AU632320B2 - Improvements relating to metal detectors - Google Patents

Description

COMMONWEALTH OF AUSTRALIA Patents Act 1952-1969 Form COMPLETE SPECIFICATION

(ORIGINAL)

pa 63z3zu FOR OFFICE USE: Class Int. Class Application No Lodged Complete Application No Specification Lodged Published Priority: Related art: 0s S 0 0*99 *r *4 c *i: TO BE COMPLETED BY APPLICANT hi 4 r tact t

I.

St Name of Applicant: Address of Applicant: Actual Inventor: Address for Service: Main Lobethal Road, Basket Range, State of South Australia, Commonwealth of Australia BRUCE HALCRO CANDY Care of COLLISON CO., 117 King William Street, Adelaide, South Australia, 5000 BRUCE HALCRO CANDY Complete Specification for the invention entitled: IMPROVEMENTS RELATING TO METAL DETECTORS The following statements is a full description of this invention, including the best method of performing it known to me: 9- /9 PATENT, TRADE MARKS 29 S /9 DESIGNS SUB-OFFICE 29 MAR 1990 SOUTH AUSTRALIA L ,400275 -I _II J~_ SQUARE PULSE METAL DETECTOR.

This invention relates to conducting metal discriminating detectors.

The problem to which the invention is directed relates to difficulties associated with discriminatory detection of target object;3 when within an environment that provides either substantive magnetic roactive or resistive components of re-transmitted magnetic signals where it has been hitherto difficult to distinguish a target signal from a background signal.

Such an environment can be typically ironstone magnetic soils or salt water or indeed both.

The object of this invention is to achieve a method and apparatus by which 15 greater sensitivity can be achieved in such difficult environments with *ea .equipment that can be economically manufactured.

Concept of the Invention: According to one form of this invention there is provided a conducting metal e gdiscriminating detection apparatus for detecting a metal target in a target region comprising: .4 transmission means for transmitting a continuous pulsed voltage waveform to provide an alternating magnetic flux in the target region, the 25 continuous pulsed voltage waveform comprising abrupt transitions from one substantially steady state voltage to at least one other substantially steady state voltage; a detector coil for detecting a rate of change of magnetic flux in the target region, the rate of change of magnetic flux being a function of the alternating magnetic flux in the target region; measurement means for measuring at least two time averaged measurements of the detected rate of change of magnetic flux at spaced apart time intervals, the measurement means being adapted to measure the time averaged measurements after a duration of an initial time interval such that the initial time interval starts directly after one of the abrupt transitions and finishes when the rate of change of eddy currents, due to the effects of the alternating magnetic flux upon conductive soils within the target region, is substantially zero; p 1 7 2 processing means for comparing the time averaged measurements against two pre-determined responses, one resulting from an instantaneous component due to the effects of the alternating magnetic flux upon magnetic soils within the target region and the other from a historical component due to the effects of the alternating magnetic flux upon magnetic soils within the target region; and assessment means for providing an assessment signal from the comparison of the time averaged measurements, the assessment signal being indicative of the metal target within the target region.

In preference, the measurement means comprises: synchronous demodulator means for synchronous demodulating the rate of change of the magnetic flux at spaced apart time intervals; and 15 low pass filter means for low pass fiitering the synchronous S demodulated rate of change of the magnetic flux at spaced apart time o: intervals.

In preference, the transmission means providing the continuous pulsed voltage waveform such that the assessment signal is substantially S independent of the reactive to resistive response ratio component of the Sa magnetic soils within the target region at frequencies below 100kHz.

S In preference, the continuous pulsed voltage waveform comprises two 25 different pulse periods, these being a first pulse period and a second pulse a. period, the first pulse period being substantially of a shorter duration than the second pulse period, and the measurement means is adapted to synchronously demodulate the rate of change of the magnetic flux during a first measurement period occurring in the last quarter of the second pulse period to form a first signal, and the measurement means is further adapted to synchronously demodulate the rate of change of the magnetic flux during a second measurement period occurring in the first half of the first or second pulse period to form a second signal, wherein the second measurement period occurs after the rate of change of eddy currents in the conductive soils is substantially zero, and the processing means is adapted to form a ground balanced channel signal by scaling and subtracting the first signal from the second signal, the processing means being further adapted ii h 3 to compare the sign of the ground balanced channel signal with the sign of the first signal for the assessment means to provide a signal indicative of a ferrous target within the target region.

Alternatively, according to another form of this invention there is provided a conducting metal discriminating detection apparatus for detecting a metal target in a target region comprising: transmission means for transmitting a continuous pulsed voltage waveform to provide an alternating magnetic flux in the target region, the continuous pulsed voltage waveform comprising abrupt transitions from one F substantially steady state voltage to at least one other substantially steady state voltage; a detector coil for detecing a rate of change of magnetic flux in the 4 ,CI target region, the rate of change of magnetic flux being a function of the alternating magnetic flux in the target region; measurement means including synchronous demodulation means for 4 synchronous demodulating at least three measurements of the rate of L change of magnetic flux during at least three spaced apart time intervals, the measurement means being adapted to measure the rate of change of S 20 magnetic flux after a duration of an initial time interval such that the initial time interval starts directly after one of the abrupt transitions and finishes when the rate of change of eddy currents, due to the effects of the alternating magnetic flux upon conductive soils within the target region, is substantially zero; S 25 a filter means adapted to low pass filter the synchronous demodulated measurements of the magnitude of the rate of change of magnetic flux at spaced apart time intervals to provide at least three low pass filtered synchronous demodulated measurements; a processing means adapted to process the low pass filtered synchronous demodulated measurements to provide at least three time averaged values these being a first time averaged value, a second time averaged value and a third time averaged value; a calculating means adapted to: subtract the third time averaged value from the first time averaged value to provide a first resultant value; subtract the third time averaged value from the second time averaged value to provide a second resultant value; 16 CASE FEATURE *o i D a O 1 m O D o p e Q, Q s o* 4 multiply the second resultant value by a pre-determined value to provide a third resultant value; and subtract the third resultant value from the first resultant value to provide a value which is substantially zero for all iron oxides; and an assessment means for providing an assessment signal from the subtracting of the third resultant value from the first resultant value, wherein the assessment signal is indicative of a metal object within the target region.

In preference, the continuous pulsed voltage waveform comprises two different pulse periods, these being a first pulse period and a second pulse period, the first pulse period being substantially of a shorter duration than the second pulse period, wherein at least one of the spaced apart time intervals occurs in the first pulse period and at least one other time interval occurs in the second pulse period.

In preference, the synchronous demodulation means comprises n synchronous demodulators and is operably configurable to satisfy the equation: n 20 _SiPi= 0, i=1 where Si is the relative effective gain of the ith synchronous demodulator the output of which is to be added and low-pass filtered, where the sign of Si is consistent with the polarity of the pulse being demodulated, and Pi is the relative period for which the said synchronous demodulator is "on".

In preference, the calculating means is adapted such that the following equation is satisfied: a. for a purely magnetic component of any detected signal, Gk Mnak3k 0 a=1 and b. for an historical component of any detected signal i:I G; CnakPk 0 P3=1 where Gk is the gain of the demodulator passing the Kth contribution, Mn(xp is the demodulated and averaged magnetic component, due to magnetic soils, between the times a and p following the n th transition from 1 0 one substantially steady state voltage to at least one other substantially steady state voltage, and Cnap is the demodulated and averaged component resulting from the component of magnetic soils dependent on the historical component, due to magnetic soils, of the applied field between times a and 3 following the nth 15 transition from one substantially steady state voltage to at least one other substantially steady state voltage, whereby the output signal is substantially independont of the effect of any non-electrically conducting ferrite constituents existing in the target region.

In preference, the transmission means providing the continuous pulsed voltage waveform at a frequency such that the assessment signal is substantially independent of the reactive to resistive response ratio component of the magnetic soils within the target region when the frequency 25 is below 100kHz.

