Metal Detector PCB: Coil Drive and Signal Processing
What the Board Does
A metal detector board drives a search coil with an alternating current, listens to the tiny change that a metallic object causes in the coil’s magnetic field, amplifies that change and decides whether it represents a target. The signal of interest is buried in noise that is many orders of magnitude larger, and the board is the instrument that recovers it.
That makes the design an analogue problem before it is a digital one. The microcontroller and the display are straightforward; the receive chain, the shielding and the phase stability are what separate a detector that finds a coin at 20 centimetres from one that finds it at 30.
The Detection Methods
Very low frequency, or induction balance. A transmit coil produces a continuous sinusoid and a receive coil is arranged so that its signal cancels, or nulls, in the absence of a target. A metal object disturbs the coupling and produces a small residual signal whose amplitude and phase carry information about the target. The phase difference between the transmit signal and the received signal is what allows the detector to distinguish a ferrous object from a non ferrous one.
Pulse induction. A short, high current pulse is sent through the coil, and the board measures the decay of the eddy currents in the target during the interval between pulses. Pulse induction is less sensitive to ground mineralisation and is the usual choice for a beach or a saltwater environment, at the cost of a more demanding power stage.
Beat frequency oscillator. Two oscillators, one of which includes the search coil, are mixed and the difference frequency is monitored. It is simple and inexpensive, and it is what most beginner designs use.
The method determines the board architecture, and in particular whether the design is dominated by a continuous wave transmit path with a phase sensitive receiver or by a high current pulse driver with a fast, low noise sampling receiver.
The Blocks on the Board
Transmit stage. In a continuous wave design it is an oscillator and a driver that deliver a stable sinusoid to the coil. In a pulse design it is a switch, usually a MOSFET, that dumps a defined current pulse into the coil, with the DC link capacitance and the damping network around it. In both cases the stability of the transmitted waveform sets the limit of the whole instrument.
Coil interface. The connector, the matching network and the damping components that shape the coil current. The coil itself is the sensor, and the interface has to be designed for the impedance the coil presents rather than for a convenient value.
Receive preamplifier. This is the most demanding block on the board. The received signal is very small, so the amplifier needs a low noise front end, high gain and a bandwidth that passes the signal while rejecting out of band interference. The input stage should be as close to the coil connector as the layout allows, and its feedback network should be chosen for stability rather than for bandwidth.
Demodulator and filter. A phase sensitive detector, either an analogue mixer or a sampled measurement in a pulse design, converts the received signal into the two quadrature components that carry amplitude and phase. The subsequent filtering is what averages the noise down and gives the detector its depth.
Digitisation and processing. An analogue to digital converter and a microcontroller or digital signal processor run the discrimination, the ground balance and the target identification. In a modern detector the ADC resolution and the stability of its reference have a direct effect on the achievable depth.
Audio and user interface. The tone generator or a small amplifier drives a speaker or headphones, and the display and the controls provide the settings. Audio is part of the measurement chain, because the operator is listening for a change in tone rather than reading a number.
Power. Battery management and low dropout regulation for the analogue rails, with attention to the noise that a switching supply would inject into the receive chain.
Design Rules That Matter
Treat the receive chain as an instrument. Every microvolt of noise added before the demodulator reduces the usable depth. A low noise amplifier, metal film resistors in the critical positions, a clean supply and a ground plane that is continuous under the signal path are the basics. The input trace should be short and guarded, and the amplifier should be placed before any digital routing crosses the area.
Separate the transmit and receive domains. The transmit stage is a large signal, whether it is a continuous sinusoid or a high current pulse, and the receive stage is looking for a signal a million times smaller. They need separate return paths, separate supply decoupling and, in a pulse design, a physical separation that keeps the pulse current out of the sampling window. A shared return path is the most common reason a prototype detector works on the bench and becomes unstable in the field.
Mind the phase. Discrimination depends on measuring the phase difference between the transmit and receive signals accurately, so any component in the receive path that has a temperature coefficient or a drift with time shows up as a false target. Stable components, thermal symmetry between the two channels and mechanical rigidity around the coil connector all matter.
Win the shielding battle. The search coil and its cable are excellent antennas, and the board is a noise source. A shielded coil cable, a proper shield connection to the board ground and filtering at the connector reduce the electromagnetic interference that otherwise appears as chatter. On the board side, a compact transmit loop, a ground plane and the avoidance of long unterminated traces reduce what the design radiates.
