GB2461994A

GB2461994A - An AC MOS switch arrangement with both zero-voltage turn-on and zero-current turn-off

Info

Publication number: GB2461994A
Application number: GB0912521A
Authority: GB
Inventors: Frank Ainscow
Original assignee: INTERFACING Ltd
Current assignee: INTERFACING Ltd
Priority date: 2008-07-18
Filing date: 2009-07-20
Publication date: 2010-01-27
Also published as: GB0912521D0

Abstract

An AC switch for controlling the application of power from a source 22 to a load 24 comprises two separately-controlled MOSFETs TA and TB in series. Comparators 32V and 32C provide indications of the polarity of the source voltage and the load current, so that the controller 26 can switch each MOSFET on during the half-cycle when the voltage across it is substantially zero, and so that the controller can switch each MOSFET off during the half-cycle when current through the device is either zero or is flowing through its anti-parallel diode. The circuit can thus switch inductive or capacitive loads without knowledge of the load, and power-consuming protective circuits such as snubbers or current-limiting resistors can be omitted. The transistors may be turned on briefly to determine whether the load is inductive or capacitive, so that an appropriate phase angle control regime may be used (figures 5 and 6). The power consumption of the load may be monitored and transmitted to the central controller of an energy management system (figures 7 and 8).

Description

SWITCH

[001] The present invention relates to a switching device and, in particular, to a switching device for use in an energy management device. The present invention also relates to an energy management device incorporating such a switching device, and an energy management system comprising one or more such energy management devices.

[002] With rising costs in fuel, and due to environmental concerns, energy management systems are increasingly important as a way to reduce energy consumption in domestic buildings, and to reduce costs in commercial buildings where the charging tariff is related to peak demand.

[003] Known systems for managing energy usage in domestic or commercial buildings seek to limit the cumulative or peak energy consumption of the building by monitoring energy consumption and initiating automatic or manual load-shedding when usage reaches prescribed levels.

[004] Such systems typically operate by measuring the total power consumption and shedding predefined categories of load, such as lighting or air conditioning, or sometimes by withholding power from specified areas of the building.

[005] Thus, the actual power consumed by each type of load and in each area of the building is not monitored.

Accordingly, the loads/areas to be denied power can only be selected on the basis of settings defined by an operator in advance, in accordance with predicted power consumption patterns, and with no reference to where and how power is actually consumed at the time.

[006] Such systems do not, therefore, have the ability to react to the actual power consumed by different categories of load or in specific areas of the building, which means that the system has only crude control over the way power is consumed, and is unlikely to react in the most effective manner to unusual or unpredicted situations.

[007] The effectiveness of such systems could be improved by providing the ability to monitor and control the supply of power at the level of individual appliances. However, in order to control the power supply to appliances, a separate switching device for each individual appliance would be required.

[008] Various types of semiconductor switch are available for the purpose of switching an electrical appliance on and off. However, limitations associated with the power consumption of such switches and/or restrictions on the type of load with which such switches can be used has previously made the provision of individual switching devices for controlling a plurality of electrical appliances within an energy management system unworkable.

[009] Semiconductor switching devices suitable for switching an electrical appliance include SCR5, triacs, IGBT5 and MOSFET5. For currents up to about lOAmps (2.4kw at 240V), MOSFET5 (or Metal Oxide Semiconductor Field Effect Transistors) are preferred because channel resistance can be made very low, drive requirements are low and switching is very fast.

[0010] A circuit 10 for a typical MOSFET based device for switching AC current is illustrated in figure 1. The circuit comprises a pair of MOSFET5 Ti, 12 connected in series between an AC supply 12 and a load 14. The construction of a MOSFET results in a parasitic diode in parallel with the channel, with similar current capability to the transistor itself. This diode shares the reverse current with the transistor when reverse voltage is applied and so prevents the device blocking in the reverse direction. Thus, in the circuit of figure 1, each transistor Ti, 12 has, in effect, a corresponding parasitic diode Di, D2 as shown. If the parasitic diodes are not present or are inadequate, discrete diodes may be incorporated. Accordingly, when the gate voltages of the transistors Ti, 12 are driven high, for "on", current will flow through Ti and D2 in parallel with 12 during positive half-cycles of the AC supply and through 12 and Di in parallel with Ti during negative half-cycles. When the gate voltages of the transistors Ti, 12 are driven low, for "off", the connection between the two transistors is held by the diodes at a voltage which is approximately iV more positive than the more negative terminal of the supply. During positive half cycles, therefore, Ti supports the full supply voltage but 12 experiences a small negative voltage equal to the forward voltage drop of the diode D2. During negative half cycles 12 supports the full supply voltage but Ti experiences a small negative voltage equal to the forward voltage drop of the Di. The forward voltage drop of the diodes is typically iv, compared to the peak voltage of the supply which is typically about 350V, so it is regarded as substantially zero. Switching is achieved by means of a control circuit (not shown) connected at control input 16, which drives the gate voltages for Ti and 12 together.

[0011] Simple switching devices of this type are suitable for remote switching of resistive loads such as light bulbs. However, a problem occurs in cases where the load is reactive i.e. inductive or capacitive. In this respect, any appliance which comprises a transformer, such as low voltage lighting, or which comprises a motor, such as automated blind or window control mechanisms, will constitute an inductive load. On the other hand, appliances such as fluorescent lights and computers typically represent capacitive loads.

[0012] In the case of an inductive load, the voltage V is given by equation (1) V = L dI/dt (1) where L is the inductance in the load and I is the current.

[0013] On switch-off, the current I falls rapidly to zero from whatever value it had at the moment of switching.

