BACKGROUND TO THE INVENTION

The present invention relates to a system for coupling electrical power from a source to a load.

Commonly, electrical equipment is connected to a power source by means of a plug and socket or by other arrangements involving direct physical contact between conductors, as in, for example, a domestic light fitting.

There are a number of situations in which the use of these conventional power coupling systems is inappropriate.

In medical electronics, the need to ensure the safety of patients connected to the equipment often makes it undesirable that there should be a direct connection from equipment to the mains. In hazardous areas such as mines, oil refineries and chemical plants, there are considerable risks associated with the possibility of sparks occurring as contact is made. Even in domestic settings such as the kitchen or bathroom, the presence of water makes the use of conventional plugs and sockets potentially hazardous.

SUMMARY OF THE INVENTION

According to a first aspect of the present invention, there is provided a power coupling system comprising a source unit arranged to be connected to a power supply and including a first winding and a load unit including a second winding, in use power being inductively coupled from the first winding to the second winding, and the source unit and load unit being readily separable, characterised in that the second winding is connected in a resonant circuit, and in that an AC voltage applied to the first winding is tuned to the resonant frequency of the load circuit including the secondary winding leakage inductance.

Preferably the system includes a variable tuning circuit arranged to control the frequency of the voltage

applied to the first winding in dependence upon the said resonant frequency of the load circuit.

The power supply system of the present invention replaces conventional direct conduction of current from the source to the load by inductive coupling between a primary winding connected to the source, and a secondary winding connected to the load. The primary and secondary windings are not formed around a common core as in a conventional transformer, but are located in physically separable units and power is inductively coupled across an air-gap. Since there is no requirement for a direct physical connection, the different units can be entirely encapsulated, making the system safe for use in hazardous environments such as those discussed above. it has previously been suggested, in a few specialised fields of application, to use inductive coupling of a power source to a load. For example, it has been proposed to provide a charger for an electric vehicle in which a primary winding in the charging unit is brought into the proximity of a secondary winding of a circuit connected to the vehicle batteries. The system operates at the mains frequency of 50 Hz. However, such systems suffer a number of serious limitations and so hitherto have not found wide acceptance. By comparison with conventional transformers in which the two windings are formed on a common core, the amount of flux leakage is greatly increased and so the efficiency with which power is coupled across the system is very poor. In practice, therefore, such systems have only been used where the power requirement of the load is small. The present invention overcomes these disadvantages to provide a power coupling system of sufficiently high efficiency to be of use in a wide range of applications. This achieved by providing in the source unit a control circuit arranged to drive the primary winding at a frequency substantially matched to the resonant frequency of the load circuit including the secondary winding leakage inductance. Operation in this manner provides power

coupling with high efficiency despite the presence of significant leakage inductances. At the same time, the relatively high frequencies involved makes it possible to reduce the size of the windings to the extent that a load unit can readily be accommodated within many otherwise conventional items of electrical equipment. Because it relies upon tuning to a resonant frequency within a certain range, a system embodying the present invention will not in general couple power, e.g., to a metal object which accidentally comes into contact with the source. The system therefore offers improved safety.

Preferably the control means include a detector responsive to the phase difference between voltage and current in the source circuit or between the source unit voltage and the load unit current as referred to the source unit and arranged to produce a control output dependent on that difference.

Preferably the source unit includes an inverter switched at a frequency determined by the control means. Using the system of the preferred aspects of the present invention, as the load unit is brought nearer the source unit the load impedance is reflected into the source circuit, reducing the phase difference between voltage and current in that circuit. A phase comparator detects this change and in response to it changes the operating frequency of the inverter so that it becomes closer to the load circuit resonant frequency. As a result, the load circuit driving impedance is reduced, as the frequency of the inverter approaches the resonant frequency of the load circuit.

Preferably the control circuit is arranged to drive the primary circuit at a high frequency in the absence of a load, and to reduce the operating frequency of the inverter as the load approaches. With no load present operating at high frequency, the primary reactance is high resulting in low primary current and low resultant magnetic field strength.