Alternatively, according to another form of this invention there is provided a method of conducting metal discriminating detection of metal target in a target region comprising the steps of: transmitting a continuous pulsed voltage waveform to provide an alternating magnetic flux in the target region, the continuous pulsed voltage waveform comprising abrupt transitions from one substantially steady state voltage to at least one other substantially steady state voltage, and the target region contains magnetic soils and electrically conductive soils; 6 detecting a rate of change of magnetic flux in the target region, wherein the rate of change of magnetic flux results from the rate of change of the alternating magnetic flux, rate of change of eddy currents flowing in the metal target, rate of change of eddy currents flowing in the electrically conductive soils and a rate of change of alternating magnetic flux modified by two components within the magnetic soils these components being an instantaneous component dependent upon the instantaneous values of the rate of change of the alternating magnetic flux and a historical component 1 0 dependent upon previous values of the alternating magnetic flux; measuring at least two time averaged measurements of the detected rate of change of magnetic flux at spaced apart time intervals, the measuring occurring after a duration of an initial time interval such that the initial time interval starts directly after one of the abrupt transitions and finishes when S1 5 the rate of change of eddy currents in the conductive soils is substantially zero; comparing the time averaged measurements against two predetermined responses, one resulting from the instantaneous component and the other from the historical component, wherein the pre-determined response is the expected response from the target region due to the effects of magnetic soils; and effecting an assessment signal from the comparing of the time averaged measurements, the assessment signal being indicative of the metal target within the target region.

t. S, In preference, the detecting further comprises the steps of: synchronous demodulating the rate of change of the magnetic flux at spaced apart time intervals; and low pass filtering the synchronous demodulated rate of change of the magnetic flux at spaced apart time intervals.

In preference, the method is further characterised by the target region having a reactive to resistive response ratio component of the magnetic soils within the target region which is substantially independent of an interrogating frequency below 100kHz.

In preference, the continuous pulsed voltage waveform comprises two different pulse periods, these being a first pulse period and a second pulse period, and the first pulse period is substantially of a shorter duration than the second pulse period, the method further including the steps of: detecting and synchronously demodulating the rate of change of the magnetic flux during a first measurement period occurring in the last quarter of the second pulse period to form a first signal, detecting and synchronously demodulating the rate of change of the magnetic flux during a second measurement period occurring in the first half of the first or second pulse period to form a second signal, wherein the second measurement period occurs after the rate of change of eddy currents in the conductive soils is substantially zero, and processing to form a ground balanced channel signal by scaling and 15 subtracting the first signal from the second signal, S* comparing the sign of the ground balanced channel signal with the sign of the first signal to provide a signal indicative of a ferrous target within the target region.

S 20 Alternatively, according to another form of this invention there is provided a method of conducting metal discriminating detection comprising the steps of: transmitting a continuous pulsed voltage waveform to provide an alternating magnetic flux in the target region, the continuous pulsed voltage waveform comprising abrupt transitions from one substantially steady state .i 25 voltage to at least one other substantially steady state voltage, and the target region contains magnetic soils and electrically conductive soils, detecting a rate of change of magnetic flux in the target region, wherein the rate of change of magnetic flux results from the rate of change of the alternating magnetic flux, rate of change of eddy currents flowing in the metal target, rate of change of eddy currents flowing in the electrically conductive soils and a rate of change of alternating magnetic flux modified by two components within the magnetic soils these components being an instantaneous component dependent upon the instantaneous values of the rate of change of the alternating magnetic flux and a historical component dependent upon previous values of the alternating magnetic flux; 'jfg I' ti V: 4e 1 j z synchronous demodulating at least three measurements of the rate of change of magnetic flux at spaced apart time intervals, the synchronous demodulating occurring a duration of an init*il time interval such that the initial time interval starts directly after one of the abrupt transitions and finishes when the rate of change of eddy currents in the conductive soils is substantially zero; low pass filtering the synchronous demodulated measurements of the magnitude of the rate of change of magnetic flux at spaced apart time intervals to provide at least three low pass filtered synchronous demodulated measurements; processing the !ow pass filtered synchronous demodulated measurements to provide at least three time averaged values these being a *1~ first time averaged value, a second time averaged value and a third time 04 15 averaged value; 9* *4 subtracting the third time averaged value from the first time averaged r value to provide a first resultant value; subtracting the third time averaged value from the second time averaged value to provide a second resultant value; 20 multiplying the second resultant value by a pre-determined value to provide a third resultant value; :subtracting the third resultant value from the first resultant value to provide a value which is substantially zero for all iron oxides; and effecting an assessment signal from the subtracting of the third resultant value from the first resultant value, the assessment signal being indicative of a metal object within the target region.

In preference, the continuous pulsed voltage waveform comprises two different pulse periods, these being a first pulse period and a second pulse period, the first pulse period being substantially of a shorter duration than the second pulse period, wherein at least one of the spaced apart time intervals occurs in the first pulse period anrid at least one other time interval occurs in the second pulse period.

S1;

V

6 f 9 In preference, the step of synchronously demodulating is accomplished by n coupled synchronous demodulators which satisfy the following equation: n |SiPi 0, i=1 where Si is the relative effective gain of the ith synchronous demodulator the output of which is to be added and low-pass filtered, where the sign of Si is consistent with the polarity of the pulse being demodulated, and Pi is the relative period for which the respective synchronous demodulator is "on".

1 0 In preference, the calculating further satisfies the following equation: a. for a purely magnetic component of any detected signal, 1•N Gk MnakPk 0 a 15i =l S, and b. for a historical component of any detected signal 2 20 Gk Cn(kPk =0 6 where Gk is the gain of the demodulator passing the Kth contribution, Mncap is the demodulated and averaged magnetic component, due to magnetic soils, between the times a and j following the nth transition from one substantially steady state voltage to at least one other substantially steady state voltage, and Cniap is the demodulated and averaged component resulting from the component of magnetic soils dependent on the historical component, due to magnetic soils, of the applied field between times a and P following the nth transition from one substantially steady state voltage to at least one other substantially steady state voltage, i 9A whereby the output signal is substantially independent of the effect of any non-electrically conducting ferrite constituents existing in the target region.

In preference, the method is further characterised by the target region having a reactive to resistive response ratio component of the magnetic soils within the target regionwhich is substantially independent of an interrogating frequency below 100kHz.

1 0 It is understood throughout this specification that such an output can then be used in combination with other detection techniques and discriminating techniques to further assist in detection of materials.

t' 'In preference, the discrete periods used lie within the range of .5ms to t C* tAn alternating magnetic field metal detector consists of electronic circuitry in 15 which an c 11 e t i ttt i lO alternating current signal is produced which is fed to a transmitting coil, and detection electronics which compares a loaded emf signal induced in a receiver coil to the transmitted signal. The induced signal in the receiver coil results from two sources, namely, from alternating currents flowing in the transmitting coil, and from retransmitting magnetic sources in the local environment under the influence of the transmitted magnetic field.

Consider for the sake of clarity, an ideal situation where the capacitance between windings of both the transmitting coil and receiving coil can be considered negligible.

Also consider that the load presented to the receiving coil by the detection electronics is effectively infinite unless otherwise stated. Furthermore consider that induced eddy currents in the transmitter or receiver coil may be considered for the following analysis to be negligible. Thus for the analysis which follows, the received signal may be considered as the induced emf resulting from alternating magnetic flux with an amplitude directly proportional to the rate of change of flux passing through the S° receiver coil.

Under these circumstances, the component of the received signal resulting from 2 currents flowing in the transmitting coil is such that for each Fourier component transmitted, the induced emf in the receiver coil has a phase angle of 90 degrees relative to the component's current flowing in the transmitter coil. This induced component will be called the "'magnetic'" or "reactive" component.

Any induced received Fourier component with the same phase as the transmitted L i current or exactly the opposite phase depending on the sign sense, will be called the "loss" or "resistive" component. Received signals resulting from local retransmitting environmental sources induce both resistive and reactive components in the receiving coil. Two sources dominate in most ground. One results from ironstone in which the reactive component is usually much greater than the resistive component (usually by S more than 30 times), and the second results from mildly electrically conductive C' components such as moist ground containing salts.

The most difficult ground for detecting highly conductive metal targets, such as coins, gold, underwater pipes etc., is that containing large concentrations of ironstone whose resistive to reactive ratio varies spatially, and worse still if the ground also contains moderately electrically conductive components. The best of the existing metal detectors transmit a roughly sinusoidal signal (distortion -20dB) at between a few component aue to tne effects of the alternating magnetic flux upon magnetic j A soils within the target region and the other from a historical component due J I to the effects of the alternating magnetic flux upon magnetic soils within the /2 11 kHz to a few 10's of kHz. The received signal is synchronously demodulated and passed through a low-pass filter to remove both noise and carrier related signals.