Given that the switching time of the device is very short, the voltage so generated is limited only by breakdown of the transistor, leading to early failure of the device.

[0014] In the case of a capacitive load, the current flowing into the capacitance at switch-on is given by equation (2) I = C dV/dt (2) where C is the capacitance in the load.

[0015] At switch-on, the voltage V rises from zero to whatever is the instantaneous mains voltage at the time.

Unless that happens to be zero, the current I is limited only by the series resistance of the circuit, which is necessarily very small, leading to overcurrent and consequent failure of the device.

[0016] A number of ways for protecting against these effects are known. However, all have associated disadvantages.

[0017] For example, a snubber circuit may be connected to the switching device. Snubber circuits are passive circuits which contain resistance and capacitance (RC type), and sometimes diodes (RCD type), that limit the voltage generated at switch-off and absorb the energy of the pulse, and thus protect the switching device in the case of an inductive load. However, snubber circuits dissipate large amounts of energy, making their use undesirable, particularly in the context of an energy management system, where the primary aim is to save energy. RC snubber circuits are particularly inefficient, as they dissipate power continuously, even when the switch is off. RCD snubber circuits require two additional semiconductor diodes of similar rating to the transistors, thereby substantially increasing the cost of the device.

[0018] To protect a switching device in the case of a capacitive load, series resistance or a thermistor may be incorporated into the circuit to limit the current at switch-on. This also results in high levels of spurious power dissipation.

[0019] The inefficiency of these solutions is a particular problem in cases where it is necessary to protect the switching device against the effects of both types of reactive load, since both types of protection will be required, thereby potentially doubling the energy dissipation of the device, and increasing its cost and size.

[0020] In some cases a switch may be provided to enable a user to select the correct protective circuit for the type of load with which the switching device is to be used. However this solution is not ideal, because the switching device will be destroyed if the user inadvertently selects the wrong circuit for the type of load.

[0021] Another strategy is to employ zero-point switching. In this case, the device is constrained to switch on or off only at the zero-point of either the voltage or the current waveform, thereby protecting against one or other of the effects described above, but not both. A switch may be provided to select zero-point switching of either the voltage or the current waveform.

Again, however, this solution is not ideal, because the switching device will be destroyed if the user inadvertently selects the wrong form of protection for the type of load.

[0022] Zero-point switching represents a more efficient method for protecting a switching device than snubber circuits or series resistance. However, this method can only work effectively if the voltage and current waveforms are free of noise. In any practical application, noise will be present. As a result, the timing of the zero-point can vary significantly from cycle to cycle, and there may be multiple zero-crossings in each half-cycle. For this reason, switching devices which employ zero-point switching always do so in conjunction with a snubber circuit, and thus suffer the associated disadvantages of high power dissipation, high component costs and increased size and weight.

[0023] It will thus be appreciated that previously known arrangements for protecting MOSFET based switching devices seriously increase the power dissipation of the device, significantly increase the cost, size and weight of the device, and/or impose restrictions on the use of the device. For these reasons, the provision of a large number of individual switching devices to enable the implementation of an energy management system which monitors and controls the power supplied to individual appliances, has previously been impractical.

[0024] It is an object of the present invention to overcome the problems associated with the prior art.

[0025] According to a first aspect of the present invention there is provided a switching device for controlling the supply of AC current from a power supply to a load, the device comprising:-a pair of switchable elements, each having a channel for carrying current in forward and reverse directions, and being switchable between an on state for allowing current to flow through the respective channel and an off state for blocking the flow of current therethrough, wherein the switchable elements are connected such that, in said on state, current is carried in the forward direction of the respective switchable element during respective alternate half cycles; and control means for switching the pair of switchable elements between said on and off states to control the flow of current from the supply to the load, wherein the control means is configured to switch one of the switchable elements at a different instant to the other switchable element.

[0026] Accordingly, each of the first and second switchabiLe elements can be switched on independently of the other, during the half cycle when the forward voltage across the respective switchable element is substantially zero. This prevents failure of the device in the case of a capacitive load. Similarly, each of the switchable elements can be switched off independently of the other, during the half cycle when the current through the respective channel is either zero or flows in the reverse direction for the respective switchable element, and therefore flows through, or is shared with, a parasitic diode in parallel with the channel. On switch off, all of the load current is carried by the diode with minimal change in the voltage. This prevents failure of the device in the case of an inductive load. Moreover, each switchable element can be switched at a time well away from the zero-crossing of the respective voltage or current waveform, thereby avoiding associated noise issues.

[0027] The control means is preferably configured to switch the switchable elements to said on state, when the voltage across the respective switchable element is substantially zero. For a 50Hz supply, the switching time is preferably between approximately 3 and 7 ms, more preferably approximately 5ms, after the preceding zero-crossing of the supply voltage waveform. For a 60Hz supply, the corresponding time is between 2.5 and 6 ms, preferably approximately 4.2ms. Thus a time of approximately 4.8 ms satisfies both situations.

[0028] The control means is preferably configured to switch the switchable elements to said off state when the current through the channel of the respective switchable element is zero or flows through the channel in the reverse direction of the switchable element. Preferably, between approximately 3 and 7 ms, more preferably approximately 5ms, after the preceding zero-crossing of the load current waveform.

[0029] The control means is preferably configured to switch the state of the switchable elements such that one or both of rules I) and II) are satisfied, wherein rule I) requires each switchable element to be switched on at an instant when the voltage across the respective switchable element is substantially zero, and rule II) requires each switchable element to be switched off at an instant when the current through the channel of the respective switchable element is zero or flows in the reverse direction of the switchable element.

[0030] By satisfying rule I), the device is protected in the case of a capacitive load. By satisfying rule II), the device is protected in the case of an inductive load.

