GB2159360A - Power supplies - Google Patents

Abstract

A power supply circuit comprises first and second SCRs (8,9) operative to pass a cyclically reversing current through a load circuit (17,1,2) from a d.c. supply (Vs). The SCRs are turned on alternately by a circuit comprising a first transformer (36P,36S) having a primary winding section (39) through which free wheel current passes during commutation of the first SCR (8), and a secondary winding (36S) coupled to the gate of the second SCR (9) so that cessation of the free wheel current through the primary winding section (39) of the first transformer induces a signal in the secondary winding (365) for firing the second SCR (9). A second transformer (27P,27S) has a primary winding section (59) through which free wheel current passes during commutation of the second SCR (9), and a secondary winding (275) coupled to the gate of the first SCR (8) so that cessation of the free wheel current through the primary winding section (59) of the second transformer induces a signal in the secondary winding (27S) for firing the first SCR. This configuration ensures that switching of the SCRs takes place in a zero current condition, so that generation of interference is greatly reduced. <IMAGE>

Description

SPECIFICATION Power supplies This invention relates to power supplies, and particularly, but not exclusively, to switched mode power supplies for supplying power to discharge lamps, such as fluorescent lamps, and to power supplies in which power factor correction is provided.

Switched-mode power supply circuits for operating fluorescent lamps have previously been proposed in which thyristors, silicon controlled rectifiers (SCRs), bipolar transistors or power FETs are used to pass an alternating current through a transformer to which one or more lamps are connected.

However, the known power supply circuits suffer from a number of problems. Firstly, the switching devices operate in such a manner that the switching takes place at a time when current is still flowing in the circuit. The resulting sudden interruption of the current gives rise to radio-frequency interference of a level which is far higher than that permitted under the various standards such as VDE 0871, CISPR, and BS800.

A second problem, which is possibly even more important than the high level of interference, is the low power factor which such circuits present to the supply mains. As is well known, the electricity supply authorities require that the power factor of any load connected to the supply system shall be as near to unity as possible. Wattless volt-amperes load the generators and lines of the system, but are not paid for by the consumer. The earning capacity of the system is therefore reduced. Furthermore, the voltage regulation of the transmission lines of the supply system is degraded by low power factor loads.

It is an object of the present invention to provide an improved power supply.

According to the present invention, there is provided a switched-mode power supply for supplying alternating current to a load from a direct current supply, comprising first and second semiconductor switching devices connected in series across the d.c. supply with a junction therebetween; load supplying means connected to the junction; and means to switch the devices on alternately to supply the alternating current to the load; wherein the means to switch the devices on alternately comprises a first transformer having a primary winding section arranged to pass free wheel current during commutation of the first device, and a secondary winding coupled to a control electrode of the second device so that cessation of the free wheel current through the primary winding section of the first transformer induces a signal in the secondary winding for switching the second device; and a second transformer having a primary winding section arranged to pass free wheel current during commutation of the second device, and a secondary winding coupled to a control electrode of the first device so that cessation of the free wheel current through the primary winding section of the second transformer induces a signal in the secondary winding for firing the first device.

An embodiment of the invention will now be described, by way of example, with reference to the accompanying drawings, in which: Figure 1 is a circuit diagram of a switched mode power supply for supplying power to one or more fluorescent lamps; Figure 2 illustrates waveforms occurring in the circuit of Fig. 1; Figure 3 is a simplified schematic diagram of a switched mode power supply, for use in explaining the theory of a power factor correction circuit; Figure 4 is a voltage-current curve of a modified current transformer; Figure 5 is a simplified schematic diagram of a switched mode power supply including a modified power factor correction circuit; and Figure 6 illustrates current and voltage waveforms.

Referring to Fig. 1 of the drawings, a power supply circuit for supplying power to lamps 1 and 2 comprises a half-bridge inverter 3 which is energised by power from a bridge rectifier circuit 4. The circuit 4 is connected to an alternating current mains supply 5. Capacitors 6 and 7 help to suppress mains-borne interference.