Such art shall be called single frequency detectors.

The phase angle of the synchronous demodulator is set so that the detector is insensitive to components with a phase near to the reactive component, but offset by usually less than several degrees towards the resistive component. This phase angle can be varied, and is so in most detectors manually by means of the user varying a potentiometer. The user is said to "ground balance" the detector to a local area of 1 0 ground, so that the detector is relatively insensitive to the local area. This occurs when the detector's demodulator "object" channel reference phase is at quadrature to the resultant ground vector of the local reactive and resistive vectors. The "resultant ground vector" varies spatially owing to both variations in the mildly conductive component as well as variations in ironstone resistive component relative to the 1 5 reactive component of the soil. This adjustment need be made frequently for best results.

Another type of metal detector transmits a magnetic transient pulse rather than a sinusoidal signal. These are called pulse induction detectors. Usually a voltage of the order of 10 volts is switched on to the transmit coil though a series resistor for j longer than several milliseconds, whereafter the switch is turned off and the resulting back e.m.f. from the transmit coil is clamped by means of a voltage clamp such as a zenner diode. The voltage is clamped typically to a few hundred volts. Both the receive coil and transmit coil are critically damped with resistors. In some detectors only one '25 coil is used to both transmit and receive. Decaying eddy currents following this pulse I.I induced in conducting metal target objects induce signals in the receive coil and the amplified receive signal is gated by means of a S.D. to an averager and thence passed to indicator means.

4 30 One method for substantially reducing the signals arising from the ground is described in provisional patent application" AU PH7889/86 and AU PJ991/88. In these applications an apparatus adapted tc transmit at least two substantially sinusoidal signals is described. The apparatus is adapted to select linear combinations of reactive and resistive signals of at least two transmitted signals such that the mildly conductive ground components are substantially cancelled or the ironstone "resultant ground vector" is substantially cancelled, or both, while maintaining sensitivity to interrogated target objects. Such art shall be called multiple frequency detectors. However, even though in practice M.F. detectors are a considerable improvement on the S.F. detectors described above, M.F. detectors require a substantial number of stable and accurate electronic components which may need fine trimming and hence these are costly. No commercially available P.I. detectors have ground balance features let alone intrinsicall' -'r,cel the ground components. A provisional patent AU PJ2261/89 describes means of achieving both these features utilising P.I. techniques. However, P.I. detectors have a disadvantaged in that the power supply requirements are substantially greater than either S.F. or M.F.

counterparts.

1 0 The subject of this invention is a metal detector apparatus in which the voltage signal applied to the transmit coil is essentially square in shape, where the period between transitions may consist of several different periods in a basic sequence. As will be shown, power requirements can be kept low. Induced signals in the receive coil are amplified and synchronous demodulated and thence passed by low-pass filters.

1 5 These low-pass filtered signals are then processed further for interpretive means. It is easiest to consider such apparatus from the point of view of time domain analysis rather than frequency domain or phasors as is easiest for S.F. or M.F. art. Time domain analysis of pulsed transmit signal detectors can be adapted to be considered as counter-parts to the sinusoidal transmitting S.F. of M.F. apparatuses, although the final '*20 response to interrogated objects are intrinsically different. Means of this implementation is the subject of this invention, where the apparatus will be categorised as that is "time domain detectors." Furthermore, the apparatus described herein can be adapted to advantage over existing art to furnish information about the conductive nature of the interrogated metal target object, and hence to discriminate between different target objects.

To assist with an understanding of the present invention, definitions and back ground physics will be presented: "First order objects" refers to target objects that can be represented magnetically as a 6 ,C single inductor L, to which the interrogation field is loosely coupled, loaded with a single resistor R. For these objects the characteristic frequency w is defined as R/L.

Mildly electrically conducting soils, not containing "magnetic soils" such as ironstone can be represented by a continuum of first order objects where the distribution of w on scale sizes of the order of 1m is only significant at high frequencies and peaks at the order of 1 MHz.

j 4 r 13 "Ground channels" refer to low-pass filters and associated S.D.s whose outputs are substantially sensitive to magnetic soils. (In the case of S.F. detectors these are called reactive channels.) "Object channels" refer to low-pass filters and associated S.D.s whose outputs are substantially insensitive to magnetic soils (but not necessarily c mpletely) and are relatively sensitive to electrically conducting metal target objects. (In the case of S.F.

detectors these are called ground balanced or resistive channels.) It is important to note that received purely magnetic component follows the applied field "instantly," that is, the value of the purely magnetic component is simply proportional to the instantaneous value of the applied field. Thus it is possible to cancel out the purely magnetic component by selecting S.D. gains and references with 1 5 respect to the transmitted signal such that the summed averaged contribution from low-pass filtered S.D. is substantially nulled (non-trivially) to interrogation of the purely magnetic component. This is in fact achieved in S.F. detectors by the appropriate *setting of the ground balance control. This said null is achieved by virtue of the fact that the received response is predictable, except for the basic magnitude which is *20 dependent on the amount, position and type of material within the influence of the interrogating field, which may be achieved by addition of one averaged portion of the received response to another averaged but different portion of the received response, such that the two said averages are substantially equal but opposite in sign. In the case of half wave S.D.s in S.F. detectors two such said portions could be the low- S"2 5 passed synchronous demodulation taken between t=-i7n/(2w) and t=0 added to that taken between t=0 and t=nn/(2w) where k sin(wt) is the transmitted signal, or in other words, the low-passed synchronous demodulation taken between t=-7in/(2w) and 't=7tn/(2w) is nulled to the purely magnetic component.

In contrast to this, received loss components are dependent on the history of the applied field and are not proportional to the instantaneous applied field. However it is possible to substantially effect a null to any loss component which responds in a linear predictable manner to a specified transmitted signal. This can be achieved in a similar way to effecting a null to the purely magnetic component: namely; by selecting S.D.

gains and S.D. references with respect to the transmitted signal such that the summed averaged contribution from low-pass filtered S.D. is substantially nulled (non-trivially) to interrogation of the said loss component. This said null is achieved by virtue of the fact that the received response is predictable, except for the basic magnitude which is N "r o 14 dependent on the amount, position and type of material within the influence of the interrogating field, which may be achieved by subtraction of one averaged portion of the received response 'rom another averaged but different portion of the received response, such that the two said averages are substantially equal.

Furthermore it is possible to effect a substantial (non-trivial) simultaneous null to two or more predictable linear components, for example the purely magnetic and a specific loss component by a similar method described above. This said null is achieved by virtue of the fact that the received response is predictable, except for the basic magnitude of each said component, which are each dependent on the amount, position and type of each material within the influence of the interrogating field, which may be accommodated by selection of a linear combination of three different averaged portions of the received response with respect to the transmitted signal. It should be noted that as at least three degrees of freedom are required to achieve this said null, at least two degrees of freedom (ignoring magnitude) must be available from the *o transmitted signal in the frequency domain. For example, single sinusoidal transmitted signal would not suffice as this only has one degree of freedom (ignoring magnitude and phase), but two sinusoidal signals would suffice or any signal whose Fourier components consist of more than one frequency component, such as a square wave.

We have noted that non-electrically conducting soil has reactive to resistive component ratios that are frequency independent to within a fraction of a percentum below frequencies of the order of 100s of kHz. This is typical property of all ferrites and indeed the major magnetic soil contributor is a ferrite, namely Fe30 4 It is assumed ,225 that H is sufficiently small to substantially avoid any magnetic field saturation. The j relaxation temporal characteristics of all ferrites may be described as a multiplication constant, K, times a function of time, for the same change in the applied magnetic field. It is important to note that F(t) is the same for for all ferrites for the same applied field history except in the high frequency domain. K is different for different chemical ferrites, all else being equal, and indeed is dependent on the quantity and proximity of c V a particular ferrite to the transmit and receive coils. These "relaxation temporal characteristics" following (not instantly) a change in the applied field shall be called "ferrite relaxation tails," or F.R.T.