[0031] The control means may be configured to satisfy both rules I) and II) [0032] Alternatively, the control means may be configured to satisfy rule I) in the case of a capacitive load, and rule II) in the case of an inductive load.

[0033] In this case, the switching device preferably comprises load analysing means, for determining whether the load is inductive or capacitive.

[0034] The switchable elements are preferably MOSFET5, and are preferably connected in series.

[0035] According to a second aspect of the present invention, there is provided an energy management device for measuring and controlling the supply of power to a -10 -load, the device comprising energy monitoring means for determining the power consumed by the load, and a switching device, as defined in the claims, for controlling the supply of power to the load.

[0036] According to a third aspect of the present invention, there is provided an energy management device for measuring and controlling the supply of power to a load, the device comprising:-monitoring means for monitoring the voltage and current waveforms for the load, and determining a measure of the power consumed by the load by aggregating instantaneous values of the voltage x current product; switching means for controlling the supply of power to the load; load analysing means for determining the nature of the load from the relative phase of said voltage and current waveforms; and control means for switching said switching means in accordance with the determined nature of the load.

[0037] That is to say, the energy management device uses the relative phase of the voltage and current waveforms to determine the impedance of the load, ie, to measure any inductive or capacitive component.

[0038] Existing power control devices do not measure the load impedance, because the extra complexity results in higher power consumption and higher component costs, which are not thought to be justified. However, with the present invention, the phase information is additionally used to provide control means with information about the nature of the load, so that it can operate the switching device in accordance with the nature of the load. The same switching device can thus be used with resistive, inductive and capacitive load types, without the need for -11 -snubber circuits and the like to protect the switching device from the effects of high transient currents and voltages on switch-on and switch-off. The resultant saving in terms of component cost, product volume and power consumption more than compensates for the extra complexity of the circuitry required for measuring the load impedance. The switching means preferably comprises a pair of switchable elements, each having a channel for carrying current in forward and reverse directions, and being switchable between an on state for allowing current to flow through the respective channel and an off state for blocking the flow of current therethrough, wherein the switchable elements are connected such that, in said on state, current is carried in the forward direction of the respective switchable element during respective alternate half cycles; wherein the control means is configured for switching the pair of switchable elements between said on and off states to control the flow of current from the supply to the load, wherein the control means is configured to switch one of the switchable elements at a different instant to the other switchable element.

[0039] The control means is preferably configured to switch the switchable elements to said on state, when the voltage across the respective switchable element is substantially zero. Preferably, between approximately 3 and 7 ms, more preferably approximately 5ms, after the preceding zero-crossing of the supply voltage waveform.

[0040] The control means is preferably configured to switch the switchable elements to said off state when the current through the channel of the respective switchable element is zero or flows in the reverse direction of the switchable element. Preferably, between approximately 3 -12 -and 7 ms, more preferably approximately 5ms, after the preceding zero-crossing of the load current waveform.

[0041] The control means is preferably configured to switch the state of the switchable elements such that one or both of rules I) and II) are satisfied, wherein rule I) requires each switchable elements to be switched on at an instant when the voltage across the respective switchable element is substantially zero, and rule II) requires each switchable element to be switched off at an instant when the current through the channel of the respective switchable element is zero or flows in the reverse direction of the switchable element.

[0042] The control means is preferably configured to satisfy rule I) when the load is determined to be a capacitive load, and rule II) when the load is determined to be an inductive load.

[0043] The switchable elements are preferably MOSFET5.

[0044] In accordance with a fourth aspect of the present invention, there is provided an energy management system, comprising a central controller and one or more energy management devices, as defined in the claims.

[0045] The present invention will now be described with reference to the accompanying drawings in which: Figure 1 shows a circuit for a known MOSFET based device for switching AC current; Figure 2 illustrates a circuit 20 for a switching device in accordance with a first embodiment of the present invention; Figure 3a compares the supply voltage with the voltage across the transistors A and T of figure 2, when both A and TE are off; -13 -Figure 3b compares the load current with the current that flows through the channels of transistors A and B of figure 2 when both TA and TB are on; Figure 4 shows an example of a control circuit, constructed from discrete logic blocks, for controlling the operation of the transistors A, B, in accordance with the first embodiment of the invention; Figure 5 illustrates a circuit for a switching device in accordance with the second embodiment of the present invention; Figure 6 shows waveforms which illustrate the operation of the circuit of figure 5; Figure 7 shows an energy management system which comprises a plurality of energy management devices which incorporate the switching device of the present invention; and Figure 8 shows one of the energy management devices of figure 7 in more detail.

[0046] As explained above, in the absence of protective circuits, the conventional MOSFET switching device of figure 1 will fail at switch off when connected to an inductive load, due to high transient voltages, and will fail at switch-on when connected to a capacitive load, due to high transient currents.

[0047] The present inventor has established that both of these modes of failure can be eliminated if the switching device is operated in accordance with the following two rules: -I) switch on only when the forward voltage across the transistor is substantially zero; and II) switch off only when the forward current through the transistor is zero or flows in the reverse direction of the switchable element.

-14 - [0048] With reference to the switching device of figure 1, when the gate voltages of the transistors 11, 12 are driven high, to turn the transistors on, current will flow through the first transistor Ti and the parasitic diode D2 associated with the second transistor T2 in parallel with transistor T2 itself during positive half cycles of the supply frequency, and through the second transistor T2 and the parasitic diode Dl associated with the first transistor Ti in parallel with transistor Ti itself during negative half-cycles. Thus, in any half-cycle, load current flows in the forward direction in one transistor, and reverse current passes through the diode associated with the other transistor in parallel with transistor itself. Similarly, when the gate voltages of the transistors Ti, 12 are driven low, to turn the transistors off, the supply voltage will appear across only one transistor while the diode of the other device limits the reverse voltage to near zero.

