Characterising Different Metal Detector Types

Electromagnetic Induction Detectors are often characterised as either being 'Time Domain' or 'Frequency Domain'.

Time domain versus Frequency Domain Metal Detectors

Pulse Induction/Time Domain

Alternative names for this type of detector is Time Domain, Pulse Induction (PI), Pulse, Pulsed Mode, or Transient. 

These are characterised as detectors that transmit pulses on a Transmit coil and, at specific times following each pulse, they sample the amplitude of the resultant voltage or current on a Receive coil.  Analysis doesn't involve analysing any specific frequencies or characteristics of any specific frequencies in the received signal.  In fact, with almost all of today's current detectors, analysis doesn't involve any analysing of received signal frequencies whether discrete or not.  That is likely to change in the future as the likes of 100 million sample per second A/Ds with good resolution become more economical.

Strictly speaking the term 'time domain' will end up incorrect when new versions of Pulse Induction detectors incorporate techniques such as wavelet functions that allow the analysis of frequencies in the received signal of the detector.  They can still be differentiated from Frequency Domain detectors as even with these techniques, analysis is still not likely to be looking at 'discrete' or specific frequencies.

Due to the above problems with the term Time Domain, we will refer to them as Pulse Induction or PI detectors but we will continue to use the term Frequency Domain Detectors to cover all non Time Domain detectors and use either VLF or Induction Balance to cover the subset of Frequency Domain Detectors that use a Transmit and Receive coil with Induction Balance.

With Pulse Induction detectors, the Transmit and the Receive coils are typically the same coil.  One advantage of Pulse Induction detectors compared to Induction Balanced detectors is that the PI detectors have very simple requirements for the coil(s).

Pulse Induction detectors have problems detecting low conductivity objects.  This includes metal such as stainless steel.  For many purposes, such as gold prospecting, this is a good feature.

Low conductivity background and nuisance items, such as sea water and thin foils have a very short decay time. A pulse induction detector, which is tuned to sample only a specific portion of the received signal is easily made insensitive to them by an appropriate choice of the delay (some tens of µsec) between switch-off and sampling of the Receive coil. A similar argument applies to purely magnetic but non-conductive targets, which are magnetised by the transmit pulse but demagnetise just as promptly after switch-off.

Pulse systems are normally the detector of choice when it comes to working in salt water or strongly mineralised soils.

Most Pulse Induction detectors have a relatively slow current build up followed by an extremely quick cut off of current.  This quick cut off of current creates a quickly changing magnetic field and that creates strong eddy currents.

It is more difficult to obtain characterisation information from a Pulse Induction metal detector.  Until relatively recently, they couldn't even differentiate between ferrous and non ferrous metals.

Today's Pulse Induction metal detectors generally differentiate between different non-ferrous anomalies by measuring how well eddy currents flow. This is determined by an anomalies 'time constant'.  This is sometimes referred to as conductivity or resistance but that is not strictly correct as conductivity/resistance is only one of two properties of an anomaly that determines its time constant.  The other property is the anomalies inductance. The inductance may be thought of as the effective quantity of eddy currents, and is related to the size of the eddy current path.

Historical performance results of Time Domain based detectors has shown better Independence of result to the surrounding soil and rocks.  Generally, the small grains that make up the ground and many rocks have a very short time constant while the time constant of most anomalies of interest are very much longer.  This simple definition based in the time domain makes separation with Pulse Induction systems very easy.

Unfortunately, the definition in the time domain of other characteristics of anomalies is extremely difficult and the mathematics poorly developed.  The concepts being used or attempted by designers toward better characterisation of anomalies look more like art or folklore than rigorous mathematically developed science.

In mathematical theory, anything that can be done in the time domain can be done in the frequency domain and vice versa.  Unfortunately, this doesn't seem to translate to practicality.

Both proponents of Time Domain and proponents of Frequency Domain Detectors declare their detectors as showing less problems from interference from ambient electromagnetic noise.  The detectors handle interference in different ways.  Pulse Induction detectors perform an averaging of the values received from different pulses over time.  Frequency Domain Detectors such as VLF detectors filter the frequencies that they analyse to narrow band widths.  Noise sources generally produce noise over a wide bandwidth with the energy associated within each narrow bandwidth being very small.

With most Pulse Induction detectors, the percentage of the overall time that Pulses are being produced is very small.  Normally, the length of time between pulses is far greater than the time it takes to produce a pulse.  On the other hand, the actual current being produced at the peak of each pulse is generally far greater than that being used with Frequency Domain detectors.