This control arrangement can adjust the operation of the inverter to suit a wide range of load impedances and air gaps. When the source unit has "acquired" the load unit it will operate with air gaps of 0 to about 10% of the total magnetic path length. Circuit efficiency is high when operating in this resonant mode, and with appropriate design of the magnetic core components, stray magnetic fields can be reduced to acceptable levels both when the load unit is present and absent. When no load is present the primary current is small.

Preferably the load winding is formed on a core having a re-entrant portion and the source winding is formed on a core having at least a portion which in use locates in the re-entrant portion of the load core as the load and source are brought together.

Possible applications for the power transfer system include emergency lighting in hazardous areas, plug-in lighting systems for such areas, rechargeable lanterns and torches for marine and industrial use and lanterns or emergency lamps for domestic use, and ultra low leakage power supplies for electronic equipment in, e.g., medical applications. In the case of lighting, luminaries incorporating a non-contact power supply system can be re- lamped without it being necessary for the entire power system to be shut down. The power transfer system is also of particular value in enabling isolation from the effects of large supply transients and common-mode effects which are common, for example, in electric railways.

The systems embodying the present invention are particularly useful in enabling the coupling of both power and data between units which move with respect to each other. Such a system may be used, for example, with a button or other equipment mounted on a sliding door. Another important field is in the use of transducers mounted in a rotating wheel or a rotating turbine. Systems embodying the present invention facilitate the transmission of electrical power from a fixed power supply to the load

unit which is mounted with the rotating transducer inside the wheel or turbine.

As will be described in further detail below, power coupling systems embodying the present invention can readily be adapted for the transmission of data using frequency-shift keying techniques without disturbing the transmission of power.

According to a second aspect of the present invention, there is provided a method of coupling power inductively from a source to a load characterised by bringing separable respective source and load units into proximity, the load unit including a winding connected in a resonant circuit, and driving the source at a frequency substantially matched to the resonance frequency of the load, including the load winding leakage inductance.

BRIEF DESCRIPTION OF THE DRAWINGS

A system in accordance with the present invention will now be described in further detail with reference to the accompanying drawings, in which:

Figure 1 is a simplified circuit diagram;

Figure 2 is a diagram showing a rechargeable lantern incorporating the circuit of Figure 1;

Figure 3 is a diagram showing a lumped-circuit model of the transformer;

Figure 4 shows the control characteristic in a first embodiment;

Figure 5i and 5ii shows a current monitoring circuit for a third embodiment; Figure 6 shows current waveforms for the circuit of Figure 5i and 5ii;

Figure 7 shows the control characteristics of a second embodiment; and

Figure 8a is a schematic a bridge inverter for use in a source unit and Figures 8bi,8bii and 8ci-8civ are detailed circuit diagrams;

Figures 9a to 9d show different alternative configurations for the cores and windings in the source and load units;

Figure 10 is a circuit diagram of a load unit adapted for transmitting a simple logic signal;

Figure 11 is a circuit diagram of a load unit adapted for a continuous data transmission;

Figure 12 is a diagram showing schematically a power coupling system for use underwater; and Figure 13 is a diagram showing schematically a source and load unit used to provide a power connection to a wheelchair battery charger.

DESCRIPTION OF EXAMPLES A power coupling system comprises a source unit 1 and a load unit 2, each of which is normally contained within a separate housing. The load unit may, for example, be incorporated in the base of a housing of a rechargeable electric lantern L. The source unit 1 in this case is located in the housing of a charging unit for the lantern.

The source unit comprises a primary winding 3 formed on an associated core 4. The winding is driven via a half- bridge inverter 5. Alternatively, for circuits handling higher power levels, a full bridge inverter may be used. A control circuit 6 including a current monitor CM and phase comparator 7 controls the frequency of switching of the inverter 5. The phase comparator employs exclusive-OR logic devices to provide an output which, after filtering, increases linearly from 0 as the phase difference between the input signals arises. Phase comparison might alternatively be carried out using other logic devices in an equivalent circuit, or integrated circuits such as phase-lock loops. An appropriate device of this latter sort is device 74HC4046A from the standard 74 Series Logic family. The control circuit may be constructed using conventional discrete logic and linear components, or

alternatively may be formed as an ASIC, or a thick film hybrid assembly.