The inverter 3 comprises series-connected silicon controlled rectifiers (SCRs) 8 and 9 and series-connected capacitors 10 and 11 between positive and negative d.c. supply lines 1 2 and 1 3. The anode of the SCR 9 and the cathode of the SCR 8 are interconnected at a junction 14, and the capacitors 10 and 11 are interconnected at a junction 1 5. A primary winding 1 6 of a transformer 1 7 and a primary winding 1 8 of a transformer 1 9 are connected in series between the junctions 14 and 1 5.

Heater electrodes 20 and 21 of the lamp 1 are connected to tappings on a secondary winding 22 of the transformer 16. It is assumed, for the present, that tappings X and Y on the winding 22 are interconnected. Similarly, heater electrodes 23 and 24 of the lamp 2 are connected to tappings on a secondary winding 25 of the transformer 16. The d.c. supply is smoothed by a capacitor 26.

In operation of the circuit so far described, the capacitors 10 and 11 are charged from the d.c. supply lines 1 2 and 13, and are alternately discharged through the transformer windings 16 and 18, via the switching SCRs 9 and 8 which are fired alternately by circuitry which will be described later. An alternating current therefore flows through these transformer windings.

The capacitors 10 and 11 and the windings of the transformers 1 7 and 1 9 together form a resonant circuit at the operating frequency of the inverter, which may, for example, be in the 20KHz to 25KHz range. The current through the windings is, therefore, substantially sinusoidal.

Resonance in the circuit causes a voltage to be generated across the winding 1 6 so that a relatively high voltage is applied between the electrodes 20 and 21. Heating current flows through the electrodes from the respective sections of the winding 22. The high voltage produced on the secondary winding causes ionisation of the gas in the iamp, and the lamp lights. Once the lamp has struck, the voltage between the electrodes falls to the normal lamp running voltage, which is determined by the lamp characteristic. The relatively low impedance of the lamp when struck damps the resonant circuit.

The circuit for firing the SCR 8 includes a transformer secondary winding 27S, one end of which is connected to the cathode of the SCR. The other end of the winding is connected, via a resistor 28 and a capacitor 29 connected in parallel, to the gate of the SCR. A capacitor 30 is connected across the winding 27S, as are a diode 31 in series with a zener diode 32. A capacitor 33 is connected between the gate and the cathode of the SCR 8, together with a resistor 34 and a capacitor 35 in series. A transformer primary winding 36P has a tapping 37 to which the anode of a diode 38 is connected. The cathode of the diode 38 is connected to the d.c. positive line 1 2. The winding 36P is divided into two sections 39 and 40 by the tapping 37.The free end of the section 39 is connected to the cathode of the SCR 8, and the free end of the section 40 is connected to the cathode of a diode 41. The anode of the diode 41 is connected to the gate of the SCR 8 via a resistor 42. A resistor 43 is connected between the line 1 2 and the junction 14.

The firing circuit for the SCR 9 has resistors 45-47, capacitors 49-52 and diodes 53-56 which correspond to the components in the SCR 8 circuit. A transformer winding 36S which corresponds to the winding 27S of the latter circuit is magnetically coupled to the winding 36P.

A winding 27P which corresponds to the winding 36P is magnetically coupled to the winding 27S. The winding 27P has a tapping 57 which divides the winding into sections 58 and 59.

The SCR 9 firing circuit has additional components for starting up the circuit when the apparatus is first switched on. These components comprise a capacitor 61 and a resistor 60 connected in parallel between a junction 62 and the negative d.c. line 13; a resistor 63 connected between the junctions 62 and 14; a resistor 64 and a diode 65 connected in series between those junctions; and a diac 66 and a resistor 67 connected in series between the junction 62 and the gate of the SCR 9.

