GB2037462A - Stabilised switched mode power supply - Google Patents

Abstract

A switched mode power supply circuit of the forward converter type which has a plurality of output circuits is provided with a control circuit 28, for stabilising an electrical quantity, such as voltage, at the output terminals 32, 33 of at least one of the output circuits. The relevant secondary winding 2 of the converter transformer feeds a flywheel circuit 14, 20 via a rectifier 13 and a saturable choke 27 which is provided with a control winding 28. During each "off" period of the converter the induced voltage across the control winding is clamped using feedback 31 at a value proportional to the output quantity to be stabilised thereby controlling the demagnetization of the core material of the choke which occurs during each "off" period, and hence the magnetization which it has when the next "on" period starts. In an alternative arrangement a steady current determined by the value of the quantity to be stabilised is passed through the saturable choke control winding. <IMAGE>

Description

SPECIFICATION Switched mode power supply The invention relates to a switched mode power supply circuit arrangement of the forward converter type, comprising a series com bination of a controllable switching element and an input winding of a transformer, which series combination is connected across a pair of direct-voltage input terminals, first and sec ond output circuits each having output termi nals for connection to a load, and a control circuit having its output coupled to a control input of said switching element so that said control circuit will render said switching ele ment periodically and alternately conductive and non-conductive in operation, each said output circuit comprising an output winding of the transformer, a rectifying diode, and a flywheel circuit comprising a diode and a choke, the first said output circuit being provided with means constructed to adjust and stabilise an electrical quantity at the output terminals of said first output circuit.

A known such power supply arrangement is disclosed in German Auslegeschrift 26 08 1 67. This German specification describes a forward converter having two electrically iso lated outputs. The voltage at one of these outputs is stabilized and adjustable independently of the other converter parameters by means of a controlled switching series-transistor in the output circuit. To this end the arrangement includes a separate electronic control circuit for the switching series-transistor. A variable part of the switching period of the transistor is not used effectively for building up and maintaining the output voltage, so that pulse-width control is effected in the corresponding output circuit, thereby obtaining the required adjusting and stabilizing action.

German Offenlegungsschrift 26 34 193 also describes a power supply arrangement of the type specified in the preamble. Pulse width control in each of a plurality of output circuits is in this case provided by a corresponding electronic circuit which controls a thyristor, which thyristor also serves as the rectifier in the corresponding output circuit.

Fig. 1 of this German Specification also indicated that not all the output circuits which are present need be provided with adjusting and stabilizing means, and that the input circuit may include two switching elements in series with the input winding, these being bypassed by a flyback diode. The two switching elements in this known arrangement are periodically turned on at the same time by a pulse generator.

The output control circuit in these known power supply circuit arrangements are of fairly intricate construction and comprise many electronic components including generally, driven transformers so as to enable the switching transistors or switching thyristors to be controlled correctly.

It is an object of the invention to provide a simpler arrangement.

The invention provides a switched mode power supply circuit arrangement of the forward converter type, comprising a series com bination of a controllable switching element and an input winding of a transformer, which series combination is connected across a pair of direct-voltage input terminals, first and second output circuits each having output terminals for connection to a load, and a first control circuit having its output coupled to a control input of said switching element so that said control circuit will render said switching element periodically and alternately conductive and non-conductive in operation, each said output circuit comprising an output winding of the transformer, a rectifying diode, and a flywheel circuit comprising a diode and a choke, the first said output circuit being provided with means constructed to adjust and stabilise an electrical quantity at the output terminals of said first output circuit, characterised in that the adjusting and stabilising means comprise a saturable second choke having a power winding and a control winding, and a second control circuit having an input for said electrical quantity and an output connected to said control winding, said power winding being connected in series with the output winding and the rectifying diode of said first output circuit and said second control circuit being constructed to apply a control signal to said control winding so as to control, in accordance with the value of said electrical quantity, the initial value which the magnetisation of the core material of said second choke has each time said switching element closes.

Provision of such a saturable choke enables the leading edge of each rectangular voltage pulse on the relevant output winding of the transformer to be effectively delayed in time before it is transmitted to the relevant flywheel circuit, the delay being proportional to the difference between the flux in the core of the saturable choke when the switching element closes and the saturation flux thereof.

(For a constant voltage across the power winding of the choke the flux change therein is effectively directly proportional to time.