The load unit 2 includes a secondary winding 8 on an associated core 9, together with a rectifier and any further smoothing and regulating circuitry required by the particular load. A series resonating capacitor CR is connected between the secondary winding and the output circuitry.

As will be described in further detail below, the reactance of the resonating capacitor, in conjunction with the leakage inductance and any load inductances associated with the secondary winding creates a resonant circuit. The control circuit 5 in the source unit is effective to control the frequency of the source so that it generally corresponds to the resonant frequency of the load circuit. The frequency of operation of the source will typically be within 10% of the resonance frequency and may match the resonance frequency to within 1%. Typical frequencies of operation are from around 100 kHz to 1 or 2 MHz, according to the power requirements of a particular load.

Figure 8a shows in further detail the half bridge inverter 5 of the source unit 1. Parasitic transistor elements are shown in broken lines. The transistor switches in the two arms of the bridge are both power MOSFETs. Suitable devices for a 200 W load are available as model no. IRF 840 manufactured by International Rectifier. Such devices offer high switching speeds but in conventional use suffer the disadvantage of significant power dissipation in the switches. As conventionally operated, Mosfet switches QI and Q2 are driven by anti-phase square wave signals so that the switches conduct alternately. Capacitors Cla and Clb are relatively large so that the voltage at point X remains sensibly constant at Vd/2. This implies that IL/(4fCl) << Vd

where IL is the load current as shown on the diagram, f is the switching frequency and Cl = Cla = Clb.

Suppose that switch QI is conducting and load current is flowing as shown in the diagram.

The parasitic capacitance CQ1 of switch QI will be uncharged (voltage across switch effectively = 0) while CQ2 will be charged to Vd.

Now let QI be turned off and Q2 be turned on. The voltage at point Y must fall from Vd to effectively zero.

CQ1 must charge to voltage Vd while CQ2 must discharge to zero. The charge and discharge currents must flow via Q2 and power dissipation can be high.

When, at the end of the next half cycle of operation, Q2 turns off and QI turns on, a similar action occurs but with energy being dissipated by QI.

In contrast with the mode of operation described above, if the load is highly inductive then, when QI is turned off the load current will continue to flow, the driving force being provided by the energy stored in the magnetic field associated with the inductive load. Current transfers from switch QI to capacitors CQ1 and CQ2, charging QI and discharging Q2. When the voltage at Y falls to just below zero, diode D2 conducts "catching" the voltage at Y and preventing further fall in voltage.

Transistor Q2 can now be turned on without power loss since the switching action will occur with diode D2 conducting and, therefore, almost zero voltage across the transistor.

The current through the load will then fall to zero at a rate set by the circuit parameters and then reverse in polarity, flowing via Q2.

In systems embodying the present invention, by using a load coupling transformer with low magnetising inductance the magnetising current is made high and operation in the mode just described can be ensured. The circuit therefore operates with near-zero switching losses in QI and Q2. The increased magnetising current leads to increased on-state

losses in the switches and in the transformer primary winding resistance but this is offset by the reduction in switching loses, especially at high switching frequencies as are used in NCET. Figure 2 shows one construction for the transformer cores, and the location of the cores in relation to the housing for the source unit and load unit. The use of a shape having a re-entrant portion for one of the cores and a complementary projecting shape for the other of the cores serves to reduce the stray flux as the load and source unit are brought together. This increases the efficiency of the system, and also serves to reduce radio-frequency interference.

Figures 9a to 9c show further alternative constructions for the transformer cores. Again the shape and configuration of the cores is chosen to maximise the efficiency with which the flux is coupled across the air gap. In Figure 9a, the ferrite core 91 is U-shaped with a winding, indicated by the hatching, formed around the base of the U. The second core 92 is also generally U-shaped but with smaller dimensions than the first core so that it fits within the re-entrant portion of the first core. The tips of the arms of the smaller U-shaped core 92 are cut away as shown in the Figure so that the surface of the arm is closer to the winding on the first core 91 in the region towards the edge of that winding. This serves to increase the uniformity of the field and reduce the flux leakage in that region towards the edge of the winding. Either of the cores may be used as the primary, that is to say for the source unit, but it is preferred to use the smaller core 92 for the source because the source magnetic assembly is then contained by the load assembly and stray field leakage is reduced.