The switching action of the SCRs will now be described. When the supply is first switched on, the potential of the junction 1 4 rises to a level between the potentials of the positive and negative lines 1 2 and 13, due to the potential dividing action of the resistors 61, 63 and 43 and the capacitor 61. The capacitor 61 charges up, and when the voltage at the junction 62 reaches the breakover voltage of the diac 66 the diac conducts and a trigger signal is applied to the gate of the SCR 9 at a time t1, and the SCR is fired. The voltage Va at the junction 14 falls substantially to zero, as shown in Fig. 2A.Current i flows from the capacitors 10, 11, and through the windings 18 and 1 6 and the SCR 9 to the line 1 3. Due to the resonant nature of the circuit, the current rises substantially sinusoidally to a peak and then falls to zero at a time t2, causing commutation of the SCR 9 to begin. The current then overshoots and reverses, as shown in Fig. 2B, the "free-wheel" diode 55 providing a path for the reverse current from the line 13, through the section 59 of the winding 27P, to the junction 14. The voltage Va also overshoots and becomes negative. The reverse current produces a magnetic flux in the core of the transformer formed by the windings 27P and 27S, so that a voltage Vb(Fig.2c) is induced across the winding 27S.The phasing of the windings is such that the voltage Vb, at this point in time, is negative at the upper end of the winding 27S as viewed in Fig. 1. The voltage is clamped by the diodes 31 and 32. This negative voltage is applied to the gate of the SCR 8 with respect to its cathode. At the same time, a negative voltage is induced in the section 58 of the winding 27P. This is applied via the diode 56 and the resistor 47 to the gate of the SCR 9 to speed up the commutation.

At a time t3 the reverse current flow in the winding 27P reaches zero, and the magnetic flux in the transformer core collapses. This causes a relatively large "flyback" voltage to be induced in the winding 27S, making the upper end of the winding positive. This applies a gating pulse to the SCR 8. The SCR fires, and the current i starts to flow through the SCR and the transformer windings 1 6 and 1 9 from the line 12 into the capacitors 10, 11. The current between the time t3 and a time t4 has the same waveform as between the times t1 and t2, but flows in the opposite direction. The voltage Va at the junction 14 rises at t3 to a level which is less than the d.c. supply voltage Vs by an amount equal to the voltage drop across the conducting SCR 8. V, stays at that level until the time t4, and then overswings as the SCR 8 turns off. The current i overswings from the time t4 to a time t5, similarly to t2 to t3 but in the opposite direction of flow, the reverse current flowing through the free wheel diode 38 and the section 39 of the winding 36P. Just as explained above for the transformer 27P,27S, a negative voltage VG(Fig.2D) is induced across the secondary winding 36S, holding the SCR 9 off. A negative voltage is induced in the winding section 40 which-speeds up the commutation of the SCR 8. When the reverse current falls to zero at the time t5, the magnetic field in the transformer core collapses. A positive voltage is induced in the winding 36S, turning on the SCR 9. The inverter continues to operate cyclically in this manner.

The shape of the voltage V9 applied to the gate of the SCR 9 is shown in Fig. 2E. The voltage applied to the gate of the SCR 8 will be the same, but displaced in time.

It should be noted that each SCR is turned off and the other turned on at instants when the current i is at zero. Hence, a great reduction in radio-frequency interference is achieved.

The circuit comprising the diac 66, the resistors 60, 63 and 67 and the capacitor 61 is operative, as previously explained, to ensure that the SCR 9 turns on soon after the supply 5 has been connected to the apparatus. However, it is essential that that circuit shall not operate again during the cycling of the inverter, otherwise miscommutation of the SCR 9 would occur.

The resistor 64 and the diode 65 ensure that that circuit remains inoperative by discharging the capacitor 61 every time the voltage V5 of the junction 14 goes to zero. The voltage of the junction 62 cannot, therefore, rise sufficiently to fire the diac 66.

As is well-known, some SCRs are prone to arbitrary firing due to the parasitic capacitance between the anode and the gate. If, at switching off, the rate of change of the anode/cathode voltage is high, sufficient current can be fed into the gate through the capacitance to cause the SCR to fire again. This must be prevented, and in the present circuit the networks 34, 35 and 46, 51 are provided to apply a negative bias to the gate of the respective SCR to prevent this occurring.

The capacitors 33 and 52 act as decoupling capacitors to smooth out noise spikes which may appear in the respective SCR gating circuits, and which could otherwise cause unwanted firing of the 5CRs.

The transformers 27P,27S and 36P,365 can comprise windings on quite small, lowpermeability toroidal cores, or on ferrite cores with gaps to reduce their effective permeabilities.

If the inverter is overloaded by the connection of too many lamps thereto, the freeqheel current will be reduced and will cause a reduction of energy in the cores of the transformers, so that SCR firing pulses will not be delivered by the secondary windings 27S and 36S. Cycling of the inverter will therefore cease, and overload protection is thereby provided.

