AU2017100579A4

AU2017100579A4 - Pulse induction metal detector using a convolution error feedback amplifier

Info

Publication number: AU2017100579A4
Application number: AU2017100579A
Authority: AU
Inventors: Paul Gerard Moody
Original assignee: Individual
Current assignee: Individual
Priority date: 2016-06-18
Filing date: 2017-05-21
Publication date: 2017-06-22
Anticipated expiration: 2025-05-21

Abstract

PATENTS ACT 1990 COMPLETE SPECIFICATION FOR AN INNOVATION PATENT Name of Applicant: Paul Gerard Moody Actual Inventor: Paul Gerard Moody Address for Service: 55 Hunter Street, Brunswick West, 3055, Victoria, Australia Invention Title: PULSE INDUCTION METAL DETECTOR USING A CONVOLUTION ERROR FEEDBACK AMPLIFIER The following statement is a full description of this invention, including the best method of performing it known to me. The whole of provisional patent application, titled : "PULSE INDUCTION METAL DETECTOR USING A CONVOLUTION ERROR FEEDBACK AMPLIFIER", Australian application number: 2016902385 is wholly incorporated by reference in respect of the present invention priority date, description, drawings and claims set out herein. PULSE INDUCTION METAL DETECTOR USING A CONVOLUTION ERROR FEEDBACK AMPLIFIER A pulse induction metal detector having at least three sequential states, each state 5 being a non zero period of time, a first state where a magnetic pulse, due to the flow of electric current controlled by a first switch, is emitted from a first coil, a second state where a back EMF being a high voltage is present across the first coil for a period of time due to abrupt cessation of electrical current flowing in the first coil, a third state being a receive state where residual current from the first and .0 second states and induced currents from conductive targets, mains frequency, earth magnetic field and ground in vicinity of the first coil result in a signal voltage across the first coil, includes the aforementioned coil for pulsed emission and detection of a time varying magnetic field, a means for damping a back EMF across the aforementioned coil, a first diode conducts during the aforementioned second .5 period protecting input electronics from the high voltage back EMF that may exist across the first coil, a first amplifier configured as an integrator performing an integration of the voltage across the coil during each receive period, a second switch connecting the output of the first amplifier during at least one receive period to the input of a second amplifier configured as an integrator, the second switch and .0 second amplifier thus demodulating and reconstructing a time varying voltage waveform, representative of the unwanted convolution error signal voltage, at the output of the second amplifier which is connected to a summation point at the input to the first amplifier forming a negative feedback loop, the loop gain having a negative sign, where the sample timing of the second switch and the time constant .5 of the second integration amplifier are adjusted by a suitable control means as known in the art such that unwanted convolution spectral components are demodulated and fed back into a summing point at the input of the first amplifier and thus the magnitude of unwanted signals in the output of the first amplifier is reduced. Page 2 of 15 PULSE INDUCTION METAL DETECTOR USING A CONVOLUTION ERROR FEEDBACK AMPLIFIER CD + Iwt C41± Page 13 of 15

Description

PULSE INDUCTION METAL DETECTOR USING A CONVOLUTION ERROR FEEDBACK AMPLIFIER 2017100579 21 May 2017

TECHNICAL FIELD

The present invention relates to pulse induction metal detectors, eddy current testers, and magnetic proximity detectors as known in the art.

5 BACKGROUND ART

The principle of pulse induction metal detectors is well known in the art.

With reference to Fig 1 as taught in US 5576624, a coil is damped by a resistor connected across the coil, the receive signal from the coil is then connected to the input of an amplifier. The output of the amplifier is connected to three switches SI, S2 and S3 which .0 perform the synchronous demodulation function. The output of SI, S2 and S3 are then connected to the input of an integrator. The output of the integrator is a waveform containing target information. A problem can occur due to the need to amplify the very small signals induced in the detector coil where the first amplifier in the receive electronics, being also presented with .5 large periodic signals due to the transmit and back EMF periods, causes the wanted signals to undergo convolution or modulation by the unwanted signals as a result of a non linear amplifier response. Unwanted signals include EMI, mains power, magnetic fields and other electrical or magnetic noise. Due to convolution of the wanted and unwanted signals target detection is made more complex. !0 A further problem occurs where the means of ground balance, as known in the art, is more difficult due to the convolution of wanted signals (targets ) with unwanted signals ( Earth field, mains frequency interference, EMI and motion of the coil itself. A further problem occurs where the residual ringing of the coil during a receive period after the back EMF period makes target detection more difficult. 25 A further problem occurs in the sampling method used to resolve a target signal from the combined effects of Earth magnetic field and ground effects due to ferrite content in the ground matrix. As taught in the prior art, a multiple sample, null summation technique is used to cancel Earth field and achieve ground balance. The balance condition can be hard to achieve in difficult ground conditions. 30 The inventor hereby discloses an improved pulse induction metal detector that eliminates or reduces the aforementioned problems .