Figure 9b shows another arrangement for the cores. In this example, the cores are generally cylindrical. The first larger core 93 again has a re-entrant portion which receives the second smaller core 94. The arms of the core

93 bounding the re-entrant portion are generally annular in plan. The outer edges of the smaller core 94 and the opposing portions of the larger core 93 are cut away as shown. In the core structure shown in Figure 9c, a dual E- core structure 95, 96 is used for one of the windings, say the primary, and the other winding 97 on ferrite core 98 is inserted into a slot formed between the two E-cores. This arrangement is particularly suitable for a power coupling systems to be used as a substitute for a conventional plug and socket for users having limited dexterity. The load unit incorporating the secondary winding can then be formed as a simple card or plate which is dropped into the slot provided in the source unit. If the transformer comprising the primary and secondary winding is considered as a lumped circuit element as shown in Figure 3, the inductance at the secondary output terminals as seen by the secondary-side circuitry is given by:

Lout = L12 + Lm2 m parallel wi th Lll * ^ljm ^

Lml (1)

The secondary winding open-circuit voltage is given by:

V2 (oc) = Lm2 Lml + Lll \ lotnl

(2)

The resonant frequency of the secondary circuit is:

/ RES =

2π- LoutCR

(3)

The air-gap between the two parts of the transformer controls the primary and secondary magnetising inductances Lml and Lm2 and the leakage inductances Lll and L12. As the air-gap is changed, the open-circuit secondary voltage,

the source impedance presented to the secondary circuit and the resonant frequency of the secondary circuit will all vary.

On the primary side, the low magnetising inductance of the primary winding requires a high magnetising current in phase-quadrature with the applied voltage. If a square- wave drive voltage is employed the current will not only be in quadrature but will be of triangular waveform.

The transformer primary voltage and current are monitored by the source unit control circuitry which includes a phase comparator to provide a signal which is a function of the phase difference between voltage and current. In a first embodiment, this signal is used to modify the operating frequency of the source unit inverter so that the relation is as shown in Figure 4.

The circuit operation is arranged so that if power is applied to the source unit with no load unit present, the inverter operates at a frequency lower than the resonant frequency defined in Equation 3, and with the primary current lagging the applied voltage by, sensibly, 90°. As the load unit is brought near to the source unit, a capacitive impedance is reflected into the primary circuit and the phase difference is reduced. This causes the operating frequency of the inverter to be increased so that it becomes closer to the secondary circuit resonant frequency. In consequence the load circuit driving impedance is reduced. The control circuit adjusts the inverter drive so that operation occurs close to the resonant frequency using an exclusive - OR type phase comparator with the inputs being the sense of the voltage applied to the primary winding (source unit) and the inverse of the sense (polarity) of the primary current. This produces a 50% duty-cycle output when the current lags the voltage by 90° increasing linearly with phase to 100% duty-cycle when the current is in-phase with the voltage. The output of the phase comparator is filtered to produce a dc level which is used as the input to a voltage

controlled oscillator (VCO) , the output of which controls the power MOSFET switches in the source unit.

In a second embodiment of the control system, operation takes place so that with no load present the frequency at which the primary is driven is higher than the resonant frequency of the load. As before, the magnetising current for square-wave excitation is of triangular wave shape and lags the driving voltage by 90°. The current amplitude required is less than in the first embodiment because of the increased operating frequency. This second embodiment uses an exclusive OR type phase comparator with the inputs being the sense of the voltage applied to the primary winding (source unit) and the sense (polarity) of the primary current. This produces a 50% duty-cycle when the current lags the voltage by 90° decreasing linearly with phase to 0% duty-cycle when the current is in-phase with the voltage. The output of the phase comparator is filtered to produce a dc level to which is used as the input to a voltage controlled oscillator (VCO) the output of which controls the power MOSFET switches in the source unit. The control characteristic for this second embodiment is as shown in Figure 7.