For all ferrites, the reactive (purely magnetic) component is much larger than the i| resistive component, and is substantially frequency independent below 100kHz. Thus materials for which the reactive to resistive component ratio is frequency independent also have the property that the loss per cycle about the loop per interrogation L. magnetic intensity is substantially frequency independent at frequencies below 100kHz. K is proportional to the loss per cycle, all else being ,qual.

Thus it is possible to cancel both the loss component and purely magnetic component of all ferrites simultaneously as these components are predictable, except for the magnitude of each component, so long as the transmitted signal has more than one frequency component.

In the analysis which follows, for the sake of simplicity, it is assumed that the electronic arrangement consists of a transmit signal generator connected to a transmit coil and receiver coil which are substantially nulled to the transmit coil in "free space." A preamplifier is connected to the receive coil. S.D.s follow the preamplifier which are synchronized to the transmit pulses, and low-pass filters are connected to the outputs of the S.D.s. Also, the systems are linear, and the low-pass filters are temporally 1 5 matched. Further the low-pass filters filter out transmit related Fourier components, and that the outputs of the low pass filters are connected to further processing means and a o. thence the output indicator means.

S.D. "gain" refers to the relative gain of one S.D. compared to the others. This is .20 defined as the relative signal change at the output of the low-pass filter connected to the said S.D. resulting from a change in the mean input to the said S.D. during the "on" period of the said divided by the effective gain of the said low-pass filter. Here the effective impedance of the S.D.s connected to the said low-pass filter need be taken into account to calculate the said low-pass filters gain. For this said definition, all other 25 S.D.s connected to the said low-pass filter must have unchanging inputs during their respective "on" times.

S '.In the descriptions of T.D.D.s which follow, there are 2 possible voltage values; +V1 and -V2. The time period between a voltage transition between one and another said applied voltage value in all cases may consist of several different period values, but in all cases the resultant pulse sequence is of fixed repetitive pattern. The voltage transitions between one value and another is small compared to the duration for which the said applied voltages are applied to the transmit coil. The S.D. "on" and "off" times are synchronised to the transmit signal and also may consist of several different periods, where the "on" and "off" sequence is of fixed repetitive pattern. Table I contains a summary of different novel combinations utilising these possible T.D.D.

concepts.

I

j 1 CASE FEATURE number (c) I 4 x x 11 X x III X X Table I.

where: Feature donates "conventional" S.F. type ground balance facility.

Feature donates several S.D. reference periods in a repetitive sequence effecting insensitivity to F.R.T..

Feature donates several transmit and S.D. reference periods in a repetitive 1 5 sequence effecting insensitivity to F.R.T..

In order to avoid sensitivity to magnetised ground or the earths magnetic field as the detector interrogation head is passed across the ground, and to accommodate the well known electronic drift problem resulting principally from noise, the receiving o °.20 preamplifiers are A.C. coupled. The problem with any filtering such as A.C. coupling is a0 that received signals contain memory of preceding signals which then contaminate future signals. Part of this memory is derived from (non-linear) preamplifier slew rate limited periods following voltage transitions during transmission.

In all the T.D.D. concepts which follow, unless otherwise specified, it is assumed that the following concept is incorporated, namely: a There will be at least two S.D.s where the sum of, the gain of each S.D. times it's total "on" period, summed over all the contributing S.D.s is substantially zero. To effect this, the contribution from some S.D.s is of opposite sign to others, that is there is an It;, 30 effectively subtraction. This can be achieved by means of a subtractor, or by the preamplifier feeding the inputs of some S.D.s and an unity gain invertor amplifier following the preamplifier feeding the inputs of other S.D.s, where the demodulated signals are then added. In summary, a linear sum of the signals passed by each of the said synchronous demodulators is added such that n GiPi 0, (i) i=1 where Gi is the relative effective gain of the ith synchronous demodulator which passes it's output to be added and low-pass filtered, where the sign of Gi is consistent with the "phase" polarity of the pulse being demodulated, and Pi is the relative period for which the said synchronous demodulator is Under these conditions, it is not necessary for the amplifiers to be A.C. coupled. Such a system, if electrically balanced, cancels out any induced signals resulting from the movement of the interrogation coils relative to static magnetic fields, nor does it suffer from electronic drifts, while maintaining no loss in sensitivity to target objects. In other words, the net final lowpassed signal contains no component resulting from asynchronous received signals.

Another problem with the current art of pulsed detectors is their inability to effect ground balance. This is mainly due to the A.C. coupling effects described above.

However as stated above, ironstone produces relaxation tails whose magnitude varies according to the soil being interrogated. Effecting universal balance to soil is the 1 5 principal subject of this patent.

o, Theory The generalised response from the purely magnetic component will now be 0 determined and a generalised equation determining means to select the proportions of contributions that are to be low-pass filtered, which are derived from synchronously demodulating different periods of the received waveform with respect to the transmitted waveform will be defined.

.25 Let the time constant of the generating transmitter-cum-transmit coil be If the transmit voltage transitions from one polarity to the other occur at t=T1, 2,2, 3, then the received e.m.f. across an effectively unloaded but critically damped receive coil (which may be amplified but uncompensated in the frequency domain) resulting from the interrogation of only the purely magnetic component which when demodulated and averaged between the nth and (n+1)th transition yields a signal Scontribution Mnap proportional to s 13 n n Mnap J n exp(-lTj) }exp(-Qt)}dt. (ii) t=a i=1 j=i t. where t=0 .t the nth transmit coil voltage transition, and the demodulator is enabled between t=cc and t=p, where there are no said voltage transitions between a and p.

^3 i0 3. A mnndiri itinn matoI rfio;m;n u This contribution in practice is fed to the averager, that is a low-pass filter, on a repetitive basis, and is in general added to other such contributions derived from different demodulated periods of the received signal with respect to the transmit signal which are also fed to the said low-pass filter on a repetitive basis where the proportions of the said contributions are selected such that net average is substantially nulled to the purely magnetic component. That is;

I

X GKMnaK3 0 (iii) K==1 where the most basic but complete sequential repetitive transmit signal cycle consists of m periods between m transitions, where there are a total of I different contributing demodulated periods in each said cycle where GK is the gain of the demodulator passing the Kth contribution to the said average which is demodulated between t=a and t=P 3 K where cx and 3 lie between two transitions in the said basic sequence and has a sign consistent with the demodulated "phase," and t is defined as zero at the first of the said two transitions (note; t=0 m times in the basic cycle). A null to the purely St, magnetic component occurs when the GK are selected so that (iii) is satisfied.

t The generalised response resulting from only the loss component of ferrite will now be determined: If g(t)=exp(-wt) and f(t)=exp(-Qt)-g(t) the received e.m.f. across an effectively unloaded but critically damped receive coil resulting from interrogation of ferrite ignoring the purely magnetic component which when demodulated and averaged between the nth and (n+1)th transition yields a signal contribution Cna3 proportional to V 4 S o n n Cnap3 f f(Ti) j (g(tj)]{wexp(-wt)-Qexp(-.t)}/(Gw(w-K (iv) 30 t=a w=0 i=1 j=i r rit S "This contribution in practice is fed to the averager, that is a low-pass filter, on a repetitive basis, and is in general added to other such contributions derived from different demodulated periods of the received signal with respect to the transmit signal which are also fed to the said low-pass filter on a repetitive basis where the proportions of the said contributions are selected such that net average is substantially nulled to to the loss in ferrite. That is; i- i 19 SGKCnoKpK 0 K=1 If both (iv) and are satisfied, then the final said averaged contributions yield a null to ferrite.

Examples of electronic arrangements using different transmit waveforms.

There are several different concepts of T.D.D. described below which are listed case by case. In each case it is assumed that equation is satisfied, that is the T.D.D. is adapted to be substantially nulled to asynchronous components. It must also be noted that one complete basic transmit waveform sequence is shown plus the beginnings and ends of the prior and subsequent cycle sequence.

Case I.

A novel means of effecting ground balance, which can be considered as a pulsed *counterpart to S.F. art, is to select a linear combination of an object channel signal and 20 a ground channel signal. The addition of these two signals may occur at the S.D.-lowt s pass filter interface, as described in patent application AU PH7889/86.

