellent.
20:1–80:1 reduction, zero backlash, self-locking (the VVC cannot back-drive the motor). Eliminates position drift from vibration during high-power transmission. Harder to source but worth the effort for a permanent QRO installation.
Print in PETG for best stability. Multiple gear stages can achieve 10:1–20:1 easily. Check for and minimise backlash — it corrupts stored frequency presets. Consider a spring-loaded anti-backlash gear for critical applications.
Helical or Oldham coupler absorbs motor-to-VVC shaft misalignment. Use as the final coupling stage after a gear reduction box. Avoid jaw couplers — they have significant backlash that compounds positioning errors.
Loftur Jónsson TF3LJ (also VE2AO) has released an open-source magnetic loop controller that has become the standard choice for serious homebrew automated loops. It implements predictive tuning — the controller knows the capacitor position required for any frequency based on its calibration table, and pre-tunes without any RF transmission required. Frequency presets allow instant band-changes.
Supported transceivers: Icom (CI-V: IC-7300, IC-7610, IC-706, etc.), Kenwood (CAT), Yaesu (CAT), Elecraft K3/KX3, and any rig with USB/serial frequency output.
Find the firmware and documentation by searching GitHub for TF3LJ magnetic loop controller or visit TF3LJ's QRZ page. The project has active community support.
3D printing has transformed homebrew MLA construction. Use PETG or ASA outdoors — PLA degrades in UV and heat.
Protective housing and mounting bracket for vacuum variable capacitors. Includes strain relief for capacitor terminals and wire management. Print in PETG or ASA for outdoor installations.
→ View on ThingiverseComplete printable magnetic loop assembly by K1FM. Structural parts, capacitor mount, coupling loop support, and cable management all included. Well-documented build notes with photos.
→ View on ThingiverseCompact magnetic loop with printable frame and capacitor mount. Covers 10–20m. Good starting point for a portable design — includes full assembly instructions and BOM.
→ View on ThingiverseModular Faraday coupling loop support — adjustable coupling loop holder allowing precise positioning inside the main loop. Essential for multi-band coupling optimisation without dismantling the antenna.
→ View on ThingiverseFaraday coupling with Vernier positioning — fine adjustment of coupling depth without dismantling. Very useful for achieving optimum match across multiple bands from one fixed installation.
→ View on ThingiverseLightweight portable design for field use — compact, quick breakdown and reassembly. QRP rated (not for 400W), but an excellent reference design for mechanical structure and coupling loop integration.
→ View on Thingiverse| Part | Recommended Material | Infill | Notes |
|---|---|---|---|
| Capacitor housing (outdoor) | ASA or PETG | 40–60% | ASA superior UV resistance. PETG is acceptable for sheltered outdoor use. |
| Stepper motor bracket | PETG | 40–50% | Must be non-conductive to mechanically isolate motor from loop conductor. |
| Gear reduction parts | PETG or ABS | 50–70% | Higher infill for mechanical load. Target tight tolerances to minimise backlash. |
| Coupling loop support | PETG | 25–35% | Lightweight. Must be non-conductive — verify filament has no carbon loading. |
| Controller enclosure | Not recommended | — | Use a metal enclosure (aluminium project box) to contain stepper driver RFI. |
| Indoor loop parts only | PLA acceptable | 20–40% | PLA degrades in UV and heat above 60°C — outdoor use not recommended. |
As of 2026, no commercial MLA is rated for 400W+ continuous operation. The QRO niche requires homebrew construction.
| Antenna | Max Power | Freq Range | Diameter | Approx Price | Notes |
|---|---|---|---|---|---|
| AlexLoop Walkham | 25W | 7–30 MHz | 0.90m | ~$220 | Portable, manual tune, Faraday coupling. No motor. Popular for SOTA. |
| Chameleon F-Loop 2.0 | 25W SSB / 5W CW | 7–30 MHz | 0.56m | ~$200 | Very compact portable. Bandwidth extremely narrow — requires constant retuning. |
| MFJ-1786 | 150W | 10–30 MHz | 0.91m | ~$450 | Motor-driven air variable. Upper HF only. Built-in controller. 150W limit enforced by air cap. |
| MFJ-1788 | 150W | 7.3–21.4 MHz | 1.07m | ~$550 | Larger loop covers 40m. Same air cap limitation — 150W is the hard ceiling. |
| Alpha Antenna MLA | 100W SSB | 3.5–54 MHz | 0.91m | ~$500 | Wide claimed coverage with manual tuning. 100W max. Build quality variable. |
| Ciro Mazzoni Baby Loop | 100W | 7–30 MHz | 0.70m | ~€600 | Italian manufacture. 100W limit. Better build quality than MFJ. Premium pricing. |
| Homebrew (vacuum VVC) | 400W–1 kW+ | 7–30 MHz | 1.0–1.5m | £200–£500 | Vacuum variable + TF3LJ controller + copper tube. Full QRO. This guide. |
The QRO gap: Every commercial MLA uses an air variable capacitor, limiting them to 150W maximum. Operating an MFJ-1786 at 400W would destroy its capacitor instantly — the rated voltage is approximately 3 kV; the actual voltage at 400W would be ~13 kV. For 400W+ operation there is no off-the-shelf option. Homebrew with a vacuum variable capacitor is the only route.
Three video presentations covering MLA theory, construction, and QRO operation by Rael Paster M0RTP.
Official RSGB publications, EMC guidance, and technical references.
Official RSGB EMC guidance document covering magnetic loop antenna installation, common-mode RF management, and interference avoidance. Essential reading for any indoor or restricted-space loop installation.
RadCom and RadCom Plus have published multiple technical articles on magnetic loop design and operation. Search the RSGB publications archive for "magnetic loop" or "small transmitting loop" to find a comprehensive reading list covering theory, construction, and operating practice.
Dedicated chapters on electrically small antennas including magnetic loops — theory, construction, and practical operating guidance. Available from RSGB publications. A useful supplement to online resources, particularly for the theoretical treatment.
Authoritative guidance on RFI management for all antenna types — particularly relevant for indoor MLA installations in flats or multi-occupancy buildings. Their ferrite choke and cable management recommendations are consistent with K9YC's guidance.