Design for the environment. A detector is used on wet grass, salt sand, or in a factory. Moisture on the coil or the connector changes the capacitance and the balance point, and a board without protection will drift. Conformal coating or potting, a sealed connector and a layout that avoids exposing high impedance nodes to the environment are the practical measures. Our notes on PCB design and layout cover the practices behind these rules.
Keep the power quiet. If a switching regulator is used for efficiency, its output has to be filtered and its switching node kept away from the analogue section, or the receive chain will hear it. For a small handheld, a linear regulator with a modest quiescent current is often the better trade.
Board Types and Materials
A beginner or do it yourself design can be built on a single or double layer FR-4 board, and the layout challenges are manageable because the depth expectation is modest. A serious instrument uses four layers, with a solid ground plane under the receive chain, a clean analogue supply plane and the digital section kept on its own layer and its own area.
Standard FR-4 is adequate for the frequency range these detectors use, and the losses are not the limiting factor. Where a design does operate at higher frequency, a low loss laminate reduces the loss and the drift, at a cost premium. Copper weight matters mainly in a pulse induction transmitter, where a thicker copper reduces the resistive loss and the heating in the coil drive path. ENIG is a common finish because it is flat and gives a reliable solder joint on the fine pitch analogue devices, and potted or coated assemblies are used for waterproof products. Our notes on PCB manufacturing describe how these options are produced.
Testing
A metal detector board cannot be validated with a continuity test. The bring up procedure starts with the transmit stage, checking the waveform and the current in the coil, then the receive chain with a signal generator or a calibrated target, then the null adjustment, which sets the balance point of an induction balance coil. The test then moves to the real target response: a set of test objects at a defined distance, measured through the audio output, because that is what the operator uses.
Two further tests catch the failures that a bench test misses. The first is a drift test over time and temperature, which exposes a false signal in a receiver that changes as it warms up. The second is a field test over mineralised ground, because ground mineralisation is a real signal to the receiver and the ground balance circuit exists to cancel it. Our PCB assembly group builds these boards, and our notes on PCBA testing describe how functional checks of this kind are structured.
What Drives the Cost
The circuit complexity and the analogue components dominate, not the laminate. A simple beat frequency design on a two layer board is an inexpensive item at volume. A four layer design with a low noise receive chain, a fast ADC and a processor capable of real discrimination costs several times more, and an industrial or security product with a ruggedised enclosure, a waterproof coil and a full functional test costs more again.
At volume the price falls mainly because the components and the assembly are amortised, and the panel layout can be optimised. A quick prototype, by contrast, carries the set up cost and the engineering time, which is why the first few boards of a custom design are always the most expensive ones.
Where These Boards Are Used
Security screening uses metal detection in walk through portals, hand held wands and baggage inspection, where the requirement is reliability and repeatability rather than depth. Mining and geological exploration use detectors to find mineralisation and to locate objects in the ground, and industrial production lines use metal detectors to check food and pharmaceutical products for contamination, which is a different discipline with its own standards.
The hobby and treasure hunting market is the largest consumer of the classic handheld detector, and it is also where most of the design innovation in coils and signal processing appears.
FAQ
Which detection method is best? Very low frequency induction balance gives the best discrimination and is the usual choice on land, while pulse induction is more tolerant of salt water and mineralised ground but discriminates less well.
Why is the receive amplifier so important? Because the target signal is extremely small, and any noise added in the first stage cannot be removed later. The noise floor of the front end sets the practical detection depth.
Does the board need special materials? Usually not. Standard FR-4 works well at these frequencies. A low loss laminate only helps when the design operates at a high frequency or requires exceptionally low drift.
Why does a detector drift after it warms up? Because the phase and gain of the receive chain change with temperature. Stable components, thermal symmetry and a mechanical layout that keeps the heat sources away from the analogue section reduce the drift.
How is the coil matched to the board? The coil has a defined inductance and resistance, and the board carries the matching and damping network designed for that specific coil. Changing the coil without changing the network changes the performance.
Conclusion
A metal detector PCB is an instrumentation board built around a very small signal. The transmit stage has to be stable, the receive chain has to be quiet, the two have to be separated, and the phase has to hold as the board warms and the environment changes. The processor and the display are the easy part. Design the analogue front end and the shielding with the care the signal deserves, and the detector will find what the coil can reach.