[0049] However, because the two transistors are controlled simultaneously by the same gate driver, the application of rules I) and II) with the conventional switching device of figure 1 would require switching to take place precisely at the zero-crossing instant of the voltage/current waveforms. In practice, however, the timing of the zero-crossing point can vary significantly from cycle to cycle due to the presence of noise, and there may be multiple zero-crossings in each half-cycle.

Accordingly, the application of rules I) and II) to the conventional switching device of figure 1 is not possible in any practical situation, where noise will inevitably be present.

[0050] Figure 2 illustrates a circuit 20 for a switching device in accordance with a first embodiment of the present invention. The circuit 20 comprises a pair of -15 -MOSFETs TA and I connected in series between an AC supply 22 and a load 24. Switching of the MOSFETs is controlled by a control circuit 26.

[0051] Unlike the circuit of figure 1, the gate voltages of the transistors A, B are controlled by separate gate drivers 28A and 28B, such that the transistors can be turned on and off independently of one another.

[0052] Each gate driver 28A, 28B is connected to the gate terminal of the respective transistor A, B, and drives the gate voltage at the respective terminal high to turn the transistor on, and low to turn the transistor off, in response to logic level signals GateA and GateB supplied as inputs to the respective gate drivers.

[0053] The levels of signals GateA and GateB thus correspond to the on/off state of the respective transistors This information is fed back within the control circuit by supplying signals GateA and GateB as inputs to combinational logic blocks 30A and 30B, respectively.

[0054] GateA = 1 corresponds to A = on, and GateA = 0 corresponds to A = off. Similarly, GateB = 1 corresponds to = on, and GateB = 0 corresponds to = off.

[0055] A voltage polarity sensing amplifier 32V is connected across the supply terminals 22 for determining the polarity of the supply voltage V. A current sensing resistor 34 is connected between the two transistors 1, and a current polarity sensing amplifier 32C is connected in parallel with the resistor, for determining the polarity of the load current.

-16 - [0056] Each of the voltage and current polarity sensing amplifiers 32V, 32C has a respective output, which carries a logic level signal Po1V, Po1C which indicate the polarity of the supply voltage and the load current respectively. In this respect, when the supply voltage is positive, Po1V = 1 and when the supply voltage is negative, Po1V = 0. Similarly, when the load current is positive, Po1C = 1 and when the load current is negative, Po1C = 0. Po1V and Po1C are provided as inputs to both combinational logic blocks 30A and 30B.

[0057] Thus, Po1V, Po1C and GateA are provided as inputs to the first combinational logic block 30A, whose output A is represented by the Boolean expression: A = --GateA Po1V + GateA -PolC (3) [0058] Similarly, Po1V, Po1C and GateB are provided as inputs to a second combinational logic block 30B, whose output B is represented by the Boolean expression: B = -GateB --PolV + GateB Po1C (4) where indicates logic inversion.

[0059] Signals A and B are provided as respective inputs to delay blocks 36A and 36B, which delay signals A and B by approximately Sms. The delayed signals are provided as inputs to input terminals CK of respective D-type latches 38A and 38B.

[0060] A logic level ON/OFF command signal which represents a command from an external unit (not shown) to switch power to the load on or off, is provided as an input to both D-type latches 38A and 38B. A command to switch power to the load on is represented by ON/OFF = 1.

-17 -A command to switch power to the load off is represented by ON/OFF = 0. The ON/OFF command is clocked on the delayed rising transition of A and B by the respective latch, to generate signals GateA and GateB, at respective output terminals Q. [0061] In operation, each gate driver 28A, 28B, drives the gate voltage of the respective transistor TA, TB high to switch the respective transistor on, and low to switch the respective transistor off, in response to the level of the signals GateA, GateB output from the pair of D-type latches 38A, 38B.

[0062] Figure 3a compares the supply voltage V with the voltage VA, VB that appears across the transistors TA and TB when both TA and TB are off (GateA = 0, GateB = 0) Due to the above described construction of a MOSFET, the voltage across TA will be substantially zero during the first half-cycle of the supply, where the polarity of the supply voltage is positive (P01V = 1), whilst the voltage across TB will be substantially zero during the second half cycle of the supply, where the polarity of the supply voltage is negative (P01V = 0) [0063] Under rule I) the first transistor TA can be safely switched on at any time where Po1V = 1 and the second transistor TB can be safely switched on during the following half cycle, where Po1V = 0.

[0064] Figure 3b compares the load current I with the current A, B that flows through the channels of the respective transistors TA and TB when both TA and TB are on (GateA = 1, GateB = 1) . As can be seen from the figure, the current flowing through the channel of TA will be substantially reduced because some of the current flows through the parasitic diode Dl during the first half- -18 -cycle where the polarity of the load current is positive (P0ILC = 1), whilst the current flowing through T will be substantially reduced because some of the current flows through the parasitic diode D2 during the second half cycle where the polarity of the load current is negative (P01C = 0).

[0065] Under rule II) the first transistor A can be safely switched off at any time where Po1C = 1 and the second transistor TE can be safely switched off during the following half cycle, where Po1C = 0.

[0066] The time interval during which each of the transistors TA, TE can be switched on/off thus corresponds to a complete half-cycle of the supply frequency. This is typically lOms for a 50Hz supply and 8.3 ms for a 60Hz supply, which is ample to ensure that switching takes place at a time well away from the zero-crossings and thus avoids any associated noise issues. Thus, with the arrangement shown in figure 2, it is straightforward to implement both rule I) and rule II) to automatically protect the switching device from failure, irrespective of whether it is connected to an inductive or a capacitive load.