Thus the time taken for the flux to change from its initial value to its saturation value is directly proportional to the difference between these values). Thus for a given transformer output winding voltage an adjustable time delay can be obtained by controlling the value which the magnetisation of the core material of the saturable choke has each time the switching element closes. This value can be adjusted by means of a simple control circuit, a parameter to be adjusted and stabilised, such as the output voltage or the output currant, being for example compared with a reference and the resulting difference signal being used to adjust said value by, for exampie, controlling a current passed through the control winding or a voltage applied to it.

Such an arrangement may have the following advantages with respect to the known arrangements: -tow cost price of the components used; expensive switching transistors or thyristors and an intricate control circuit can be avoided, -low dissipation. The power winding of the choke can have a very low resistance, whilst losses in the control winding, magnetic circuit and second control circuit can be very low, -a simple linear second control circuit may be employed.

It is conductive to high efficiency if the core material of the saturable choke has a remanence which is low with respect to the satura tio magnetisation of said material. If this is not the case it may be necessary to pass an additional control current through a winding of the saturable choke, possibly through a further control winding provided specifically for this purpose so as to magnetise the core of the choke in such a way that a steep part of the B-H-curve is traversed when the arrangement is in operation. Such a shift of the B-H curve would be likely to be necessary if a magnetic material having a rectangular loop (where the remanent magnetism substantially corresponds to the saturation flux) were used as the core material.

The second control circuit may comprise a controllable voltage source a control input of which is connected across said control winding via a series diode in such manner that when said switching element is closed said series diode will be reverse-biassed by the output voltage of said voltage source and when said switching element opens said series diode will become forward-biassed by the voltage then induced across the control winding.

Such a construction enables the initial value of the core material magnetisation to be determined by the integral of the output voltage of the voltage source with respect to time, taken over the period for which the series diode is forward-biassed, and hence can enable optimum use to be made of the magnetic energy stored in the saturable choke. With such a construction the integral of the power winding voltage with respect to time when the switching element is closed will be equal to the integral of the control winding voltage with respect to time when the switching element is open.Moreover, with such a construction it is not necessary to have a continuous flow of bias current in the control winding, and the voltage source can be constructed in a simple manner, for example as an amplifier having inverting and non-inverting inputs, a reference voltage source having its output connected to one input of the amplifier, and means for supplying a signal representative of said quantity to the other input of the amplifier.

The series diode may be constituted by the base-emitter junction of a transistor, the emitter and collector of said transistor being connected to respective ends of said control winding. Such an arrangement can enable the greater part of the demagnetizing current of the saturable choke to flow outside the amplifier, which is an advantage if the output of the amplifier can take up only a small external current. (The output of an operational amplifier normally supplies current.) If the first control circuit is a pulse-widthcontrollable pulse train generator the arrangement may include a comparator to a first input of which is coupled the output of said second output circuit, to a second input of which is coupled the output of a reference voltage source, and the output of which is coupled to a pulse width control input of said generator so as to adjust the width of the output pulses of said generator in such a sense as to maintain the voltage at the output of the second output circuit substantially constant. Such an arrangement can enable interaction between the first and second output circuits as a result of load variations to be made small and, moreover, the saturable choke to have small dimensions.With such an arrangement the second control circuit may be arranged to dissipate hardly any energy, for example less than 3 per cent of the nominal output power. The adjusting range of each output voltage may be for example + and 5%, whilst the stabilization obtained may be for example substantially better than 1%.

It should be noted that the use of a saturable choke for control purposes is known.

United States Patent Specification 3,087,107 describes a direct voltage supply source, which is energized from the a.c. mains via a transformer. A saturable reactor is included between the secondary winding of the transformer and a rectifying diode, which reactor receives a magnetic bias by means of two control windings. The bias is determined on the one hand by the load current and on the other hand by the output current from an amplifier circuit which compares the output voltage with a reference. The rectifying circuit is of the half-wave type. The full-wave version is disclosed in United States Patent Specification 3,105,184. Both these known power supply circuits operate with sine-wave input voltages and employ peak rectification, so that the delay produced by the reactor cannot become effective until after the sine-wave peak has occurred. As it is also stated that the sine-wave voltage is obtained from a supply mains whose frequency is 50 or 60 Hz, this and the foregoing imply that the reactor has to produce a long delay and thus that a large reactor or choke has to be used. Furthermore, losses in the known circuit tend to be fairly large, inter alia owing to the use of a magnetic material having a rectangular loop for the core material of the choke, so that a bias current has to be applied to a control winding of the choke. The present invention on the other hand, relates to multiple-output forward converters whose switching frequency, as is known, can be high.Moreover rectangular voltage waveforms are normally generated across the output windings of the transformer in such converters, so that the delay produced by the saturable choke in such an arrangement can be effective starting from the leading edges of such waveforms. The core material of the choke may, if desired be one which exhibits a relatively narrow B-H loop.