Page 3 of 15

BRIEF SUMMARY OF THE INVENTION

According to a first aspect of the present invention, there is provided a pulse induction metal detector including the steps of : 5 At a first instant in time causing a magnetic pulse to be emitted from a first coil by initiating the flow of electric current from a power source through a first coil.

At a second instant in time, ceasing the flow of electrical current flowing from a power source causing a resulting back EMF or high voltage present across the first coil for a period of time, a suitable damping resistor in series with a first diode connected across the .0 first coil providing a means of damping and protecting the input amplifier, a first amplifier, from excessive voltage and current. A first diode ceasing conduction at a third instant in time after the back EMF has substantially decayed so that residual current from the transmit state and back EMF state and induced currents from targets, mains frequency, earth magnetic field and ground in .5 vicinity of the first coil result in a signal voltage across the first coil.

Applying the signal voltage across the first coil to the input of a first amplifier configured as an integrator and performing an integration of the voltage across the coil during each receive period. The voltage transitions caused by the coil transmit and back EMF events at the input to the amplifier serve to hold integrator in a reset state except during a receive !0 period.

Operating a second switch as a demodulator to connect the output of the first amplifier during at least one receive period to the input of a second amplifier configured as an integrator, the second switch and second amplifier thus demodulating and reconstructing a time varying convolution error voltage waveform representing unwanted signals of lower 25 frequency content than target decays ( for example Earth field, AC hum, etc )

Applying the reconstructed convolution error voltage to a summation node at the input to the first amplifier forming a negative feedback loop where the sample timing of the second switch and the time constant of the second integration amplifier are adjusted by a suitable means as known in the art such that unwanted frequency components are demodulated 30 and fed back into the first amplifier and thus the convolution of unwanted and wanted signals in the output of the first amplifier is reduced.

Connecting the output of the first amplifier to a signal processing system where and X and R samples, as known in the art, are integrated and differenced to achieve ground balance and target detection with greatly reduced error due to the reduction in convolution 35 artefacts.

Page 4 of 15

In one preferred embodiment the present invention provides a pulse induction metal detector with improved immunity from Earth Field, EMI and low frequency noise such as mains power interference.

In one further aspect of a preferred embodiment, the present invention provides a pulse 5 induction metal detector with improved means of ground balance by means of a null summation of an X sample, taken substantially during a flyback period, and an R sample taken substantially during a period immediately after the flyback period.

In one further aspect of a preferred embodiment, the present invention provides a pulse induction metal detector with improved minimisation of coil ringing during receive periods .0 by means of the low input impedance of an integrating amplifier connected via a damping means to the coil.

In one further aspect of a preferred embodiment, the present invention provides a pulse induction metal detector where all timing, switch control, signal processing, user interface and target indication is performed by a suitable control means including the use of one or .5 more microprocessors, programmable logic, discrete electronic components and transducers as known in the art.

According to another aspect of the present invention there is provided a circuit that will perform the first aspect and/or its aforementioned forms.

This disclosure is based on the principle that those skilled in the art will, with the benefit of :o this disclosure, be able to realise embodiments in addition to those described in the following descriptions. However in so far as they are utilising the disclosed structures, methods and systems for performing these derivative embodiments, the claims disclosed herein also apply to these derivative embodiments.

The realisation of any embodiment or derivative embodiment of the present invention may 25 be performed on the understanding that the method of performing the invention will rely on well understood principles of physical laws and electronic devices known to those skilled in the art.

The principles of the present invention are explained by the drawings and description of one or more embodiments contained within this disclosure. 30

Page 5 of 15

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 shows a prior art pulse induction metal detector signal chain.

Figure 2 shows one preferred embodiment of the present invention.

Figure 3 shows transmit and receive waveforms 5 Figure 4 shows waveforms with unwanted convolution and error corrected.

Figure 5 shows the voltage vs time waveform for no target, target and ferrite responses at the output of first integrating amplifier. Indicative sample timing for ground balance and target detection by a signal processing system is also indicated.