[0067] Rules I) and II) are applied by combinational logic blocks 30A and 30B, in accordance with boolean expressions (3) and (4) . In this respect, when TA and TB are off, A = 1 and B = 0 when the supply voltage is positive, whilst A = 0 and B = 1 when the supply voltage is negative (from the first terms of expressions (3) and (4)).

[0068] When TA and TB are on, A = 0 and B = 1 when the load current is positive, whilst A = 1 and B = 0 when the -19 -load current is negative (from the second terms of expressions (3) and (4)).

[0069] Delay circuits 36A, 36B, delay signals A and B such that the transition between 0 and 1 on the signal output from the delay blocks is shifted away from the transitions between 0 and 1 on Po1V and PolO, which correspond respectively to the zero crossing points of the voltage and current waveforms. The delay is set to -5ms, which is about half the duration of a half cycle of the supply. This ensures that the transition between 0 and 1 for the delayed signals A and B takes place well away from the zero-crossing points and any associated noise issues.

[0070] At the D-type latches 38A, 38B, the ON/OFF command signal is clocked by each latch on the delayed rising transition on A and B, and applied to the corresponding gate driver 28A, 28B as signals GateA and GateB respectively.

[0071] The gate drivers then drive the gate voltages of the transistors TA, TE high or low to switch the respective transistor on or off, in accordance with the level of signals GateA and GateB.

[0072] Figure 4 shows a more detailed example of a control circuit, constructed from discrete logic blocks, for controlling the operation of the transistors TA, TB, in accordance with the present invention.

[0073] The control circuits of figures 2 and 4 are illustrated as discrete logic blocks for the purpose of demonstrating the operation of the control circuit. In practice, the control circuit would typically be implemented in a small programmable logic device or a -20 -small microprocessor. Such devices are small and cheap, and operate at low power.

[0074] The embodiment described above in relation to figures 2, 3a, 3b and 4 is an on/off device, which enables the appliance to which it is connected to be powered on and off from a remote location. However, for certain types of load, such as lighting or motor controls, it is also desirable to regulate the effective voltage applied to the load so as to control the amount of power drawn by the load during use. This is particularly desirable in lighting circuits for the purpose of light dimming.

[0075] This can be achieved with the switching device of figures 2 and 4, by repeatedly switching the transistors TA and TB with a predetermined timing in relation to the supply frequency, such that power is applied to the load for less than the full half-cycle.

[0076] However, this requires one of the two rules I) and II) described above to be broken. The present inventor has established that, depending on the nature of the load, one or other of the rules (but not both) can be broken safely.

[0077] Specifically, if the load is capacitive, switch-on must occur while the voltage across each transistor is near zero, as occurs when the associated diode is conducting, but switch-off may be at any time. On the other hand, if the load is inductive, switch-on may occur at any time, but switch-off must occur during the period when the current through the transistor channel is zero or flowing in the reverse direction. Thus, in the case of a capacitive load, only rule I) need be applied, -21 -whilst in the case of an inductive load, only rule II) need be applied.

[0078] In a second embodiment of present invention, the switching device is able to automatically select which of these modes of operation should be used. To do this, the switching device must be able to determine the nature of the load.

[0079] In general, it is possible to determine the inductive/capacitive nature of a load by observing the relative timing of the voltage and current zero-crossings. The phase of the current relative to the voltage can then be determined, and hence the inductive/capacitive nature of the load. However, once power is supplied to the device, it may be too late in the case of a capacitive load.

[0080] With the present invention, the nature of the load may be determined by "probing" the load.

[0081] Figure 5 illustrates a circuit 50 for a switching device in accordance with the second embodiment of the present invention.

[0082] As in the circuit of figure 2, the circuit of figure 5 comprises a pair of MOSFET5 TA and TB connected in series between an AC supply 22 and a load 24. The gate voltages of the transistors TA, TB are controlled by separate gate drivers 28A and 28B, such that the transistors can be turned on and off independently of one another.

[0083] A voltage polarity sensing amplifier 32V is connected across the supply terminals 22 for determining the polarity of the supply voltage V. A current sensing -22 -resistor 34 is connected between the two transistors A, TE and a current polarity sensing amplifier 32C is connected in parallel with the resistor, for determining the polarity of the load current.

[0084] Each of the voltage and current polarity sensing amplifiers 32V, 32C has a respective output, which carries a logic level signal Po1V, Po1C which indicate the polarity of the supply voltage and the load current respectively.

[0085] Po1V and Po1C are provided as inputs to an input selector 55. A supervising microprocessor (not shown) provides a selecting signal to the input selector, for selecting one of Po1V and Po1C to provide a synchronising input to a counter 57. The counter is locked to the selected input either by phase-locking or by direct triggering.

[0086] Comparison registers 56A and 56E are connected to the counter 57. The comparison registers are loaded by the microprocessor, to emit logic signals when the counter value matches their content. In this way, a logic level signal having a controllable time relationship with a selected one of Po1V and Po1C is provided as an input to each of input terminals CK of respective D-type latches 38A and 38B.

[0087] These signals are used to load the latches 38A, 38B with a logic level ON/OFF command signal, which then controls the states of the switching transistors TA, TB.

[0088] An operational amplifier 59 is connected in parallel with the resistor 32. The operational amplifier amplifies the voltage developed across the current sensing resistor, and applies the amplified signal to an -23 -analogue to digital converter 51. The digital signal qenerated by the ana]oque to diqita] converter is input to the microprocessor.

