Complete Guide to Building a Pulse Induction Metal Detector Circuit with Diagram

Start with a single-transistor driver stage paired with a power MOSFET (IRF840 or IRFP460) to generate clean 12–15 μs pulses at 100–300 Hz. A 47 μH air-core coil (0.5–0.8 mm wire, 100–150 turns) ensures minimal ring-down, critical for target separation in conductive soils. Use a fast-recovery diode (UF4007) directly across the coil to clamp voltage spikes under 40V, preventing MOSFET damage.

The receiver requires a two-stage amplification chain: an initial low-noise op-amp (TL072) with 40 dB gain, followed by a second stage (LM358) tuned to 10–20 dB. Install a 10 kΩ trimpot on the first stage to adjust baseline sensitivity without overloading the ADC. A 4-pole Butterworth active filter (Sallen-Key topology) centered at 1.2–3.5 kHz cuts noise while preserving decay signals from 10–30 g coins at 12–18 cm depths.

For timing control, a 555 timer in monostable mode triggers the pulse and arms the sample-and-hold circuit (LF398) 30–50 μs after pulse cutoff. A second 555 in astable configuration generates a 5 kHz reference clock for the PIC16F628A microcontroller, which digitizes the decay envelope with 8-bit resolution at 20 μs intervals. Storage capacitance (100 nF) on the ADC input pins must match the microcontroller’s settling time to avoid baseline drift.

Power distribution demands separate rails: 12V for the pulse driver, regulated 5V (LM7805) for logic, and 9V for analog stages, each decoupled with 10 μF tantalum capacitors. Ground loops are eliminated by star-point grounding at the battery negative terminal. For battery life, implement a low-power mode that reduces pulse frequency to 50 Hz when no targets are detected for >30 seconds, cutting current draw from 120 mA to 25 mA.

Building a Pulse Induction Locator: Schematic and Key Components

Select a dual-transistor arrangement for the transmitter stage–2N3055 and MJE13007 offer the required current handling. Pair them with a 1000μF electrolytic capacitor to store energy during pulse generation; this ensures consistent signal strength down to 30cm in mineralised soil.

Use a 555 timer IC in monostable mode to control pulse width. Set the timing resistors to produce 150μs bursts–longer pulses improve depth but reduce discrimination. A 10kΩ potentiometer adjusts sensitivity; calibrate it to avoid false triggers from wet sand or iron-rich clay.

Wind the search coil on a 25cm diameter former using 0.5mm enamelled copper wire–25 turns yield optimal inductance. Shield the coil with aluminium foil, grounding one edge to prevent electrostatic interference. A ferrous core increases sensitivity but narrows detection field; air-cored designs perform better in trashy areas.

The receiver stage demands a low-noise op-amp–TL072 or NE5532 work well. Configure it as a bandpass filter centred at 3kHz to reject powerline hum and high-frequency noise. Add a 1MΩ feedback resistor to set gain; excessive amplification will saturate the signal in highly conductive soils.

Integrate a sample-and-hold circuit with a 1μF capacitor and CD4016 analog switch. Capture the induced voltage precisely 50μs after the transmit pulse ends; this timing avoids transient spikes while retaining target response. Use a quad comparator LM339 to process the signal–one channel triggers on ferrous targets, another on non-ferrous.

A 12V sealed lead-acid battery powers the unit efficiently. Regulate voltage with a 7812 IC to maintain stable operation; a 1N4007 diode prevents reverse polarity damage. For portability, add a 3.3V Zener diode to power an LCD; it draws minimal current while displaying target depth estimates.

Test the assembly with a buried 25mm iron washer–it should register at 20cm depth. Fine-tune the threshold potentiometer to eliminate background noise. Replace the search coil with a 15cm elliptical version for precision work; it resolves small artefacts at 10cm but sacrifices maximum depth.

Key Components Required for a Pi-Based Search Coil Assembly

Start with the Raspberry Pi 4 Model B or later–its quad-core processor handles signal processing at 1.5GHz, essential for real-time pulse discrimination. Earlier models like the Pi 3B+ lack sufficient GPIO speed, introducing latency of up to 30ms during waveform analysis. Use a 32GB SanDisk Ultra microSD card; cheaper alternatives corrode under frequent write cycles, throttling performance.

Select a coil former machined from 10mm-thick HDPE. Polyethylene resists moisture absorption (0.01% @ 24h), unlike acrylic which fractures under thermal cycling. Wind the transmitter using 0.5mm enameled copper wire–derate current by 20% to prevent overheating. The receiver loop requires 0.2mm Litz wire, bundled in 60 strands to suppress skin-effect losses above 15kHz.

The pulse generator demands a IXYS IXTA100N120P IGBT. Its 1200V breakdown voltage exceeds MOSFET alternatives, mitigating transient spikes during coil discharge. Pair it with a 10Ω 25W ceramic resistor; film resistors degrade at 1W/cm², causing drift. Capacitors should be polypropylene: a 1µF 630V unit for pulse shaping and a 100nF 1kV for transient suppression.

• Current sensor: ACS712 (30A variant), calibrated to ±1.5% accuracy. Hall-effect alternatives introduce 5µs sampling delays, corrupting small-target signatures.
• ADC: ADS1115 (16-bit resolution) via I²C. Avoid cheaper 12-bit converters–target thresholds below 5mV become unreadable.
• DAC: MCP4725 (12-bit) for reference voltage. Its internal EEPROM retains settings after power cycles.

Shield all signal traces with 2oz copper ground planes. Routing conventional 1oz traces introduces 23dB crosstalk at 25kHz. Use through-hole vias every 15mm; stitch capacitance drops stray inductance to 18nH, critical for rejecting 50Hz mains interference.