Page 6 of 15

PULSE INDUCTION METAL DETECTOR USING A CONVOLUTION ERROR FEEDBACK AMPLIFIER DETAILED DESCRIPTION OF THE INVENTION 2017100579 21 May 2017

One aspect of a preferred embodiment of the present invention is shown in Fig 2. A suitable means such as a microprocessor executing a stored program, as known in the art, controls the timing of the switches 202, 209 and signal processor 213. At a first instant in 5 time switch 209 is open circuit and switch 202 closes and coil 203 commences to store energy in a magnetic field as result of current supplied by power source 201. At a second instant in time switch 202 opens interrupting the current flow in coil 203. The sudden interruption of the current flowing through coil 203 results in a large reactive voltage or back EMF developing across coil 203. The back EMF may have a peak magnitude of .0 several hundred volts or more. Resistor 204 commences to conduct current as diode 205 is biased on. The resistance of resistor 204 is of a value suitable for damping coil 203. ( For example a 300 microhenry coil would require a damping resistance of approximately 500 ohms.) Diode 205 protects the input of amplifier 208 from excessive voltage and holds the integrator capacitor 207 in a reset condition. At a third instant in time the voltage across .5 coil 203 falls below the threshold voltage of diode 205 and diode 205 turns off. With diode 205 now turned off a fast integrating amplifier is formed from resistor 204, capacitor 207 and amplifier 208. The voltage across the coil is now integrated over the period of time until switch 202 closes again to initiate another transmit pulse. The feedback capacitor 207 begins to charge up due to the input voltage to amplifier 208. The capacitor charges up !0 at a rate determined by the RC time constant, (τ) of the series RC network formed by resistor 204 and capacitor 207. ( For example using a 500 ohm resistor a capacitance of ten to several hundred picofarads would be suitable.). At a fourth instant in time switch 209 closes and a sample of the output of amplifier 208 is connected to the input of the integrator formed by resistor 210 capacitor 211 and amplifier 212. At a fifth instant in time !5 switch 209 opens and capacitor 211 effectively stores the integrated value of successive samples from switch 209. The sample time of switch 209 is can be adjusted to commence at any instant and have any duty time (ie sample width ) with a rule that the sample begins and ends during a single period of time when switch 202 is open (ie coil 203 is not transmitting a magnetic pulse ).The values for resistor 210 and capacitor 211 are chosen to 30 be to such that the time constant of the series RC network formed by resistor 210 and capacitor 211 is greater than the series RC network formed by resistor 204 and capacitor 207. ( For example 2 to 3 orders of magnitude greater) The integrated samples from switch 209 are integrated by resistor 210 and amplifier 212 and capacitor 211 to form a demodulated signal which effectively is the convolution error present at the output of 35 amplifier 208. The convolution error signal at the output of amplifier 212 is connected to a summing node 206 where effectively the convolution error signal is subtracted from the input signal at amplifier 208 thus reducing convolution error in the output of amplifier 208 by means of a negative feedback loop described by the preceding description. The sample timing of switch 209 and the RC time constant of resistor 210 and capacitor 211 are 40 selected such that unwanted convolution error signals ( for example earth field and AC hum being low frequency) will be demodulated.

Page 7 of 15

With reference to Fig 3, voltage waveforms at various points of are show with respect to time. The voltage range of each waveform in Fig 3 are not to scale but ranged for the purposes of comparison. The coil transmit switch ( Fig 2, 201) is turned on for a period 301 and off for a period 302. The demodulation switch ( Fig 2, 209 ) is turned on (ie sampling ) 5 for a period 304 and remains off for a period 303. The ON period , 304 can be adjusted to sample for the whole of the flyback and receive period by a suitable control means as known in the art. In one further aspect of the present invention the demodulation switch ( Fig 2, 209 ) is ON when the transmit switch ( Fig 2, 201) is OFF and vice versa. The relative timing of the flyback voltage across the coil is shown by 305. The input waveform to the .0 first amplifier ( Fig 2, 208 ) is shown by 306 with a brief period of conduction by diode ( Fig 2, 205 ) that connects the resistor ( Fig 2, 204 ) across the coil ( Fig 2, 203 ) damping the flyback and protecting the input of the amplifier ( Fig 2, 208 ) from excessive voltage. The diode ( Fig 2, 205 ) conducts only during the flyback period and thus provides a mechanism for resetting integrator amplifier capacitor ( Fig 2, 207 ). The output of amplifier ( Fig 2, 208 .5 ) is shown by 307.

With reference to Fig 4, a comparison of the open loop (ie convolution error feedback disconnected ) waveform 401 is shown in contrast to closed loop (ie convolution error feedback active ) waveform 402 performance at the output of amplifier ( Fig 2, 208 ). In this case the coil was subjected to a 50 hertz magnetic field inducing 1 millivolt of unwanted !0 signal in the coil ( Fig 2, 203 ). In the open loop condition (ie no convolution error feedback ) the magnitude of unwanted modulation at the output of the first amplifier ( Fig 2, 208 ) is quite apparent as shown by 401. In the closed loop state the convolution error or modulation of the wanted signal by the unwanted signal is greatly reduced at the output of the first amplifier ( Fig 2, 208 ) as shown by 402 . The error voltage fed back to the summing !5 node ( Fig 2, 206 ) is shown by 403 and clearly shows the demodulated unwanted modulation waveform. Target responses are amplified by the integrating amplifier ( Fig 2, 208 ) and unaffected by the error feedback loop as the spectral components of the desired signals will lie outside the pass band of the convolution error feed back demodulator and integrator. The high frequency gain of the amplifier ( Fig 2, 208 ) is determined to some 30 degree by the reactive impedance of feedback capacitor 207 and the gain will vary with respect to the spectral components of the signal being amplified.