[0089] Depending on the specific configuration and power of the microprocessor, much of the functionality of the control circuitry illustrated in figure 5 may be contained within the microprocessor as integrated peripherals or implemented in software.

[0090] Figure 6 illustrates the waveforms of the supply voltage V. the load current I, PolV, PolC and the logic level signals applied to the gates of A and T during switch on, switch off and probing.

[0091] Po1V and PolC are derived from the voltage and current waveforms respectively. Depending on whether the switch state is on or oft, either Po1V or Po1C is selected to synchronise the counter. Alternatively, the values placed in the comparison registers may be modified to reflect the timing difference between Po1V and Po1C.

[0092] During switching ON, the registers are loaded with values which cause each latch 38A, 38B to be triggered during the half-cycle when the voltage on the corresponding transistor is near zero.

[0093] During switching OFF, triggering takes place during the half-cycle when the current is zero or flows through the channel in the reverse direction of the switchable element.

[0094] When the switching device is OFF, the load is probed just before the voltage reaches zero. The exact timing of the probing pulse will be chosen depending on the known limiting resistance of the circuit, the -24 -tolerance to single pulses of current and the accuracy with which the timing can be generated. For example, the switch-on may occur approximately O.lms before the voltage reaches zero. This results in a short, low voltage pulse to the load. The resulting current pulse is captured by the analogue to digital converter and presented to the microprocessor for analysis. If the current pulse does not exceed a predefined limit, typically approximately equal to the rated service current of the device, the probing pulse time is made incrementally earlier until the full supply voltage is applied, while continuing to analyse the current pulse.

If the current pulse still does not exceed the predefined limit, it is concluded that the circuit can safely be operated in such a way that each switching device is switched on during the period when the forward voltage across it is not zero, i.e. rule I) is broken.

[0095] On the other hand, if, as the probing time is made earlier, the probe current exceeds the predefined limit, this indicates a capacitive load. In this case, the controller must switch to the fully ON/OFF mode or, if dimming is required, to a mode in which switching ON of each transistor occurs during the half-cycle when the voltage is near zero.

[0096] An alternative strategy to probe the load, is to switch on either TA or TB during the zero-voltage phase and off just after the zero crossing. The resulting voltage pulse is then compared in a similar way with a predefined limit, as the probing pulse width is increased, to determine whether the load is inductive.

[0097] In practice a small microprocessor would be required to implement the control circuit of figure 5.

However, the savings achieved by eliminating the snubber -25 -circuits required with previously known devices significantly outweigh the cost of the control circuit, both in terms of component cost and in terms in the energy saved.

[0098] Moreover, the second embodiment of the present invention represents a universal device which may be connected to any electrical appliance, irrespective of the nature of the load, since the device will automatically determine the nature of the load and select an appropriate mode of operation to prevent failure of the device.

[0099] The reduced energy consumption of devices which embody the present invention means that the heat sink, if one is required at all, will be very small, and there will be little need for ventilation. This means that the device can be made smaller than previously known switching devices. In this respect, it is envisaged that a switching device embodying the present invention may be contained within a volume of less than 10cm3. This is considerably smaller that previously known switching devices which serve a comparable purpose.

[00100] It will be appreciated that by providing a cost effective, low power and low volume switching device which can operate with both inductive and capacitive loads, the present invention overcomes the previous obstacles to the implementation of an energy management system in which the power consumed by individual appliances is individually monitored and controlled.

[00101] An energy management system which comprises a plurality of energy management devices which incorporate the switching device of the present invention is described below with reference to figure 7.

-26 - [00102] Figure 7 shows a block diagram of an energy management system 60. The energy management system comprises a central control device 62, which communicates wirelessly or by other means with a plurality of energy management devices 70, each of which is connected to a load device 66, such as a lighting circuit, washing machine, computer, curtain motor, or other electrical appliance.

[00103] The central control module 62 comprises a receiver and transmitter (not shown) for communicating with a communication module 68 incorporated within each energy management device 70.

[00104] The energy management devices 70 may be connected to the load device 66 in several ways. For example, they may be built into the load devices at manufacture.

Alternatively, they may be manually wired in at the time of installation of the load device, for example, in the form of a module fitted in or near a light fitting, or they may be installed in the backbox of a power outlet.

They may also be built into a power cord or incorporated into a standard mains plug.

[00105] Each energy management device 70 comprises a communication module 68, which comprises a transmitter and receiver (not shown) for communicating with the central control module 62, and a monitoring/switching module 64, for monitoring and controlling the power supplied to the load.

[00106] The energy management device 70 is shown in more detail in figure 8. The monitoring/switching module 64 comprises a pair of MOSFET transistors TA, TE connected in series between an AC power supply 22 and a load 24.

-27 - [00107] A voltage monitoring module 72V is connected across the supply terminals for monitoring the voltage waveform at the load. A current sensing resistor 34 is connected between the two transistors A, B, and a current monitoring module 72C is connected in parallel with the resistor, for monitoring the waveform of the load current.

[00108] The voltage and current monitoring modules 72V, 72C observe and analyse the respective voltage and current waveforms, to calculate the value of various parameters of the waveform.

[00109] Specifically, the voltage monitoring module 72V provides as an output a logic level signal Po1V which represents the voltage polarity and an analogue signal, InsV, which is a measure of the instantaneous voltage.

Similarly, the current monitoring module 72C provides as an output a logic level signal Po1C which represents the current polarity and an analogue signal, InsC, which is a measure of the instantaneous current. The relative phase of the voltage and current waveforms may be derived from these signals.

[00110] The polarity outputs Po1V, Po1C are provided as respective inputs to a control circuit 76, which switches the transistors A, B in accordance with rule I) or rule II) or both, depending on the nature of the load, as described above in relation to figures 2-4.