Power regulation hinges on three components:

1. A Mean Well LRS-100-12 12V 8A supply–switching topology prevents ripple exceeding 50mVpp.
2. TLE5205-2 linear regulators for sensitive analog sections; dropout ≤0.5V conserves battery life.
3. A 10,000µF 25V Panasonic FC capacitor on the main rail–low ESR (

Transmitter timing precision requires a DS3231 real-time clock module. Its ±2ppm drift equates to Citizen CFV206 (load capacitance 12.5pF) for ±10ppm stability.

User feedback mandates:

• A 2.42″ OLED display (SSD1306 controller), wired via SPI at 1MHz. Slower interfaces ghost images during dynamic sweeps.
• Blue LED indicators (470nm wavelength), pulsed at 2Hz to conserve power; avoid red LEDs–phototransistors exhibit peak sensitivity at 565nm, causing false triggers.
• A 5V piezo buzzer driven by a BC547 transistor, toggled via GPIO PWM to generate 2kHz alert tones.

Step-by-Step Winding Instructions for the Pulse Induction Probe

Select a 0.5mm enameled copper wire for the search loop–this gauge balances sensitivity and durability. Secure a circular form 20–25cm in diameter (a PVC pipe segment or sturdy cardboard disc works). Drill two small holes 180° apart near the edge to anchor the wire before winding. Apply masking tape to the form’s outer rim to prevent wire slippage during the process. Begin winding tightly, maintaining consistent tension; aim for 25–30 turns total, ensuring no overlaps or gaps between loops.

Termination and Shielding

Leave 30cm tail lengths at both ends for connection. Slide the loop off the form gently, preserving its shape. Wrap the assembly in two layers of electrical tape, covering the entire surface except for the last 2cm near the terminals. This prevents stray capacitance and mechanical wear. For added interference rejection, twist the tails together tightly for 5cm before separating them for soldering. Use heat-shrink tubing over the joint to insulate the connection point.

Attach the finished probe to an impedance-matching network: pair it with a 10nF ceramic capacitor and a 1kΩ resistor in series. This configuration optimizes pulse decay detection when driven by a 5V square wave at 5–15kHz. Test continuity with a multimeter; resistance should read 3–5Ω. If values deviate, recheck turns count–excess resistance indicates broken wires, while low readings suggest shorted loops.

Mount the coil perpendicular to the handle, leaving 2–3mm clearance between the winding and any metal parts to avoid eddy current interference. Seal the assembly in a polyurethane coating for weatherproofing if deploying outdoors. Calibrate sensitivity by sweeping the probe over a known target (e.g., a coin at 10cm depth) and adjust the pulse width or frequency until the signal peaks sharply without false triggers.

Adjusting Pulse Width and Frequency for Optimal Probe Performance

Begin with a pulse duration of 100–150 microseconds for general-purpose searches in loose soil. Shorter bursts–below 80µs–excel in highly mineralized terrains, where eddy current decay occurs rapidly, while longer pulses up to 250µs enhance penetration in conductive clay or wet sand. Ensure your timing capacitor values align: 1nF yields ~120µs, while 2.2nF stretches to ~240µs at a fixed charging voltage.

Set initial repetition rates between 100–200Hz for balanced discrimination and depth. Frequencies below 60Hz deepen signal reach but reduce target separation; values above 300Hz sharpen resolution yet sacrifice sensitivity to smaller remnants. A sweep test reveals optimal tuning: adjust the rate in 50Hz increments while monitoring signal-to-noise ratio–peaks typically cluster around 120Hz and 180Hz for ferrous and non-ferrous artifacts, respectively.

Combine pulse length and interval adjustments via a dual-potentiometer setup. Pair a 50kΩ linear taper for width control with a 100kΩ logarithmic taper for rate modulation. The logarithmic response compensates for diminishing returns at higher frequencies, preventing abrupt sensitivity drops above 250Hz. Test with reference samples–1cm copper disk, 3cm iron nail–at varying depths to calibrate both controls.

Material	Pulse Width (µs)	Frequency (Hz)	Detection Depth (cm)
Gold	110–140	150–220	10–15
Silver	90–120	180–250	8–12
Iron	180–220	80–130	18–25
Aluminum	130–160	200–300	6–10

Limit pulse width to 70% of the repetition period to prevent overlap. At 200Hz, the cycle duration is 5ms–max pulse length should not exceed 3.5ms. Violating this ratio causes residual electromagnetic interference, masking shallow targets. Use an oscilloscope to verify separation; clean gaps between pulses indicate proper timing.

Fine-Tuning for Environmental Factors

Compensate for soil conductivity by reducing pulse width by 20–30µs per 10% increase in moisture content. Sandy substrates tolerate wider pulses, while saline environments necessitate narrower bursts–down to 50µs–to avoid false positives from dissolved minerals. Pre-sweep the area with a fixed frequency (150Hz) and adjust width dynamically based on soil resistance readings.

Stabilize sensitivity under temperature drift by pairing timing resistors with negative temperature coefficient (NTC) thermistors. A 10kΩ NTC compensates for ~3% pulse drift per 10°C change, maintaining consistent eddy current excitation. Test across expected ambient ranges–-5°C to 40°C–verifying response stability with a buried 5cm steel washer at 15cm depth.

Advanced Target-Specific Configuration

Narrow pulses (60–90µs) at 280–350Hz isolate thin foils or fragmented coinage, while extended bursts (200–250µs) at 80–120Hz reveal dense, deeply buried relics. Configure a mode switch to alter both parameters simultaneously: “Coin” mode (150µs/200Hz) versus “Relic” mode (220µs/100Hz). Ground balance between modes by recalibrating the threshold voltage–typically 1.2V for mineralized soils, 0.8V for neutral substrates.