To assist with the understanding of the present invention, reference will now be made to the accompanying illustrations.

Figure I shows a basic electronic block diagram of a means to effect a variable ground balance control according to a preferred embodiment.

Figure II shows voltage waveforms at various stages in figure I for case I.

Figure III shows a basic electronic block diagram of a means to effect intrinsic ground balance control according to a preferred embodiment.

Figure IV shows voltage waveforms at various stages in figure III for case !1.

Figure V shows voltage waveforms at various stages in figure I for case III.

Referring to the drawings in detail it is now noted as follows: j A means for implementing a variable ground balanced T.D.D. concept are shown in figure I and voltage waveforms at various points are shown in figure II. A receive coil 1 is connected to a preamplifier 3 and loaded with a damping resistor 2. The output of 3 Si i l I 32 7. A conducting metal discriminating detection apparatus as in claim 5 or is connected to an inverting amplifier 4 and to the inputs of S.D.s 5 and 7 (solid state switches). The output of 4 is connected to the inputs of S.D.s 6 and 8. The references to 5, 6, 7, and 8 are synchronized to the transmit pulses. (In figure II a high state turns the S.D. "on" and a low, 5 and 6 are enabled by 9 and 10 after a short delay subsequent to a transition in the transmit pulse and disabled substantially before the next transmit voltage transition. 7 and 8 are turned "on" during a later period than either 5 or 6 of a "steady" voltage state of the transmit pulse, but 7 and 8 are turned "off" before the termination of the said "steady" voltage state. 5 and 7 are turned "on" during one polarity while 6 and 8 are turned "on" during the other. The transmit coil 18 1 0 has a square waveform 20 applied by generator 19 to it. The receive signal at the output of 3 is shown as waveform 21 and the waveform of 9 is shown by 22, and that of is shown by 23. The waveform of 11 is shown by 24, and that of 12 is shown by The outputs of 5 and 6 are connected to each other and to resistor 13. The outputs of 7 and 8 are connected to each other and to variable resistor 14, which is used to control 1 5 the ground balance. The wiper of 14 is connected to resistor 15 to which resistors 13 and capacitor 16 are connected, at which junction an output 17 may be adapted for further processing. Capacitor 16 is grounded, thereby 13, 14, 15 and 16 form a lowpass filter and adder whose relative proportions may be varied by adjustment of 14.

9 o 14 may be adjusted to pass to 17 relatively more or less ground signal, which is substantially manifest at the preamplifier output during transmission. This is used to "ground balance" the T.D.D. by adding or subtracting an averaged proportion of the e soil's magnetic component from the averaged F.R.T which is manifest together with the purely magnetic component. This arrangement of ground balance is a counterpart to "25 that found in S.F. art.

o Note that the waveform 21 shows an exponential droop between transmit voltage transitions. This results from the exponential decay in E.M.F. across the "pure" 1 inductive component (Lt) of the transmit coil which has a finite "series" real resistance ,30 The exponential decay time constant is thus Rt/Lt. The S.D. 5 "on" duration shown by 22 in figure il must be very similar to that of 10 shown by 23 to effectively cancel static magnetic fields.

,i Their are several advantages to this novel T.D.D. compared to high quality S.F.

detectors: Firstly, it requires less expensive electronic components and secondly, it can be adapted to be relatively insensitive to conductive soil components such as salt water if the S.D. are "off" during a period following a transition in the transmit voltage waveform. This is due to the fact that the eddy currents induced in conducting soil 21 components decay rapidly compared to valuable target conducting metal targets.

(Note; this is only true in the situation where the size of the target metal object is more than of the order of 0.001 times in (linear) dimension the size of the transmit and receive coil.) Significant advantage may be gained to effect a ground balanced channel by selecting a portion of the receive waveform during the first quarter or less, a first period, of one transmit polarity (but not too close to the transmit transition), subtracted from an appropriate proportion of the receive signal during the last quarter or less, a second period, of the said transmit polarity. These signals are low pass filtered together with the corresponding contributions from the other transmit polarities, but these are selected with the opposite polarity sense. This exaggerates the difference between the initial peak of the eddy currents following a voltage transition and the latter part of the decayed eddy currents. If these periods are selected to be very short, 1 5 then electronic noise may become significant. Thus a useful compromise is to select approximately an eighth of period of the transmit signal of each polarity for each said period, the first occuring after a delay time suitable for the milo!y conductive soil eddy currents to effectively settle, and the second may occur immediately or just before the transmit signal transition.

Despite these advantages, this concept does not cancel both the magnetic and ferrite loss ground components. The T.D.D.s can be adapted to achieve this property by "C several different arrangements, which may be considered as T.D.D. counterparts to ,tT C The M.F. apparatus described above. In essence, as the relaxation temporal 25 characteristics are the same for all ferrites for the same change in the applied magnetic field (see above), then for a given transmit waveform, signals arising from interrogated ferrite produce a predictable function of time except for the scaling factor which is dependent on the amount of matter under the influence of the interrogating o field. Also, as the decay function is dependent on the history of the applied field, it is S 30 possible by controlling the transmit signal to determine the subsequent ferrite response function, and hence by selecting a linear combination of contributions arising from different time periods with respect to the transmit pulse, or from different periods following different transmit pulses, or both, to cancel out the signals produced by ferrite. This can be achieved by a number of different ways, M.F. art being just one special case. Here, in the examples given, the voltage applied to the transmit coil is either constant in value for periods in time, or short transients. This is because these are easy and inexpensive to produce electronically. In principle though, the wave form may be any function of time, and signals from ferrite may be cancelled.

.it 22 Case II In this case a novel T.D.D. concept in which there is cancellation of both the magnetic, conductive soil and and ferrite loss soil components is described, which may be considered as T.D.D. counterparts to the M.F. apparatus. One simple arrangement of such a system may be described by figure III where the components perform the same task as described in the above case I except for the addition of two S.D.s; hence the same labels are used for the common parts. The associated voltage waveforms are 1 0 given in figure IV.

The output of 3 is also connected to the input of S.D. 26 and the output of 4 is connected to the input of S.D. 27. The control of 26 is a signal at 28 and the control of 27 is a signal at 29. The outputs of 26 and 27 are connected together and to a variable 1 5 resistor 30 which is connected in series with a fixed resistor 31. 31 is connected to the output node 17. The low-pass filter and adder now consists of 5, 6, 7, 8, 26, 27, 13, 14, 30, 31, and 16 whose relative proportions may be varied by adjustment of 14 and 30. 19 produces a square voltage waveform 32, similar to 20 in figure II. The receive waveform, 33 is likewise similar to 21. The waveform of 9 is shown by 34, and that of 2 20 10 is shown by 37. The waveform of 11 is shown by 38, and that of 12 is shown by The waveform of 28 is shown by 36, and that of 29 is shown by 39. As shown in figure III, the time variable T is zero at the transition of the transmit voltage, and T=T1 at the 4 turn "on" transition of 5 (see 34). The turn "off" time of 5 occurs at T=T2 and the turn tc "on" time of 8 (see 35) occurs at T=T3 and then the turn "off" time of 8 occurs at T=T4.

c '25 The turn "on" time of 26 (see 36) occurs at T=T5 and then the turn "off" time of 26 occurs at T=T6. Similarly, the time variable is defined for the opposite transmit polarity period, for the associated S.D.s.

Both 14 and 30 adjust the degree of the purely magnetic component ground balance ,30 and F.R.T. ground balance. This T.D.D. can be adapted to be relatively insensitive to conductive soil components such as salt water if the S.D. are "off" during a period following transitions in the transmit voltage waveform for reasons given above.

Figure III is only one means of implementing this Case II T.D.D. concept.

Also note that it is possible to effect the same result by selecting a linear combination of two low-pass buffered outputs whose respective inputs are a linear combination of different demodulated periods between transmit transitions (the low-pass filters need b1 23 be temporally matched).

The principal advantages of this T.D.D. compared to M.F. detectors is that it requires less expensive electronic components.