[00111] InsV and InsC are provided as inputs to a power calculating module 78, which multiplies the values of these inputs and integrates the result over time to determine a measure of the energy consumed. The resulting measure is provided as an output M. -28 - [00112] Because the power calculation is based on the instantaneous product of voltage and power, with due regard to the polarity of that product, the power measurement is not affected by the level of reactive or wattless current, which flows in an inductive or capacitive load, but dissipates no power.

[00113] Output M is provided as an input to the communication module 68, which communicates this value to the central control module 62.

[00114] By determining the relative phase of the voltage and current waveforms, the power calculating unit has information available which enables it to determine the nature of the load. That is to say, whether the load is reactive, and if so, whether it is inductive or capacitive. This information is provided as signals Ni and N2, to the control circuit 76. This enables the control circuit to apply rules I) and II) selectively in accordance with the nature of the load. Note that, generally, this transfer of information takes place within a microprocessor, so signals Ni and N2 are represented by bits in memory and do not appear as external signals.

[00115] The communication module 68 of the device comprises a receiver (not shown) for receiving an ON/OFF command from the central control module 62, or another associated remote control device. The communication module converts this command to a numerical Level signal which is provided as an input to the control circuit 76.

The Level signal may be a logic level representing the ON/OFF state of power to the load, or may alternatively be a variable quantity specifying the effective voltage to be applied to the load. This enables the supply of -29 -power to the load device 24 to be controlled remotely, from the central control module 62, or another remote control device (not shown) [00116] In operation, data from the power calculating module 78 of each energy management device 64 is communicated to the central control module 62 of the system 60 by the respective communication module 68.

[00117] The central control module records the energy consumption of each of the associated load appliances over time, analyses this data, and provides output information to a user, for example by means of an integrated user interface, or via an associated PC.

[00118] This allows the user to be aware of their overall power consumption, and the amounts of power consumed by individual devices. The user is then able to switch off, or reduce the amount of power supplied to certain appliances, in response to this data, directly from the central control module.

[00119] Moreover, the central control module may be configured to automatically switch appliances on or off, or reduce or increase the power supplied to specified appliances in response to detected events. For example, specified devices can be switched off, or controlled to operate at low power, if the overall power consumption exceeds a certain level, or at certain times of day.

[00120] The power consumption data received from the energy management devices 70 is processed by the central control module, in accordance with predefined settings programmed into the control module by the operator, to identify specific appliances whose power consumption can be reduced or eliminated to achieve the optimum reduction -30 -in power consumption over the whole building, such that load shedding can be done in a controlled, strategic way.

[00121] Having identified one or more appliances whose power consumption is to be reduced or eliminated, the central control module communicates a suitable instruction to the communication module 68 of the energy management device 64 associated with the or each appliance. This instruction is passed to the control circuit 76 of the energy management device, which acts to interrupt the load current in the appliance, so as to reduce or eliminate the power drawn.

[00122] Full power may be restored to the appliance in a similar manner at a later time when the total power consumption of the building is at a lower level, at a predetermined time, or in response to a manual instruction from a user.

[00123] It is thus possible to use the central control module to implement a fully automatic load-shedding scheme based on articulated strategies, the precise details of which will depend on the requirements of the user.

[00124] The central control module is also able to determine additional information of various kinds from the power consumption information received from the energy management devices. For example, the system can adjust the commanded level setting of an appliance to bring the power consumption by that appliance to a desired level, can verify that each command has been acted upon by observing the resulting change in power consumption by the appliance concerned, or can identify an unexpected drop in power consumption due to the -31 -failure of a bulb in a lighting circuit, and may thus alert the user accordingly.

[00125] The system is also able to monitor how long an appliance has been switched on. Accordingly, the system may be configured to automatically switch off, or alert the user, when specified appliances such as an iron, or an electric oven have been left on for longer than a predetermined length of time.

[00126] More generally, the system is able to identify and alert the user in the case of unusual operation of any device. For example, excessively long or short operation, or excessive time between operation of pumps or refrigerators, etc, which could indicate potential flooding or thawing. Moreover, in some appliances, notably electric motors or fluorescent lights, the system is able to identify current consumption patterns characteristic of incipient failure, so that pre-emptive repair action can be taken.

[00127] In cases where the energy management devices are coupled to an automatic PIR controlled lighting system, the system may also be configured to act as an intruder alarm by alerting the user, or triggering an audible alarm in response to unexpected activation of the PIR devices.

[00128] A further advantage of combining energy measurement with control is in calibration of the zero point of the current measurement, which is crucial to the accuracy of power measurement. Since the load current is known to be zero when the switch is in the OFF state, the zero point can be set frequently to avoid errors due to drift over time.

-32 - [00129] The present invention has been described in terms of specific embodiments. However, it will be apparent to the skilled person that alternative implementations of the invention are possible.

[00130] In particular, the present invention has been described in terms of a 50Hz supply frequency. However, it will be appreciated that the principles of the invention apply equally to alternative supply frequencies such as 60Hz.

Claims

-33 -CLAIMS1. A switching device for controlling the supply of AC current from a power supply to a load, the device comprising: -a pair of switchable elements, each having a channel for carrying current in forward and reverse directions, and being switchable between an on state for allowing current to flow through the respective channel and an off state for blocking the flow of current therethrough, wherein the switchable elements are connected such that, in said on state, current is carried in the forward direction of the respective switchable element during respective alternate half cycles; and control means for switching the pair of switchable elements between said on and off states to control the flow of current from the supply to the load, wherein the control means is configured to switch one of the switchable elements at a different instant to the other switchable element.
2. A switching device as claimed in claim 1 wherein the control means is configured to switch the switchable elements to said on state, when the voltage across the respective switchable element is substantially zero.
3. A switching device as claimed in claim 2 wherein the control means is configured to switch the switchable elements to said on state approximately 5ms after the preceding zero-crossing of the supply voltage waveform.
4. A switching device as claimed in any preceding claim wherein the control means is configured to switch the switchable elements to said off state when the current through the channel of the respective switchable element is zero or flows in the reverse direction of the switchable element.