The transformer 1 7 is, in effect, a current transformer with a high leakage reactance. Fig. 1 shows, by way of example, two lamps connected to the transformer. The combined secondary circuits of the transformer will reflect a particular inductance into the primary winding, and the resultant inductance at the terminals of the primary winding, together with the reactances of the transformer 1 9 and the capacitors 10 and 11 will determine the resonant frequency of the inverter. The inverter will therefore switch at a given frequency.

If, now, one of the lamp circuits be disconnected, the resultant inductance at the transformer primary terminals will be increased, and the inverter switching frequency will decrease. This will reduce the inductive reactance of the transformers, and the current flowing through the primary winding 1 6 of the transformer 1 7 will decrease. By suitable design of the transformer 17, it can be achieved that the decreased current through the transformer impedance results in the reduction of current in the transformer primary winding 16, and hence the same current applied to the single lamp as would be applied to each of the previously-mentioned two lamps at the higher switching frequency.

Similarly, the inverter can be loaded up with more lamp circuits, as required, until the overload point is reached and the cycling of the inverter stops, as explained previously. The more lamps supplied by the circuit, the higher the switching frequency.

When the supply is first switched on, the voltage (e.g.800 volts) produced at resonance will appear across the secondary winding 22 or 25, and a fraction of this (say,6 volts) is tapped off at each end of the winding for the lamp heaters, the fraction being determined by the turns ratios. The heaters ensure that enough free elections are produced at each cathode to provide the required conduction through the gas, without stripping ions from the cathodes. Such ion stripping could otherwise cause premature failure of the lamp. Once the lamp 1 or 2 has fired, the voltage across the respective secondary winding 22 or 25 will be held at the running voltage across the discharge in the lamp (e.g. 1 20 volts).

It has so far been assumed that the tappings X and Y on each secondary winding 22 or 25 have been interconnected, so that the winding feeds the normal power requirement to the respective lamp. However, if required, the facility for dimming a lamp can be provided by connecting the main winding 70 of a magnetic amplifier 71 between the tappings X and Y. A control current fed through the control winding 72 of the amplifier then determines the impedance inserted between the tappings, and this impedance changes the effective impedance f the respective secondary winding 22 or 25, thereby changing the lamp current and, hence, he brightness of the lamp.

Alternatively, the tappings X and Y could remain interconnected, and a magnetic amplifier :ould be connected, in shunt mode, between the tapping X and a tapping Z on the secondary winding. In another alternative arrangement a magnetic amplifier may be connected in series vith the winding 16 to produce a similar effect.

The transformer 19 performs a very important function, in conjunction with a diode bridge 73 jind a choke 74 which are connected in series between the negative output terminal 75 of the fridge 4 and the negative supply line 1 3. As mentioned previously, known power supply 'circuits can present a low power factor load to the supply.

In order to achieve a unity power factor, two conditions must be satisfied. The current taken rom the supply system by the load must be in phase with the supply voltage, and the form actor of the current waveform must be correct.

The current and voltage waveforms of a switched mode power supply without power factor correction are illustrated in Fig. 6. It will be seen that the current waveform (Fig. 6A) comprises series of peaks. Although these peaks are almost in phase with the peaks of the supply voltage Fig. 6B), the form factor of the current waveform is very high, say 2 to 3. The power factor is :herefore very low, say 0.5.

Some known circuits have included complicated and/or undesirably large filter components in border to improve the power factor. The transformer 19, the bridge 73 and the choke 74 of the present invention provide a much improved means for power factor correction. The use of the tower factor correction circuit is not confined to inverters for lamp operation and the theory of :he operation of the circuit will now be explained in relation to any power supply circuit which ncludes a rectifier circuit for connection to an a.c. supply to provide therefrom a d.c. supply; a smoothing capacitor connected across the d.c. supply; and means to cause a cyclically reversing current to flow from the d.c. supply to a load.