United States Patent Specification 3,445,775 describes the use of saturable reactors to give pulse-width control in a control circuit for a push-pull converter.

The squarewave output signal from an oscillator is pulse-width-controlled by a magnetic amplifier the core material of which has a rectangular loop. A bias control winding for creating a fixed magnetic bias is also provided, as is a controllable voltage source.

Embodiments of the invention will be described, by way of example, with reference to the accompanying diagrammatic drawings, in which: Figure 1 is a general diagram of an embodiment; Figure 2 shows a B-H curve illustrating a possible mode of operation of the embodiment of Fig. 1; Figure 3 shows a B-H curve illustrating another possible mode of operation of the embodiment of Fig. 1; Figures 4A and 4B show time diagrams of current, voltages and flux variations, illustrating two possible modes of operation of the embodiment of Fig. 1; Figures 5A and 5B show details of parts of current-time diagrams of Fig. 4A and 4B respectively; Figure 6 shows a possible construction for part of the embodiment of Fig. 1 in detail; Figure 7 shows an alternative to the construction of Fig. 6;; Figures 8, 9 and 10 show possible additions to the construction of Fig. 6 or alternatives to part of the construction of Fig. 7, and Figure 11 shows an alternative possible construction for part of the embodiment of Fig. 1 in detail.

Fig. 1 is a diagram of a switched mode power supply circuit arrangement of the forward converter type, having four electrically isolated outputs. For this purpose there is provided a transformer with secondary windings 2, 3, 4 and 5. It is obvious that these windings may have some turns in common, or may even form part of, and/or form extensions to, the primary winding 6 of the transformer (in which cases their mutual isolation will be lost). A demagnetizing winding 7 returns the magnetic energy stored in the transformer via the diode 8 to a direct-voltage, supply source (not shown), which is connected to the terminals 9 and 10 with the polarity indicated.A first control circuit 11 supplies control pulses to a controllable switching element 12, which may for example be a transistor, and which alternately connects the winding 6 to the terminals 9 and 10 and isolates it therefrom in response to these control pulses.

Each secondary circuit comprises a rectifying diode 13, which because of the nature of the forward converter conducts when the switch 1 2 is closed, a flywheel choke 1 4 capable of storing magnetic energy, of which a part is periodically supplied to the buffer capacitor 15 and to the load 16, 17, 18 or 19, which is connected to the output, and furthermore a fly-wheel diode 20, which takes up the current through the choke 14 during the times when the rectifying diode 1 3 cannot carry this current.

The secondary circuits 2-16 and 3-1 7 are provided with adjusting and stabilizing means for a quantity at the relevant output. A further secondary circuit 4-18 produces a non-stabilized output and the output of a further secondary circuit 5-19 is controlled by influencing the primary circuit. For this purpose the output voltage of the circuit 5-19 is applied to a first input of a comparator circuit 21 a second input of which (not shown) is connected to the output of a reference voltage source (not shown). Comparator circuit 21 supplies a control signal via connection 22 to a control input 23 of control circuit 11, so that for example the duration of the conduction times of the switch 1 2 is controlled in a sense such as to maintain the voltage across terminals 38 and 39 substantially constant.

Fig. 1 indicates that a similar form of stabilization may also be obtained by means of an additional secondary winding 24, connected to a comparator circuit 25, which is connected to the line 22.

For the output circuits 2-16 and 3-17 a saturable choke 26 is included between the secondary transformer windings 2 and 3 respectively and the corresponding diode 13, a power winding 27 of said choke carrying the circuit current and a control winding 28 of said choke being used to control the initial value which the magnetisation of the core material of the choke has each time the switch 1 2 closes. This control winding is connected to the output of a respective control unit 29, which has an input 30 to which is supplied data on a quantity which is to be applied in an adjusted and stabilized manner to the relevant output terminals. This quantity may be the output voltage, the output current or the output power. In Fig. 1 this is symboli cally represented by the connections 31.Although a sense-indicating marking is shown on the winding 27 there is in fact no mag netid coupling between the transformer 1 and the choke 26.