With reference to Fig 5, Three superimposed waveforms present at three different receive periods at the output of amplifier ( Fig 2, 208 ) are shown in relation to sample timing required to achieve a ground balance condition and target detection by the signal 35 processing system ( Fig 2, 213 ). Waveform 501 represents response to a no target condition. Waveform 502 represents the response to a target is near the coil ( Fig 2, 203 ). Waveform 503 represents the response to a ferrite target or ferrite ground near the coil ( Fig 2, 203 ). The ferrite / ground response 503 lags the target 502 and no target 501 response initially so the signal processing system ( Fig 2, 213 ) takes a first 'X' sample ( as 40 known in the art) as indicated by 504, substantially during the flyback period of time. A second sample 505 , an 'R' sample ( as known in the art), is then taken by the signal

Page 8 of 15

PULSE INDUCTION METAL DETECTOR USING A CONVOLUTION ERROR FEEDBACK AMPLIFIER 2017100579 21 May 2017 processing system ( Fig 2, 213 ) just after the flyback ceases. The duty period ( sample duration ) and the ON/OFF times for each of the samples 504 (X sample ) and 505 ( R sample) can be adjusted by a suitable means ( for example a microprocessor) as known in the art and a null summation, as known in the art, performed with respect to time such the 5 effect of ferrite / ground on target detection will be minimised. The sample timings can then be adjusted to maintain a ground balanced condition.

Page 9 of 15

Claims

CLAIMS What is claimed is :

1. A pulse induction metal detector having at least three sequential states, each state being a non zero period of time, a first state where a magnetic pulse, due to the flow of electric current controlled by a first switch, is emitted from a first coil, a second state where a back EMF being a high voltage is present across the first coil for a period of time due to abrupt cessation of electrical current flowing in the first coil, a third state being a receive state where residual current from the first and second states and induced currents from conductive targets, mains frequency, earth magnetic field and ground in vicinity of the first coil result in a signal voltage across the first coil, includes the aforementioned coil for pulsed emission and detection of a time varying magnetic field, a means for damping a back EMF across the aforementioned coil, a first diode conducts during the aforementioned second period protecting input electronics from the high voltage back EMF that may exist across the first coil, a first amplifier configured as an integrator performing an integration of the voltage across the coil during each receive period, a second switch connecting the output of the first amplifier during at least one receive period to the input of a second amplifier configured as an integrator, the second switch and second amplifier thus demodulating and reconstructing a time varying voltage waveform, representative of the unwanted convolution error signal voltage, at the output of the second amplifier which is connected to a summation point at the input to the first amplifier forming a negative feedback loop, the loop gain having a negative sign, where the sample timing of the second switch and the time constant of the second integration amplifier are adjusted by a suitable control means as known in the art such that unwanted convolution spectral components are demodulated and fed back into a summing point at the input of the first amplifier and thus the magnitude of unwanted signals in the output of the first amplifier is reduced.
2. A pulse induction metal detector as claimed in claim 1 using a first diode as a switch to protect the input of receive electronics where when the diode turns OFF, the low input impedance of an integrating amplifier maintains a low effective impedance across the coil in conjunction with the damping resistor also forming part of the RC network performing the integrator function associated with the aforementioned amplifier thus reducing ringing in the coil.
3. A pulse induction metal detector as claimed in claim 1 and 2 where the timing of electronic switches utilised in the present invention are adjustable by a suitable control means, as known in the art, in respect of OFF and ON timing whilst a preferred embodiment of the present invention is operating.
4. A pulse induction metal detector as claimed in claim 1, 2 and 3 where a ground balance may be achieved by taking two time sequential samples at the output of the first integrating amplifier, a first X sample during the flyback period and a second R sample during the period immediately after the flyback period, a receive period, integrating each aforementioned sample, and performing a null summation in conjunction with adjusting the timing of the aforementioned samples, resulting in a target signal, if present, substantially unaffected by ground.
5. A pulse induction metal detector as claimed in claim 1, 2, 3 and 4 where a means of adjusting the capacitance of a capacitor utilised in an integrator, a single integrator or plurality of integrators as disclosed in the present invention, may be performed during the operation of the present invention to achieve an optimal target detection at an instant in time.
6. A pulse induction metal detector and preferred embodiments as herein before described with reference to claims 1 to 5 inclusive.