-34 -
5. A switching device as claimed in claim 4 wherein the control means is configured to switch the switchable elements to said off state approximately 5ms after the preceding zero-crossing of the load current waveform.
6. A switching device as claimed in any preceding claim wherein the control means is configured to switch the state of the switchable elements such that one or both of rules I) and II) are satisfied, wherein rule I) requires each switchable element to be switched on at an instant when the voltage across the respective switchable element is substantially zero, and rule II) requires each switchable element to be switched off at an instant when the current through the respective switchable element is zero or flows in the reverse direction of the switchable element.
7. A switching device as claimed in claim 6 wherein the control means is configured to satisfy both rules I) and II).
8. A switching device as claimed in claim 6 wherein the control means is configured to satisfy rule I) in the case of a capacitive load, and rule II) in the case of an inductive load.
9. A switching device as claimed in any preceding claim wherein the switching device comprises load analysing means, for determining whether the load is inductive or capacitive.
10. A switching device as claimed in any preceding claim wherein the switchable elements are MOSFET5.

-35 -
11. A switching device for controlling the supply of AC current from a power supply to a load, substantially as hereinbefore described with reference to the accompanying drawings.
12. An energy management device for measuring and controlling the supply of power to a load, the device comprising energy monitoring means for determining the power consumed by the load, and a switching device as defined in any preceding claim, for controlling the supply of power to the load.
13. An energy management device for measuring and controlling the supply of power to a load, the device comprising: -monitoring means for monitoring the voltage and current waveforms for the load, and determining a measure of the power consumed by the load based on the integration over time of the product of instantaneous voltage and current; switching means for controlling the supply of power to the load; load analysing means for determining the nature of the load from the relative phase of said voltage and current waveforms; and control means for switching said switching means in accordance with the determined nature of the load.
14. An energy management device as claimed in claim 13 wherein the switching means comprises a pair of switchable elements, each having a channel for carrying current in forward and reverse directions, and being switchable between an on state for allowing current to flow through the respective channel and an off state for blocking the flow of current therethrough, wherein the switchable elements are connected such that, in said on -36 -state, current is carried in the forward direction of the respective switchabiLe element during respective alternate half cycles; wherein the control means is configured for switching the pair of switchable elements between said on and off states to control the flow of current from the supply to the load, wherein the control means is configured to switch one of the switchable elements at a different instant to the other switchable element.
15. An energy management device as claimed in claim 13 or 14 wherein the control means is configured to switch the switchable elements to said on state when the voltage across the respective switchable element is substantially zero.
16. An energy management device as claimed in any of claims 13 to 15 wherein the control means is configured to switch the switchable elements to said off state when the current through the channel of the respective switchable element is zero or flows in the reverse direction of the switchable element.
17. An energy management device as claimed in any of claims 13 to 16 wherein the control means is configured to switch the state of the switchable elements such that one or both of rules I) and II) are satisfied, wherein rule I) requires each switchable elements to be switched on at an instant when the voltage across the respective switchable element is substantially zero, and rule II) requires each switchable element to be switched off at an instant when the current through the channel of the respective switchable element is zero or flows in the reverse direction of the switchable element.
18. An energy management device as claimed in claim 17 wherein the control means is configured to satisfy rule -37 -I) when the load is determined to be a capacitive load, and rule II) when the load is determined to be an inductive load.
19. An energy management device as claimed in any of claims 13 to 18 wherein the switchable elements are MOSFET5.
20. An energy management device for measuring and controlling the supply of power to a load, substantially as hereinbefore described with reference to the accompanying drawings.
21. An energy management system, comprising a central controller and one or more energy management devices, as claimed in any of claims 13 to 20.
22. An energy management system substantially as hereinbefore described with reference to the accompanying drawings.
23. A device, forming a component of an energy management system, which comprises both an energy monitoring function and a power control function, in which information derived in the course of monitoring power to an appliance is used to direct the operation of the power control function in such a way as to protect it from damage by electrical transients and maximise its efficiency and reliability.
24. A remote controlled power controlling device whose operation is supervised and verified by a co-located power monitoring device in order to provide feedback to the remote controller.

-38 -
25. A system comprising a device as claim 23 and a remote controlled power controlling device as claimed in claim 24.
26. A semiconductor-based switching device suitable for controlling the flow of AC mains power to an appliance in which two switching elements carry the load current during alternate half cycles and in which each switching element is separately controlled.
27. A semiconductor-based switching device as claimed in claim 26 in which the switching elements are MOSFET5.
28. A semiconductor-based switching device as claimed in claim 26 or 27 in which the timing of the control signals to the two elements is constrained so that switch-on always occurs when the voltage stress upon that element is near zero and switch-off always occurs when the current in that element is zero or flows in the reverse direction of the switchable element.
29. A semiconductor-based switching device as claimed in claim 26 or 27 in which the timing of the control signals to the two switching elements is adjusted in response to the electrical characteristic of the load in such a way that when the load is inductive, switch-on may occur at any time, but switch-off must occur when the current through the channel of the respective switchable element is zero or flows through the channel in the reverse direction of the switchable element, and if the load is capacitive, switch-on must occur while the voltage is near zero, but switch-off may be at any time.