The interrogation of a first order object results in a signal which is a function Obj(w) of w, which is zero for certain values of w for one or more values of w. Obj thus has a null response to certain target objects (but is not nulled to most). Different linear combinations of the proportionality constants of the said linear combinations yield 1 0 different values of w to which the resulting different Obj(w) are nulled. Thus in order to effect no nulls in the w domain, different Obj outputs need be combined in such a way that activity in an Obj oi'tput is in some way addressed. One way of achieving this is by full-wave rectification of each Obj signal, each of which may be high-passed by 4 respective high-pass filters first, and then adding these full-wave rectified signals.

1 5 Another way is again full-wave rectification of each Obj signal, each of which may be high-passed by respective high-pass filters first, and then to pass each of these fuli- Si c wave rectified signals to a selector which passes the largest instantaneous rectified I I signal to the output. This final output then may be further low- and high-pass filtered.

The final output can be used to control an audio output or trigger yet further electronics if a threshold is exceeded.

Thus it is possible to combine channel outputs in many ways to yield outputs relatively 'f insensitive to soils and yet maintain sensitivity to a large range (in w) of metal target t| objects.

If Case III There is yet another novel T.D.D. concept in which there is cancellation of the magnetic, conductive soil and and ferrite loss soil components which may be considered as T.D.D. counterparts to the M.F. apparatus: This case III form transmits at least two different pulse periods in the transmit pulse sequence. It has significantly better ability to discriminate between objects and soils than those described in the cases above, especially if the pulse periods differ more than a factor of four from each other in period length, where the net period of each transmitted period is similar. That is, several short periods are transmitted for every long period in the pulse sequence.

This yields a transmit frequency spectrum that has significant power at both low and high frequencies, unlike a simple square-wave which has powers decreasing as an inverse of the order of the (odd) harmonics. For example the second most powerful "4 B 1: 1 e. [i I 4 24 harmonic for simple square waves is the third which is three times weaker than the fundamental, and the fifth is a mere fifth of the fundamental. In addition, these said harmonics for simple square waves are too close to each other to give signals from a range of metal target objects that are grossly different to ferrite relaxation tails. Such systems may be implemented by very similar amplifiers and synchronous demodulators and low pass filters as the cases above. An example of a possible associated voltage waveform is given in figure V, and electronic configuration in figure V1.

1 0 Four quarters of a transmit sequence is shown in figure V. The first shown quarter can be considered to start at Q1, the second at Q2, the third at Q3 and the fourth with Q4, thereafter the sequence repeats itself continuously. The EMF supplying the transmit coil produces a voltage waveform 40. Switch synchronous demodulators select the non-inverted received signal to be passed to the "averaging" low pass filter when 15 signal 45 is high, and the inverted received signal when 45 is low; 45 controls the switch 71 input at 70 which selects the "phase." A first low passed signal substantially balanced to the purely magnetic component may be formed at the output of a low-pass Ot filter cum demdulator combination by passing the output of a synchronous demodulator switch controlled by signal 41 together with another controlled by 43.

This may be achieved by applying 41 to synchronous demodulator 46 at it's control input 48, and 43 to synchronous demoduiator 47 at it's control input 49. Resistors and 51, together with capacitor 52 form a low-pass filter where the signals are combined. The ratio of 51 and 50 may be selected such that the output of the said low pass filter is substantially balanced to the purely magnetic component. Such a lowpassed signal at 53 will be rich in high frequency loss components. A second low passed signal substantially balanced to the purely magnetic component may be formed at the output of a low-pass filter cum demodulator combination by passing the output of a synchronous demodulr.tor switch controlled by signal 44 together with :another controlled by 43. This may be achieved by applying 44 to synchronous demodulator 58 at it's control input 56, and 43 to synchronous demodulator 55 at it's control input 57. Resistors 58 and 59, together with capacitor 60 form a low-pass filter where the signals are combined. The ratio of 58 and 59 may be selected such that the !output of the said low pass filter is substantially balanced to the purely magnetic component. Such a low-passed signal at 61 will be rich in high and more particularly low frequency loss components. A third low passed signal substantially balanced to the purely magnetic component may be formed at the output of a low-pass filter cum demodulator combination by passing the output of a synchronous demodulator switch controlled by signal 42 together with another controlled by 43. This may be achieved

-I_

i by applying 42 to synchronous demodulator 62 at it's control input 64, and 43 to synchronous demodulator 63 at it's control input 65. Resistors 66 and 67, together with capacitor 68 form a low-pass filter where the signals are combined. The ratio of 66 and 67 may be selected such that the output of the said low pass filter is substantially balanced to the purely magnetic component. Such a low-passed signal at 69 will be rich in low frequency loss components. Yet others may be formed. For example, a low passed signal substantially balanced to the purely magnetic component may be formed at the output of a low-pass filter cum demodulator combination by passing the output of a synchronous demodulator switch controlled by signal 41 together with 1 0 an.t.her controlled by 42, except that the opposite received signal sign phase sense is selected by 42. Such a low-passed signal will be rich in both high and low frequency loss components. It should be noted that this low-passed input has a null to objects with a particular characteristic frequency and the output is relatively more independent of F.R.T. than any of the other low pass outputs described above. In fact if the resistors 1 5 and sampling periods are selected appropriately, this said low passed output may be substantially independent of both the purely magnetic component and ferrite loss 0 Ol a components. In all 3 linearly independent channels may be formed that are substantially balanced to the purely magnetic component. Linear combination of any two of these may combined to form signals substantially balanced to the ferrite loss 040 20 component as well as the purely magnetic component.

Numerous other combinations of pulse period and sequence are possible for the basic transmit sequence, so long as the T.D.D. is adapted to cancel the three background soil components. Individual S.D.s may be enabled only during transmit pulses of a particular period lengths as shown in all the waveform figures, or they may be enabled 04 during transmit pulses of various period lengths.

As in case II, it is possible to select linear combinations of time averaged currents flowing through the low-pass filter resistor of the jth S.D. pair for a given demodulation 130 waveform pattern, by selecting S.D. gains (resistor values in figure 1, III and VI), and demodulator reference waveforms so that the the time averaged signal of demodulated signals from at least 3 S.D. pairs, each synchronously demodulating the received signal with respect to a different part of the transmitted basic waveform Ssequence result in a low-passed output that satisfies the conditions for insensitivity to insensitivity to salt and the purely magnetic component. As shown in figure V, in order to obtain the best signal-to-noise, the total time spent transmitting short pulses should be similar to the total time transmitting long pulses.

?A

i :91 i rL It can be shown by theory (and practice) that high frequency and especially low frequency loss components are most accentuated at the beginning of transmit periods, which follow long transmit periods for which the transmit coil's current is relatively large at the end of such a period. Low frequency but not high frequency loss components components are accentuated following a medium delay after a transmit transition during long transmit periods, which follow long transmit periods for which the transmit coil's current is relatively large at the end of such a period. High frequency loss components are most accentuated at the beginning of transmit periods which follow short transmit periods for which the transmit coil's current is relatively small. The 1 0 purely magnetic component is most accentuated at the end of transmit periods which commence with relatively small transmit coil currents. Thus substantial advantage can be gained by passing signals via synchronous demodulation that accentuate each of these four said components. This is illustrated in the above example.

S1 5 A selected proportion of said purely magnetic balanced channels at the outputs 53, *fez 61, and 69 may be subtracted one from another to yield an output that is insensitive to ferrite relaxation tails. Thus 2 linearly independent such outputs can be obtained, 3 "different" outputs all together. The highest instantaneous absolute value of the highv passed outputs may be passed to the final output indicator means. For simplicity, channels 53, 61 and 69 need not be formed explicitly before subtraction of one from another to yield a final universally ground balanced signal; the universal ground balanced channels may be formed directly by passing at least three appropriate 'etv synchronous demodulator signals to the low-pass filter. Each channel balanced universally to ferrite can in fact be implemented by selecting just two appropriate synchronous demodulator signals as described above.

If these said receive synchronous demodulation periods are selected to be very short, then electronic noise may become significant. Thus a useful compromise is to select between an eighth and a quarter of period of the transmit signal to be passed to the 3 0 low pass filter of each polarity for each said period, the initial periods occuring after a delay time suitable for the mildly conductive soil eddy currents to effectively settle, and the latter may occur immediately or just before the transmit signal transition.

