UNCLASSIFIED
TM-TOOL-001
ADF SYSTEMS — ANTENNA DIRECTION FINDING
Watson-Watt, Doppler, Phased Array, and Rotary Loop Methods
Prepared by: Mervyn Martin, KO6NNH  •  Merced, California  •  26 May 2026
Amateur Radio / Electronics — Not for commercial use

Chapter 1 — Introduction and Scope

This manual covers four antenna direction finding (ADF) system architectures used in amateur radio and field operations: Watson-Watt two-channel amplitude comparison, Doppler frequency-shift bearing, switched phased-array (Butler Matrix), and rotary loop servo. All four systems produce a bearing to an RF source; choice depends on frequency range, required accuracy, and available hardware.

Covered frequency range: 1.8–450 MHz depending on antenna array geometry. The Butler Matrix design covers HF through VHF with a single mechanical layout scaled to frequency.

Bearing accuracy targets:

  • Watson-Watt: ±5° RMS in clear-field conditions
  • Doppler: ±2° RMS, dependent on rotation rate and SNR
  • 4-element Butler Matrix: ±3° RMS interpolated
  • Rotary servo loop: ±1° mechanical, ±2° bearing

Chapter 2 — Theory of Operation

2-1 Watson-Watt Method

Two orthogonal loop antennas (North-South and East-West) produce sine and cosine voltage components of received signal amplitude as a function of bearing. A third omnidirectional element (sense antenna) resolves the 180° ambiguity.

Bearing calculation:

bearing = atan2(V_NS, V_EW)   (degrees, corrected for magnetic declination)

Each channel must be gain-matched to within ±0.5 dB; phase-matched to within ±2° across the operating band. Gain/phase imbalance directly adds to bearing error.

2-2 Doppler Method

An element sequentially commutated around a circle of diameter d causes a frequency modulation on the received carrier equal to:

f_d = (v / λ) × cos(θ − φ)

where v = element tangential velocity = πdf_rot, λ = wavelength, θ = bearing to source, φ = current element angle. A PLL or IQ discriminator extracts θ from the FM sidebands. Rotation rate f_rot is typically 100–600 Hz; higher rates extend the capture range but increase bandwidth requirements.

2-3 Butler Matrix (4-Element Switched Array)

Four elements in a square array with λ/4 spacing feed a passive Butler Matrix beamforming network. The matrix produces four orthogonal beams at ±45° and ±135° simultaneously. RSSI comparison across the four beam ports yields a bearing estimate; interpolation between adjacent beam peaks resolves bearing to approximately 3° RMS.

The Butler Matrix is constructed from four 3 dB 90° hybrid couplers and two fixed 45° phase shifters interconnected as a 4×4 passive network. No active components in the RF path; loss is approximately 0.5–1.5 dB depending on hybrid quality.

2-4 GPS Compass Integration

Magnetic bearing from an ADF system must be converted to true bearing for navigation use. GPS compass integration provides:

  • Vehicle/platform heading reference (eliminates mount alignment error)
  • Magnetic declination correction (from GPS position + WMM model)
  • True bearing output = ADF magnetic bearing − declination + heading offset

Chapter 3 — Equipment and Materials

ItemWatson-WattDopplerButler Matrix
Antenna elements2 loops + sense whip4–8 vertical elements4 vertical elements
Element spacingLoop aperture sets sensitivityλ/4 radius circleλ/4 square
RF channels2 (plus sense)1 commutated4 simultaneous
SDR / receiverDual-channel coherentSingle channel4-channel or switched
ControllerESP32 or PCESP32 (PWM commutation)ESP32 (RSSI ADC)
CoaxMatched-length pairsSwitched relay treeFixed equal-length runs

Chapter 4 — Construction and Assembly

4-1 Watson-Watt Loop Construction

  1. Wind two identical shielded loops: 30 cm diameter, 3 turns #18 AWG, Faraday shield (gap at top, not a closed loop).
  2. Mount loops orthogonally on a common center mast; orient one loop N-S, other E-W.
  3. Run equal-length coax (within 5 mm) from each loop to the receiver switching point.
  4. Connect sense antenna (vertical whip, λ/4 at center frequency) at same switching point.