A third embodiment of the control system (Figures 5i and 5ii) includes, in the primary current monitoring circuit, two cascaded differentiator circuits and voltage amplifiers before the phase comparator. Consider with reference to Figure 6 the operation of the circuit with a light load. For the triangular magnetising current component, the first differential will be a square-wave leading the triangular current by 90°. The second differential consists of spikes of alternating polarity leading a further 90° on the current signal. The referred load current signal, being near-sinusoidal, will be unmodified in shape by the differentiators but will undergo phase inversion. The output signal from the differentiator-amplifier system, which will be as shown in Figure 6(H), is fed via a limiting amplifier, Figure 6(1),

to the phase comparator together with a signal developed from the drive voltage. The output of the comparator controls the operating frequency of the inverter, moving the frequency towards resonance. Shown in Figure 8cii is a low cost realisation of the double differentiator circuit which uses- CMOS 4069 unbuffered inverters in the linear mode. These devices are taken from the 4000 series of CMOS logic devices. The 4069 used in the linear mode has poor dc performance and requires the output to be ac coupled. The same technique may be used with high performance ac amplifiers/video amplifiers to produce a high frequency (>500 kHz) controller.

The input signal to the cascaded differentiators is taken from winding TIC, (Figure 8ci) which is a current transformer placed in series with the primary winding (see Figure 8b) with auxiliary load current compensation by means of current transformer T3. Resistors R19, R23, R24 and capacitor C25 act as a burden across which a voltage approximately proportional to the primary current is generated. This voltage is used to drive the differentiator circuits. Sketch waveforms are included in Figure 8ci.

The output of the second differentiator (IC2B) is amplified using a voltage comparator and Schmitt-triggered inverter to produce a digital waveform representing the polarity of the secondary (load unit) resonant current. This waveform IRES is applied to the clear direct (CD) input of a D-latch (IC5B) a 4013 device which is gated (Clocked) by the inverted output of the VCO (IC6) of a 4046 CMOS phase-locked-loop. The Q output of IC5B is a pulse train with a duty-cycle dependent upon the phase-lag between the clock (CLK) and Clear Direct (CD) inputs and hence on the phase-lag between the secondary circuit resonant current and the primary voltage. IC5B is used as an edge triggered phase comparator the output of which is zero if the signal applied to the CD input leads that

applied to the clock (CLK) input and rises linearly from zero output for in-phase signals to 25% duty-cycle when the signal applied to the CD input lags that applied to the CLK input by 90%. The D-latch IC5A is used to provide primary side current limiting by forcing the frequency of operation to increase if the sensed primary current exceeds a pre¬ determined threshold. If this occurs the output of IC3A goes low forcing the Q output of IC5A low and the output of IC7A high. The Q is set back to a logic 1 synchronously with the resonant circuit once the overcurrent condition has been removed.

The output of IC7A (the phase comparator output if not in an overcurrent condition) is buffered and filtered and compared with a pre-set reference. The difference between the filtered output of the phase comparator and the reference is inverted and amplified by IC30D and used to control the VCO in IC6. The output of IC6 is buffered and used to drive transformer T4 which gates the power MOSFETs Ml and M2 shown in Figure 8bi. This example of the third embodiment uses a closed feedback loop with the phase of the resonant circuit current (with respect to the primary voltage) being maintained close to a pre-set level preferably close to zero to provide a low impedance output. A second input to the VCO may be included using the OR-ing diode D17 shown in Figure 8ciii. This allows a second control-loop to be included which monitors the output voltage and modifies the frequency of oscillation to achieve output regulation; the phase-controlled loop ensuring that operation below resonance does not occur.

The capacitor C38 shown in Figure 8bii may be included to reduce the operating frequency range required to achieve output V regulation for a given load/supply voltage range. C38 changes the power circuit to a parallel loaded circuit at high operating frequencies, giving better control of the output voltage under light load.

Using this approach, it is possible to detect the phase of the referred load current even when it is small compared with the magnetising current. Very light loads are therefore acquired by the primary circuit so that efficient operation at near-resonant frequency occurs for a very wide range of load and coupling conditions.