Furthermore, the clearest indication of whether an object is ferrous or not is manifest at the end of long transmit pulse periods which follow short transmit pulses during which periods the transmit coil current is relatively low. Thus if ground channels obtained by synchronous demodulation which pass received signals at the end of long transmit periods -that commence with small transmit coil currents and which follow short S ~MVt~t!tL~~ ~2Th., t transmit pulses are compared to purely magnetically balanced channels, discrimination as to whether an interrogated metal object is made of iron or not is enhanced to advantage. The well known techniques such as sign comparison may be employed for purposes of ferrous/non-ferrous discrimination; the most important feature of metal detectors used for coin and treasure searching.

As the sensing transmit and receive coils are passed by different soils, their inductance is modulated by the varying permeability of the soils. This changes the quality factor which in turn varies the current flowing through the transmit coil. This 1 0 is most acute at the end of long transmit pulses or when the transmit coil current is relatively large. As a result this must be compensated for or else the effective "gain" of the coil will be modulated as the inductance is modulated by the varying interrogated soil. The simplest way of doing this is to measure the current flowing through the coil and compensate for it's changes. The current may be measured by monitoring a S1 5 voltage across a small valued resistance or inductance (which may be a "current" transformer) in series with the transmit coil which either deceases or increases with increasing Q depending as to where in the transmit cycle the current is measured.

Another way is to have a third receive coil well magnetically coupled to the transmit coil. The EMF across this coil then indicates the Q which increases with increasing Q.

The monitored transmit coil signal can then be used to multiply appropriate receive signals so as to effectively cancel the effects of changing Q.

It should be noted that in all the above cases, It is never possible to cancel the purely magnetic nor the loss components absolutely. Firstly no electronics is perfectly linear (typically more than 0.1% non-linearity), and secondly, all electronics drifts with time.

Thus balanced signals become unbalanced. Thus it is necessary to compensate for these drifts by adjusting the gain of at least one component of the input signal passed to the low-pass filters. This can be achieved by adjusting resistor values described V above at the low pass filter stage or adding a selected varied amount of the ground signal to "substantially" ground balanced signal. This can be done automatically along very similar lines to that described in provisional patent AU PJ991/88.

It should be noted that in all the above cases so long as only small voltages appear between the transmit generator power supply rails and the transmit coil when current is flowing into or out of the said power supply rail and the net averaged A.C. transmit flux is zero, then the power consumption will be relatively small and will depend on Q and the transmit pulse lengths. The reason for this is fundamental to basic electronic principles and is very well established in numerous electronic applications, such as

II

28 switch mode power supplies. However, unipolar flux pulses may be transmitted with the condition that only small voltages appear between the transmit generator power supply rails and the transmit coil when current is flowing into or out of the said power supply rail, if in addition a switch-mode power supply is utilised to return net current "dumped" by the transmit coil to one power supply reservoir to the source power supply.

I* tt t t It

C

S

l

Claims (13)

1. A conducting metal discriminating detection apparatus for detecting a metal target in a target region comprising: transmission means for transmitting a continuous pulsed voltage waveform to provide an alternating magnetic flux in the target region, the continuous pulsed voltage waveform comprising abrupt transitions from one substantially steady state voltage to at least one other substantially steady state voltage; a detector coil for detecting a rate of change of magnetic flux in the target region, the rate of change of magnetic flux being a function of the alternating magnetic flux in the target region; measurement means for measuring at least two time averaged 1 5 measurements of the detected rate of change of magnetic flux at spaced apart time intervals, the measurement means being adapted to measure the time averaged measurements after a duration of an initial time interval such that the initial time interval starts directly after one of the abrupt transitions and finishes when the rate of change of eddy currents, due to the effects of S 20 the alternating magnetic flux upon conductive soils within the target region, is substantially zero; processing means for comparing the time averaged measurements against two pre-determined responses, one resulting from an instantaneous component due to the effects of the alternating magnetic flux upon magnetic soils within the target region and the other from a historical component due to the effects of the alternating magnetic flux upon magnetic soils within the target region; and assessment means for providing an assessment signal from the comparison of the time averaged measurements, the assessment signal 3 0 being indicative of the metal target within the target region.

2. A conducting metal discriminating detection apparatus as in claim 1 in which the measurement means comprises: synchronous demodulator means for synchronous demodulating the rate of change of the magnetic flux at spaced apart time intervals; and low pass filter means for low pass filtering the synchronous demodulated rate of change of the magnetic flux at spaced apart time intervals. I- 770 F,*

3. A conducting metal discriminating detection apparatus as in any previous claim, further characterised by the transmission means providing the continuous pulsed voltage waveform such that the assessment signal is substantially independent of the reactive to resistive response ratio component of the magnetic soils within the target region at frequencies below 1 00kHz.

4. A conducting metal discriminating detection apparatus as in any previous claim in which the continuous pulsed voltage waveform comprises 1 0 two different pulse periods, these being a first pulse period and a second pulse period, the first pulse period being substantially of a shorter duration than the second pulse period, and the measurement means is adapted to synchronously demodulate the rate of change of the magnetic flux during a

9. first measurement period occurring in the last quarter of the second pulse 15 period to form a first signal, and the measurement means is further adapted to synchronously demodulate the rate of change of the magnetic flux during a second measurement period occurring in the first half of the first or second pulse period to form a second signal, wherein the second measurement period occurs after the rate of change of eddy currents in the conductive soils is substantially zero, and the processing means is adapted to form a ground balanced channel signal by scaling and subtracting the first signal from the second signal, the processing means being further adapted to compare the sign of the ground balanced channel signal with the sign of the first signal for the assessment means to provide a signal indicative of a ferrous target within the target region. Z A conducting metal discriminating detection apparatus for detecting a metal target in a target region comprising: transmission means for transmitting a continuous pulsed voltage waveform to provide an alternating magnetic flux in the target region, the continuous pulsed voltage waveform comprising abrupt transitions from one substantially steady state voltage to at least one other substantially steady state voltage; a detector coil for detecting a rate of change of magnetic flux in the target region, the rate of change of magnetic flux being a function of the S alternating magnetic flux in the target region; A 31 measurement means including synchronous demodulation means for synchronous demodulating at least three measurements of the rate of change of magnetic flux during at least three spaced apart time intervals, the measurement means being adapted to measure the rate of change of magnetic flux after a duration of an initial time interval such that the initial time interval starts directly after one of the abrupt transitions and finishes when the rate of change of eddy currents, due to the effects of the alternating magnetic flux upon conductive soils within the target region, is substantially zero; 1 0 a filter means adapted to low pass filter the synchronous demodulated measurements of the magnitude of the rate of change of magnetic flux at spaced apart time intervals to provide at least three low pass filtered synchronous demodulated measurements; a processing means adapted to process the low pass filtered 15 synchronous demodulated measurements to provide at least three time averaged values these being a first time averaged value, a second time averaged value and a third time averaged value; a calculating means adapted to: subtract the third time averaged value from the first time averaged value to provide a first resultant value; e subtract the third time averaged value from the second time averaged value to provide a second resultant value; .multiply the second resultant value by a pre-determined value to provide a third resultant value; and 25 subtract the third resultant value from the first resultant value to :provice a value which is substantially zero for all iron oxides; and an assessment means for providing an assessment signal from the subtracting of the third resultant value from the first resultant value, wherein the assessment signal is indicative of a metal object within the target region. 6. A conducting metal discriminating detection apparatus as in claim 5 in which the continuous pulsed voltage waveform comprises two different pulse periods, these being a first pulse period and a second pulse period, the first pulse period being substantially of a shorter duration than the second pulse period, wherein at least one of the spaced apart time intervals occurs in the first pulse period and at least one other time interval occurs in the second pulse period. (t rl3 c -l-I xi' ii 32 7. A conducting metal discriminating detection apparatus as in claim 5 or claim 6, wherein the synchronous demodulation means comprises n synchronous demodulators and is operably configurable to satisfy the equation: n _SiPi 0, i=1 where Si is the relative effective gain of the ith synchronous demodulator the output of which is to be added and low-pass filtered, where the sign of Si is consistent with the polarity of the pulse being demodulated, and Pi is the relative period for which the said synchronous demodulator is "on". 8. A conducting metal discriminating detection apparatus as in any claims 5 to 7, wherein the calculating means is adapted such that the following equation is satisfied: a. for a purely magnetic component of any detected signal, et a.e a. a reeQ a a a a. d a a,.r i a a t a II I i Ce *r a I I I t( t I a=1 Gk Mnaklk =0 and b. for an historical component of any detected signal 0=1 Gk Cnaklk 0 where Gk is the gain of the demodulator passing the Kth contribution, Mnap is the demodulated and averaged magnetic component, due to magnetic soils, between the times a and P following the nth transition from one substantially steady state voltage to at least one other substantially steady state voltage, and i, r*r.o IB 33 Cnap is the demodulated and averaged component resulting from the component of magnetic soils dependent on the historical component, due to magnetic soils, of the applied field between times oa and following the nth transition from one substantially steady state voltage to at least one other substantially steady state voltage, whereby the output signal is substantially independent of the effect of any non-electrically conducting ferrite constituents existing in the target 9. A conducting metal discriminating detection apparatus as in any claims 5 to 8, further characterised by the transmission means providing the continuous pulsed voltage waveform at a frequency such that the assessment signal is substantially independent of the reactive to resistive 1 5 response ratio component of the magnetic soils within the target region when the frequency is below 100kHz.