4-2 Butler Matrix PCB

  1. Fabricate 4-port 3 dB 90° hybrid couplers on FR4 (Z0=35.4Ω microstrip, λ/4 length at design frequency).
  2. Connect hybrids per the Butler Matrix topology: E1,E2 → Hybrid #1,#2; outputs of Hybrid #1,#2 → Hybrid #3,#4 with a 45° fixed phase shifter between Hybrid #2 output and Hybrid #4 input.
  3. Verify phase relationships with NanoVNA S21 phase measurements before installing array elements.
  4. Mount array elements at corners of a square, λ/4 side length, connected to matrix element ports E1–E4.

4-3 Doppler Commutator

  1. Mount 4 or 8 vertical elements equally spaced on a circle of radius λ/4 at operating frequency.
  2. Wire each element through an RF relay (PIN diode or mechanical) to a common output coax.
  3. Drive relay sequence from ESP32 GPIO at rotation rate f_rot (start at 200 Hz for HF).
  4. Connect common output to receiver; extract FM component in software (GNU Radio or custom DSP).

Chapter 5 — Operating Procedures

5-1 Watson-Watt Bearing Measurement

  1. Tune receiver to target signal. Confirm adequate SNR (>20 dB for ±5° accuracy).
  2. Enable 2-channel sampling. Record V_NS and V_EW amplitude values.
  3. Compute bearing: brg = atan2(V_NS, V_EW). Apply declination correction.
  4. Confirm sense: enable sense antenna and verify the 180° correct quadrant.
  5. Average 10 readings; discard outliers more than 15° from median.

5-2 Butler Matrix Bearing Estimate

  1. Read RSSI on all four beam ports simultaneously (or in rapid sequence <10 ms total).
  2. Identify the two highest-RSSI ports (adjacent beams straddle the signal).
  3. Interpolate: bearing = beam_angle_1 + 45° × (RSSI_1 / (RSSI_1 + RSSI_2)).
  4. Apply platform heading offset if mounted on a moving vehicle.

Chapter 6 — Calibration

6-1 Watson-Watt Channel Balance

  1. Inject equal-amplitude, in-phase signal into both channels simultaneously from a common splitter. Verify RSSI within 0.3 dB.
  2. If unbalanced: add fixed attenuator pad (1–3 dB) to the stronger channel at the receiver input.
  3. Inject signal into NS channel only. Verify 90° + known bearing reads correctly. Repeat for EW channel.

6-2 Butler Matrix Beam Verification

  1. Place a known CW signal at each of the four expected beam-peak azimuths in turn (0°, 90°, 180°, 270°).
  2. Verify that the corresponding beam port shows maximum RSSI and adjacent ports show ≥3 dB lower level.
  3. Record beam-center azimuths. Apply offset table in software if beam centers deviate more than 5° from design.

Chapter 7 — Verification and Acceptance

After calibration, conduct a bearing verification test using a known-location transmitter:

  1. Position a low-power test transmitter at a measured azimuth from the array (use GPS or compass, to within ±1°).
  2. Take 20 bearing readings; compute mean and standard deviation.
  3. Acceptance criterion: mean error <5°, standard deviation <3°.
  4. If failed: re-check element spacing, coax phase lengths, and channel balance.
  5. Record results in calibration log: date, frequency, test azimuth, mean bearing, standard deviation, operator.

Appendix A — Design Parameters Quick Reference

ParameterWatson-WattDopplerButler 4-el
Accuracy (typical)±5°±2°±3°
Min SNR required20 dB15 dB10 dB
Multipath sensitivityHighMediumLow
Moving platformPoorGoodGood
Hardware complexityLowMediumHigh

Appendix B — Worked Bearing Example

Watson-Watt on 14.225 MHz. Measured: V_NS = 0.82 V, V_EW = 0.57 V.

bearing = atan2(0.82, 0.57) = atan2(0.82, 0.57) = 55.2 deg
Declination (Merced, CA) = +12.3 deg East
True bearing = 55.2 + 12.3 = 67.5 deg True

Sense antenna confirms NE quadrant (not SW ambiguity). Final reported bearing: 068° True.