The basic components of such a circuit are shown in Fig. 3 of the drawings. An a.c. supply 30 feeds a bridge rectifier 81, which produces a d.c. voltage VDC across a smoothing capacitor 32. Semiconductor or other switching devices, represented schematically as switches 83 and 34, close alternately to feed an alternating current through the primary winding 85 of a transformer 86. The secondary winding 87 of the transformer feeds a load R,. The primary Ninding 88 of a current transformer 89 is connected in series with the winding 85. The secondary winding 90 of the transformer 89 is connected across two corners 91 and 92 of a second diode bridge 93. The other corners 94 and 95 of the bridge 93 are connected to the output of the bridge 81 and to a positive d.c supply line 96, respectively.The polarity of the iiodes in the bridge 93 must be such that current can flow unimpeded from the bridge 81 to :he line 96.

The following voltages and currents are considered in the theoretical explanation: V,, is the instantaneous value of the voltage on the mains supply 80, Vinpk is the peak value of that voltage, I,, is the instantaneous value of the input current from the supply 80, ijnpk is the peak value of the input current, Vsi is the voltage between the corners 91 and 92 of the bridge 93, i is the load current through the windings 85 and 88, Vac is the voltage across the capacitor 82, i, is the current in the secondary winding 90 of the current transformer 89, V5 is the voltage between the corners 94 and 95 of the bridge 93, VcT is the voltage across the winding 90.

Considering one half cycle of the mains supply voltage Vln Vinpk sin wit (1) Assuming that VDC = Vmpkt i.e. the capacitor 82 remains charged at the peak mains voltage, :hen the voltage V will be given by

The voltage Vc, = VBE (3) The instantaneous power P taken from the winding 90 is given by P = icT.VcT (4) If the transformer 89 is designed to have appreciable leakage reactance, so that the opencircuit voltage does not exceed VDC, the transformer voltage VCT versus current jCT characteristic will be linear, as shown in Fig. 4.

When the potential at 94 is equal to the potential at 95, a short-circuit exists across the winding 90. In this condition the winding delivers maximum current PK, but zero power (since VCT is zero), as will be seen from Fig. 4.

When 94 is at zero potential and 95 is at Vpk the winding 90 is in an open-circuit condition in which the maximum voltage is developed across the winding, but no current flows.

The current jCT flowing in the winding 90 must taken the shape of the voltage Vx, the current being given by jCT = ipk sin wt (5) From equation (4), the instantaneous power P is given by P = ipk sin 'ot. Vpk (1-sin sot) Hence, P = Vpk ipk (sin #t - sin#t) (6) The average power PaV flowing from the winding 90 can therefore be defined as follows

Integrating (7), we have

When cot =

Since the design parameters must meet the requirement that Pin = Pout, the peak current ipk of the transformer 88 can be chosen.

For unity power factor, the peak power input from the supply must equal the peak power of the transformer 89

The RMS power PRMS of the mains waveform is given by

Hence, the average power Psv from the winding 90 is given by

Hence, the average power in the current transformer winding is approximately 19% of the mains supply RMS power.

The peak power demanded by the current transformer from the mains supply can be obtained by differentiating equation (6) 2P i.e. = Vpk ipk (cos cot-2cos cot sin xt) (13) 2cot and equating this to zero.

Hence, peak power is obtained when cos cot = 2 cos cot sin cot (14) i.e. 1 = 2 sin cot sin cot = + .-.t = 30 or 150 Substituting in equation (6), the peak power Ppk in the transformer 1 9 is given by

Hence, the extra peak power which must be passed through the SCRs due to the power factor correction is only 35% of the RMS load current. This is well within the normal design factor of safety, so no increase in SCR size is necessary.

By interconnecting the bridge 93 and the transformer 89 as shown, it is ensured that the current taken from the supply is always in phase with the supply voltage and has the correct waveform to give unity power factor, irrespective of the current taken by the load. If, at any instant, the load current is less than the correct value for unity power factor, the current transformer 89 will 'suck" extra current out of the supply, and this is achieved by monitoring the difference in potential between the input 94 and the output 95 of the bridge 93. The extra current is merely circulatory, so does not amount to an increase in consumed power.

It will be seen that circuit of Fig. 3 for which the above results have been derived comprises the basic components of the Fig. 1 circuit. In the case of Fig. 3 the extra bridge 93 is connected in the positive supply line for the sake of clarity of explanation, whereas in the Fig. 1 circuit the bridge 73 is connected in the negative line, with the polarity of the diodes reversed. This makes no difference to the relevance of the calculated results above.