The control arrangement can adjust the operation of the inverter to suit a wide range of load impedances and air-gaps. When the source unit has acquired the load unit it will operate with air-gaps of 0 to about 10% of the total magnetic path length. Circuit efficiency is high, producing power outputs ≥ lkw at efficiencies greater than 90%, with near-unity power factors, e.g. greater than 94% at 2.5kw output, when operating in this resonant mode and stray magnetic fields are reduced to acceptable levels by appropriate design of the magnetic core components. Power efficiency in this context is defined as the ratio of dc power output in the load to dc power input in the source. When no load is present the primary current is small. The power coupling circuit may also be used for signalling from the source unit to the load unit. Figure 10 shows a load unit which in this example is designed to provide a DC output. This circuit has been modified to allow information about momentary operation of a switch, that is simple transient logic information, to be transmitted from the load unit to the source unit.

In the secondary circuit, capacitor CR resonates with the secondary-side leakage inductance and wiring parasitic inductance. Capacitor Cl acts as an AC short-circuit and isolates the smoothing inductor LI from the resonant circuit.

Resistor Rl and transistor QI (QI is a MOSFET) are added to the output circuit to enable transmission of the signal from load unit to source unit. When QI is momentarily turned on, current flows via Rl and QI, increasing the loading as seen by the primary unit. This increase is arranged, by choice of Rl to be substantially

greater than the maximum increase resulting from normal load changes. A simple threshold detector in the source unit is used to sense changes in the current demand within that unit, and thereby to recover information about operation of switch QI in the load unit.

One possible use for such an arrangement is the transmission of information relating to the operation of a push-button switch mounted, for example, in a sliding door in a railway carriage. Because of the movement of the door conventional wired connections present considerable problems, and these can be overcome by using an inductive connection with the circuit as described above.

Figure 11 shows a further example of a load circuit modified to allow digital data transmission from the load to the source. Additional components are capacitor CR2, transistors QI and Q2, and drive transformer TR1. QI and Q2 are again Mosfets. All other components are as in the original circuit described above with reference to Figure 10. Transistors QI and Q2 form a bi-directional switch (Bright switch) which is controlled by the data signal via transformer TR1 which provides signal isolation. With the switch undriven and non-conducting, the secondary resonant circuit consists essentially of capacitor CR1 and the coupling transformer secondary leakage inductance. When the switch conducts, capacitor CR2 is connected in parallel with CR1 and the resonant frequency falls.

The primary (source) unit control circuitry contains a phase-locked loop (PLL) which monitors the phase difference between the primary winding voltage and the secondary winding current as referred to the primary. The PLL produces an output voltage proportional to phase difference which controls the frequency of a voltage- controlled oscillator (VCO) so that the primary unit inverter is driven at the resonant frequency of the secondary circuit.

As the secondary (load) side switch, formed by transistors QI and Q2, is operated, the resonant frequency of the secondary circuit changes and the PLL output voltage changes in order to move the drive frequency to the required value. The PLL output voltage is passed, via a filter serving to remove system noise, to a logic output stage and the signals controlling QI and Q2 in the secondary unit are thereby recovered in the primary unit. With such a circuit, the maximum data transmission rate is approximately 10% of the lower of the two resonant frequencies of the load circuit.

Load and source units embodying the present invention can be used to provide a universal power coupling system enabling a variety of source voltages and load voltages to be handled using standard components. All that is required for such a set is that the load and source units should operate within a defined maximum power level, in a certain frequency range, and at a certain flux density. The control electronics in each source unit can then be designed to sense the presence of a resonant load circuit tuned in the correct frequency range, and then to adjust the drive frequency of the primary unit inverter to the actual resonant frequency. The ratings of the power stage electronic components would be determined primarily from the power output required and the supply voltage to be used.

The physical size of the magnetic coupling components is chosen with reference to the frequency of operation and the maximum power to be transmitted. In any given source unit, the number of turns forming the primary winding is chosen as a function of the supply voltage, frequency of operation and magnetic circuit dimensions and parameters. A number of different primary units may be constructed to operate, for example, one from a 240V 50Hz supply and another from, say, a 50V DC supply. The units operate at the same frequency and the same magnetic flux density and have magnetic assemblies of the same dimensions. The

difference in supply is then "transparent" to the load units which receive the same driving conditions from either source unit.