10. A method of conducting metal discriminating detection of metal target in a target region comprising the steps of: 20 transmitting a continuous pulsed voltage waveform to provide an alternating magnetic flux in the target region, the continuous pulsed voltage waveform comprising abrupt transitions from one substantially steady state voltage to at least one other substantially steady state voltage, and the target region contains magnetic soils and electrically conductive soils; detecting a rate of change of magnetic flux in the target region, wherein the rate of change of magnetic flux results from the rate of change of the alternating magnetic flux, rate of change of eddy currents flowing in the metal target, rate of change of eddy currents flowing in the electrically conductive soils and a rate of change of alternating magnetic flux modified by two components within the magnetic soils these components being an instantaneous component dependent upon the instantaneous values of the rate of change of the alternating magnetic flux and a historical component dependent upon previous values of the alternating magnetic flux; measuring at least two time averaged measurements of the detected rate of change of magnetic flux at spaced apart time intervals, the measuring occurring after a duration of an initial time interval such that the initial time interval starts directly after one of the abrupt transitions and finishes when zero; the rate of change of eddy currents in the conductive soils is substantially jj ;It: zero i::i-glr t i 34 comparing the time averaged measurements against two pre- determined responses, one resulting from the instantaneous component and the other from the historical component, wherein the pre-determined response is the expected response from the target region due to the effects of magnetic soils; and effecting an assessment signal from the comparing of the time averaged measurements, the assessment signal being indicative of the metal target within the target region. 1 0 11. A method of conducting metal discriminating detection as in claim wherein the detecting further comprises the steps of: synchronous demodulating the rate of change of the magnetic flux at spaced apart time intervals; and low pass filtering the synchronous demodulated rate of change of the 15 magnetic flux at spaced apart time intervals. •4

12. A method of conducting metal discriminating detection as in claim or claim 11, further characterised by the target region having a reactive to resistive response ratio component of the magnetic soils within the target region which is substantially independent of an interrogating frequency 4 below 1 00kHz.

13. A method of conducting metal discriminating detection as in any claims 10 to 12, in which the continuous pulsed voltage waveform comprises 25 two different pulse periods, these being a first pulse period and a second pulse period, and the first pulse period is substantially of a shorter duration than the second pulse period, the method further including the steps of: detecting and synchronously demodulating the rate of change of the magnetic flux during a first measurement period occurring in the last quarter of the second pulse period to form a first signal, detecting and synchronously demodulating the rate of change of the magnetic flux during a second measurement period occurring in the first half of the first or second pulse period to form a second signal, wherein the second measurement period occurs after the rate of change of eddy currents in the conductive soils is substantially zero, and i processing to form a ground baanced channel signal by scaling and subtracting the first signal from the second signal, 71 (I I ii _.-ili-4 comparing the sign of the ground balanced channel signal with the sign of the first signal to provide a signal indicative of a ferrous target within the target region.

14. A method of conducting metal discriminating detection comprising the steps of: transmitting a continuous pulsed voltage waveform to provide an 1 0 alternating magnetic flux in the target region, the continuous pulsed voltage waveform comprising abrupt transitions from one substantially steady state voltage to at least one other substantially steady state voltage, and the target region contains magnetic soils and electrically conductive soils; r A Cdetecting a rate of change of magnetic flux in the target region, 15 wherein the rate of change of magnetic flux results from the rate of change of the alternating magnetic flux, rate of change of eddy currents flowing in the metal target, rate of change of eddy currents flowing in the electrically I S, conductive soils and a rate of change of alternating magnetic flux modified o C' by two components within the magnetic soils these components being an instantaneous component dependent upon the instantaneous values of the rate of change of the alternating magnetic flux and a historical component dependent upon previous values of the alternating magnetic flux; synchronous demodulating at least three measurements of the rate of So.. t change of magnetic flux at spaced apart time intervals, the synchronous 25 demodulating occurring a duration of an initial time interval such that the to, initial time interval starts directly after one of the abrupt transitions and finishes when the rate of change of eddy currents in the conductive soils is substantially zero; low pass filtering the synchronous demodulated measurements of the magnitude of the rate of change of magnetic flux at spaced apart time intervals to provide at least three low pass filtered synchronous demodulated measurements; processing the low pass filtered synchronous demodulated measurements to provide at least three time averaged values these being a first time averaged value, a secondj time averaged value and a third time averaged value; 36 subtracting the third time averaged value from the first time averaged value to provide a first resultant value; subtracting the third time averaged value from the second time averaged value to provide a second resultant value; multiplying the second resultant value by a pre-determined value to provide a third resultant value; subtracting the third resultant value from the first resultant value to provide a value which is substantially zero for all iron oxides; and 1 0 effecting an assessment signal from the subtracting of the third resultant value from the first resultant value, the assessment signal being indicative of a metal object within the target region. A conducting metal discriminating detection apparatus as in claim 14 S 15 in which the continuous pulsed voltage waveform comprises two different pulse periods, these being a first pulse period and a second pulse period, the first pulse period being substantially of a shorter duration than the second pulse period, wherein at least one of the spaced apart time intervals occurs in the first pulse period and at least one other time interval occurs in the second pulse period.

16. A method of conducting metal discriminating detection as in any claims 13 to 15, wherein the step of synchronously demodulating is accomplished by n coupled synchronous demodulators which satisfy the following equation: j n 25 _SiPi= 0, i=1 where Si is the relative effective gain of the ith synchronous demodulator the output of which is to be added and low-pass filtered, where the sign of Si is consistent with the polarity of the pulse being demodulated, and Pi is the relative period for which the respective synchronous demodulator is "on".

17. A method of conducting metal discriminating detection as in any claims 13 to 16, wherein the calculating further satisfies the following equation: :i 37 a. for a purely magnetic component of any detected signal, S Gk MnkPk 0 a=1 and b. for a historical component of any detected signal I GkCnakl3k =0 I P =1 j 1 5 where Gk is the gain of the demodulator passing the Kth contribution, I Mnoap is the demodulated and averaged magnetic component, due to I magnetic soils, between the times a and p following the nth transition from S* one substantially steady state voltage to at least one other substantially steady state voltage, and Cnap is the demodulated and averaged component resulting from the t" component of magnetic soils dependent on the historical component, due to p it magnetic soils, of the applied field between times a and 3 following the nth transition from one substantially steady state voltage to at least one other I' substantially steady state voltage, i 25 whereby the output signal is substantially independent of the effect of r any non-electrically conducting ferrite constituents existing in the target region.

18. A method of conducting metal discriminating detection as in any claims 13 to 17, further characterised by the target region having a reactive to resistive response ratio component of the magnetic soils within the target region_which is substantially independent of an interrogating frequency below 100kHz.

19. A conducting metal discriminating detection apparatus substantially 3 5 as described with reference to and as illustrated by the accompanying drawings. -1 38 A method of conducting metal discriminating detection apparatus substantially as described with reference to and as illustrated by the accompanying drawings. DATED this 22nd day of October 1992. BRUCE HALORO CANDY By his Patent Attorneys COLLISON CO 094 44 49 4. toes' 94 4 N9. 4 t 99.4-' L