The stronger current pulses produce stronger magnetic fields and the extremely fast turn off of the current produces quicker changing magnetic fields that produce stronger eddy currents.  This allows these Pulse Induction detectors to detect anomalies at greater depth than Frequency Domain based detectors.  It also allows Pulse Induction detectors to be able to detect smaller anomalies. 

Regardless of the fact that the pulse draw more current, the total power consumption can work out to be less due to the long length of time between pulses.

All of these advantages together has increasingly over recent years made Pulse Induction detectors the detector type of choice for gold prospecting.  They are also increasingly being used in humanitarian demining, particularly in difficult areas where ground effects cause problems with detection using the current Frequency Domain Detectors.

The design of Pulse Induction systems does not lend itself to the concept of a software oriented detector.  It is difficult to use an audio card for either D/A output to control the transmit signal or A/D input to read the received signal.  The creation of a Pulse Induction module for this project is likely to involve a lot more hardware design than that required for the VLF Induction Balanced detector.

Frequency Domain.

Sometimes referred to as continuous Wave (CW).  Most actual designs are also referred to as 'Very Low Frequency' (VLF) design as most of these operate in the Very Low Frequency (3kHz to 30kHz) range.  There have been Frequency Domain detectors operating in the Ultra Low Frequency (ULF) range (300Hz to 3kHz) as well as detectors operating in the Low Frequency (LF) range (30kHz to 300kHz).  For simplicity, we will incorrectly refer to the whole ULF, VLF, LF range as VLF.

Frequency Domain detectors can be further sub classified into Frequency Shift and Fixed Frequency design.

Frequency Shift Design.

These are also called Varying Frequency Design

Some older electronic based systems and some of todays very cheap detectors are designed so that the frequency of the received signal is able to vary slightly with the changing inductance of the anomaly.  The most common of these is the Beat Frequency Oscillators (BFO).  These circuits normally use the same coil for transmitter, receiver and as an element of an oscillator circuit.  As an anomaly comes near the coil, the effect of the induced eddy currents that the anomaly produces is to change the inductance of the coil.  This changes the frequency of the oscillator.

Varying frequency design has often been very cheap to make but has not otherwise provided as good a result as the other designs.    The principles of the design do not work in well with the use of A/D output and D/A input based computer controlled design.

Fixed Frequency Design

Most are also referred to as Induction Balance (IB) detectors or VLF detectors.

The Fixed Frequency detectors are characterised as detectors that involve transmitting on a Transmit coil one or several discrete frequencies and analysing amplitude and/or phase of discrete frequencies in the received signals input from a Receive coil.

Almost all modern frequency domain metal detectors regulate their Transmission to a specific frequency or frequencies.  This results in all available information from the input being contained in the phase and amplitude of those frequencies and/or their harmonics and intermodulation frequencies.

The received signals can include one or both of the following:

• waveforms obtained by continuously reading the voltage across the receive coils and/or
• waveforms obtained by continuously reading the current through the coils.

The received waveform and in some detectors also the transmitted waveform can vary their phase as they approach or depart from an anomaly.  The actual changing of the phase is the equivalent to a small change of frequency but this only occurs while the detector is in motion relative to the anomaly.

It is possible to have a metal detector that uses both Pulse Induction and Frequency Domain techniques at the same time.  One or several patents exist on this.

If a metal target is attracted to a magnet, it is called a ferrous target.  If not, it is called a non-ferrous target. In prospecting, ferrous targets are not normally sought.  Most of the time, the ferrous targets are considered junk or trash. Consequently most prospecting detectors have discrimination modes that the user can set so that the ferrous target is ignored when it is detected.

The design of VLF Induction Balanced detectors lends itself well to the use of audio cards for performing D/A and A/D conversion.  As a consequence, it also lends itself well to the concept of a software oriented metal detector design.  Due to the use of the audio card and the similarity to that of processing audio data, a range of software modules intended for audio are able to be used in the design.

With the exception of the coils, the general electronics required to interface the coils of a VLF Induction Balanced detector to sound card is very little.  Any good quality VLF detector requires two coils which have to be Induction Balanced to some degree.  This requirement is given in more detail in the following page:

http://humanise.org/demining/induction-balance

Most of the electromagnetic phenomena that can be used to characterise anomalies is easier described in the frequency domain than in the time domain.  This makes the possibility to obtain information about the characteristics of anomalies far easier with a frequency domain based detector than that realistically possible with Pulse induction.

The reasons that the initial focus of this Open Source project is to create an Open Source Induction Balanced detector design are the following:

• The better capability of Induction Balanced detectors to provide characterisation of anomalies (which is the most required feature wanted for advancing humanitarian demining)
• The ease that Induction Balanced detectors allow a software oriented design.