For the secondary (load) units, the dimensions and number of secondary winding turns are functions of the power level, frequency, flux density, and of the structural dimensions of the source unit. The number of turns on the secondary winding is selected to produce a correct desired output voltage when coupled to the standard AC magnetic field of the source units. A range of secondary units can be designed, all of the same physical dimensions and requiring magnetic fields of the same frequency and strength, but providing different output voltages. When secondary-side power processing electronics is added the range of outputs can be very wide, e.g: 5V @ 50A dc; 100V @ 2.5A dc; 5000V @ 50mA dc; 240V, 50Hz ac.... Since for a given frequency of operation, flux density and physical size and form, any secondary "plug" will correctly couple to any primary "socket" a non-contact plug and socket range can be provided which operates from various input voltages and, if the plug is hard wired to the load, always provides the correct drive voltage for the load.

Such a system of universal connectors may be of particular value for use with underwater power tools. Equally any single load unit/source unit system embodying the present invention can offer significant advantages in this context. Conventional so-called wet-mating assemblies used hitherto require complex sealing systems which exclude the water from contact with the power conductors. With such conventional assemblies it is not safe to mate or separate the connectors with the power turned on. In practice often a key interlock system is used, with a diver having in his possession a key necessary to operate a surface power supply while he makes the physical connection between the plug and socket assemblies, and the diver then having to return to the surface to return the key before the power can be turned on.

All these practical problems can be avoided using the non-contact power transfer system of the present invention. With this system true wet-mating is possible: the load unit and source unit can be mated or separated under water and with the power applied. The physical arrangement for the mating can be a simple clip arrangement and there is no need to exclude water.

With a set of universal wet-mating connections, there is the further advantage that a range of different tools having different power supply requirements can be plugged in to a single source unit without the user having to ensure correct matching of a particular tool to a particular power source. Thus, in the example shown schematically in Figure 12, a first tool 121 operating at 12V and a second tool 122 operating at 240V both are wired to a respective secondary unit 123, 124 which generates the correct voltage when mated with the source unit 126. Power to the source unit 126 is supplied from a surface generator 127. A further example of possible fields of use for systems embodying the present invention, is in the provision of connectors for a power supply for charging the battery of an electrically operated wheelchair. Often a user of the wheelchair will have limited strength or dexterity. Accordingly manipulation of a conventional plug and socket for connecting a charger to the wheelchair battery can present considerable difficulties. These difficulties can be overcome by replacing the conventional plug and socket with a non-contact load unit/source unit constructed in accordance with the present invention. The load unit which is a charging circuit 131 for the battery 132 on the wheelchair 133 may include a slotted double-E coil as discussed with respect to Figure 9c above. This provides the "socket" the source unit 134 is the "plug" used to make the connection. This "plug" containing the source coil is card-shaped and is connected to the mains at its remote end. The user then simply has to drop the card

into the socket 135 in the load unit to establish a connection to the mains. Alternatively other non-contact assemblies may be used in different configurations.

The tables below show components for Figure 5 and Figure 8 respectively.