The current waveform obtained for the Fig. 3 circuit showing the effect of the power factor correction circuit 19 and 73 is illustrated in Fig. 6C. It will be seen that the current waveform is substantially sinusoidal and is in phase with the mains voltage (Fig. 6B). A similar current waveform would be obtained for the Fig. 1 circuit.

Referring to Fig. 1, an SCR 100 is connected between the bridge output 75 and the line 1 3 so that, when the SCR is made to conduct, it short-circuits the choke 74 and the bridge 73. A resistor 101 is connected between the gate and the cathode of that SCR, and the gate is connected to the positive supply line 1 2 via a resistor 102 and zener diodes 103 and 1 04. In the event of open-circuit or short-circuit conditions occurring in the transformer 17, the power factor correction circuit can cause the d.c. supply voltage V, to exceed 400 volts, and the correction circuit must be disabled under those conditions. The resistors 101, 102 and the zener diodes 103, 104 form a voltage-sensitive potential divider which applies a firing signal to the SCR 100 when the voltage V, exceeds a predetermined level.

An alternative way of connecting the power factor correction circuit is shown in Fig. 5. This figure shows only the basic components and is similar in form to Fig. 3. Corresponding components in the two figures are numbered correspondingly.

In Fig. 5 the separate transformers 86 and 89 of Fig. 3 are replaced by a single transformer 105 having a primary winding 106 through which the main inverter current flows, a secondary winding 107 which acts as the power factor correction winding. The main winding 109 of a magnetic amplifier 110 is connected in series with the winding 108 across the corners 91 and 92 of the bridge 93.

The control winding 111 of the magnetic amplifier is connected to a control circuit 11 2 which monitors the d.c. supply voltage at the line 96. The circuit 112 feeds a control current to the magnetic amplifier 110 such that any changes in the d.c. supply voltage are compensated for by a change in the current fed through the transformer 105, thereby maintaining the correct waveform for the current taken from the supply 80.

Various modifications of the circuit of Fig. 5 may be made. For example, the capacitors 11 3 and 114 may be replaced by further switching devices to form a full bridge inverter. A capacitor 11 5 would then be added across the d.c. supply.

The subject matter of this application is also disclosed in British Patent Application No.

8303499, from which the present application is divided.

Claims (6)

1. A switched-mode power supply for supplying alternating current to a load from a direct current supply, comprising first and second semiconductor switching devices connected in series across the d.c. supply with a junction therebetween; load supplying means connected to the junction; and means to switch the devices on alternately to supply the alternating current to the load; wherein the means to switch the devices on alternately comprises a first transformer having a primary winding section arranged to pass free wheel current during commutation of the first device, and a secondary winding coupled to a control electrode of the second device so that cessation of the free wheel current through the primary winding section of the first transformer induces a signal in the secondary winding for switching the second device; and a second transformer having a primary winding section arranged to pass free wheel current during commutation of the second device, and a secondary winding coupled to a control electrode of the first device so that cessation of the free wheel current through the primary winding section of the second transformer induces a signal in the secondary winding for firing the first device.

2. A power supply as claimed in claim 1, comprising a respective diode connected in series with each primary winding section to ensure that only free wheel current flows therethrough.

3. A power supply as claimed in claim 1 or claim 2, comprising an auxiliary firing circuit coupled to the gate of one of said devices for switching that device to initiate the alternate switching of the devices; and means operative whilst the alternate switching is proceeding to disable the auxiliary firing circuit.

4. A power supply as claimed in any preceding claim, wherein the first transformer includes a further winding which is operative during the passage of free wheel current through the primary winding section to apply a negative voltage to the control electrode of the first device to speed up the commutation of that device; and wherein the second transformer includes a further winding which is operative during the passage of free wheel current through the primary winding section to apply a negative voltage to the control electrode of the second device to speed up the commutation of that device.

5. A power supply as claimed in any preceding claim, wherein the load supplying means includes a third transformer having a primary winding through which current is passed by the devices, and a secondary winding with tapped sections at its ends to feed heater current to respective heaters of a discharge lamp.

6. A power supply as claimed in claim 5, wherein the third transformer is connected in a resonant circuit, the resonance frequency of which increases as the load increases.