TABLE 1

RESISTORS

R25 100Ω 2% 0.5W

R26 15KΩ 1% 0.125W

R17 22KΩ 1% 0.125W

R28 22KΩ 1% 0.125W

R37 10KΩ 2% 0.125W

R38 1KΩ 2% 0.125W

R39 1.5MΩ 2% 0.125W

R40 820KΩ 2% 0.125W

R41 1KΩ 2% 0.125W

CAPACITORS

C36 2n2 10% CERAMIC

C37 150pF 2% CERAMIC

C38 8.2pF 0.25P CERAMIC

C39 47pF 2% CERAMIC

C40 8.2pF 0.25P CERAMIC

C45 lOpF 2% CERAMIC

DIODES

D14, D15, D18, D19 1N 148

D13, D16 BZX79C6V2

INTEGRATED CIRCUITS

IC2 LM6364 IC4 LM311N

IC3 LM6364 IC5 40106

TABLE 2

Ref Valu Tolerance Limi ts

R3 10K 2% 0 .125W

R4 100K 5% 1W

R5 100K 5% 1W

R6 47 5% 0 .25W

R7 47 5% 0 .25W

R18 10K 5% 0 .125W

R19 33 2% 0 .125W

R20 10K 2% 0 .125W

R21 1K5 2% 0 .125W

R22 820K 5% 0 .125W

R23 33 2% 0 .125W

R24 33 2% 0 .125W

R25 47K 2% 0 .125W

R26 47K 2% 0 .125W

R27 10K 1% 0 .125W

R28 10K 1% 0, .125W

R29 10K 1% 0. .125W

R30 10K 5% 0. .125W

R31 100K 5% 0. .125W

R32 IK 5% 0. ,25W

R33 10K 5% 0. 125W

R34 10K POT

R35 10K 2% 0. 125W

R36 10K 2% 0. 125W

R37 150K 2% 0. 125W

R54 10K 2% 0. 125W

R55 2K7 2% 0. 125W

R56 IK 5% 0. 5W

R57 10K 5% 0. 125W

R58 15K 2% 0. 125W

R59 100K 2% 0. 125W

R60 10K 5% 0. 125W

R61 18K 2% 0. 125W

R62 10K 2% 0. 125W

Ref Value Tolerance Limi ts Comments

R63 10k 2% 0.125W

R64 10K 2% 0.125W

R65 15K 2% 0.125W

R66 IK 5% 0.25W

R67 IK 5% 0.125W

R68 330 5% 0.5W

R69 100 5% 0.5W

R70 10K 5% 0.125W

R71 10K 5% 0.125W

R72 10K 5% 0.125W

C8 470xF 200V >1A ripple current

C9 470/xF 200V >1A ripple current

Cll 1/ιF 630V Siemens MkP

C12 0.22/ιF 400V Siemens MkT

C13 0.22/zF 400V Siemens MkT

C14 15nF 2kV Philips 378 Series

C15 15nF 2kV Philips 378 Series

C16 15nF 2kV Philips 378 Series

C17 15nF 2kV Philips 378 Series

C18 2/ι2F 250V Philips 378 Series

C19 2μ2F 250V Philips 378 Series

C20 2/X2F 250V Philips 378 Series

C23 22μF 35V

C24 lOμF 16V Tantalum

C25 InF 5% Ceramic

C26 33pF 2% Philips 680 Series

C27 2p2F 0.25pF Philips 680 Series

C28 33pF 2% Philips 680 Series

C29 2p2F 0.25pF Philips 680 Series

C30 lOOnF

C31 InF

C36 lOOnF 10% Ceramic

C37 330pF 5% Ceramic

C38 47nF lkV Philips 378 Series

C39 lOOnF Ceramic

Ref Value Tolerance Limi ts Comments

C40 lOOnF 10% Ceramic

C41 lOOnF Ceramic

C42 33pF Ceramic

D2 1N4148

D3 1N4148

D4 45V 80SQ045

Fitted for low volt-dp and 175°C junction temperature

D5 80SQ045

D6 80SQ845

D7 80SQ845

Dll UF4001

D12 UF4001

D13 UF4001

D14 UF4001

D15 1N4148

D17 1N4148

D18 1N4148

Ml 500V IRF840

M2 500V IRF840

M3 20V ZVP2106A

M4 20V ZVN2106A

M5 20V ZVN2106A

M6 20V ZVN2106A

Q7 BC556

Ref Comments

IC1 78M150.5A+15V regulator

IC2 4000 Series 4069 inverter

IC3 LM324 op-amp

IC4 4000 Series 40106 inverter

IC5 4000 Series 4013 D-latch

IC6 4000 Series 4046 phase-lock loop

IC7 4000 Series 4093 NAND

IC8 LM311 comparator

L4 50μH differential-mode choke

Tl 1:1:30 current transformer wound on Philips 3C85

U10 core using 0.315mm enamelled copper wire T2 Power transformer

T3 3:20 current transformer wound as Tl T4 48:48:48 gate drive transformer wound on Philips

3C85 U10 core using 0.224mm enamelled copper wire