Fig. 2 shows a B-H curve for a possible magnetic core material of each saturable choke 26 of Fig. 1. This material has a low remanent magnetism, which means that the induction value Br at zero magnetic fieldstrength is small relative to the saturation induction Bs, which occurs at a field strength Hs. By means of the control winding 28 initial magnetisation or bias values B1 or B2 may for example be obtained, these corresponding to field strengths H1 and -H2 respectively.

The various bias values can be obtained by passing a steady current through control winding 28 from a controllable current source circuit included in the control unit 29 of Fig.

1 or they may be obtained by merely arranging that a current having a value such as to give rise to the correct field strength flows in the control winding only at each instant that the secondary windings 2 and 3 begin to supply voltage, i.e. when the switch 1 2 closes. It is alternatively possible to pass a constant current of a value such as to give rise to the saturation field strength through a separate winding and to employ the control winding 28 for reduction of the resulting saturation induction Bs to the required value, e.g. B1 or B2. As a further alternative the induced voltage across the control winding 28 when the switch 1 2 opens may be clamped by the control unit 29 at a value which is such that the resulting demagnetisation of the core material is to the value required, e.g. B1 or B2.The upward magnetization, caused by the secondary voltage substantially the whole of which appears across the winding 27 each time switch 12 is initially closed (provided that choke 26 is not then saturated), varies from the bias value to the saturation value, e.g. from B2 to Bs, from B1 to Bs, or from Br to Bs.

The induction law states that the voltage across the coil is equal to the rate of change of flux in the coil or that the integral of the coil voltage with respect to time is equal to the total flux change. In Fig. 2 an arrow 40 represents the induction variation Bs-B1, the arrow 41 the variation Bs-Br and the arrow 42 the variation Bs + B2, and, as is known, the flux variation is proportional to such induction variations, the constant of proportionality being n.A, where n is the number of turns which make up the coil and A is the cross-sectional area of the magnetic core material.During the upward magnetization the voltage term in the voltage-time integral corresponding to the particular flux variation which then occurs is the steady secondary rectangular voltage, so that saturation will occur a time after the secondary voltage initially appears which is proportional to the difference between the initial magnetisation and the saturation value. Thus the secondary voltage will be applied to the rectifier 1 3 after a delay which is proportional to said difference. Each time switch 1 2 is opened the magnetization may be returned to its initial or bias value by rapidly reducing the field by means of a current through a further control winding (not shown) so that a bias field, e.g.

H1 or H2 associated with the current through the control winding 28 remains, or by applying a voltage from a source to the control winding 28 of the coil when the switch 1 2 is open, so that for a part of the time when the switch 1 2 is open the field steadily decreases to the desired induction value, for example B1 or to approximately Br. In the latter case only, one control winding will be required and current will flow therein for only part of the time. It is evident that the latter method only enables the range of induction values between Br and Bs to be covered. In order to enable a larger range to be covered the B-H curve may be effectively shifted by applying an auxiliary field, such as H2, to the core material.

Fig. 3 shows a B-H curve of a rectangular loop magnetic material. This reveals that the value Br for the remanent magnetism of such a material is substantially equal to the saturation value Bs. In order to obtain a useful flux variation with such a material it will always be necessary to provide a reverse field during the open periods of the switch 1 2.

Fig. 4A and Fig. 4B show some time diagrams illustrating the operation of the converter of Fig. 1, the diagram Vsec representing the transformer secondary voltage. Fig. 4A relates to the case where demagnetisation of the choke 26 when switch 1 2 is open is obtained by then clamping the voltage across the control winding 28, and Fig. 4B relates to the case where current in the control winding 28 when switch 1 2 is open is maintained at the value which it has when switch 1 2 is closed, demagnetization then being obtained when switch 1 2 is open by means of a current in the power winding 27. At the instant t1 in Fig. 4A the switching element 1 2 in Fig. 1 is turned on (closed), so that the rectangular voltage pulse 43 is produced across the secondary windings. At the instant t4 the element 1 2 opens. For the purpose of demagnetization of the transformer 1 the diode 8 is turned on when element 1 2 opens and the supply voltage appears across winding 7. As a result of this the voltage pulse 46 appears across the secondary windings. Owing to the presence of the saturable choke which, provided it is not in the saturated state, represents a high impedance when switch 12 is initially closed, a significant part of the voltage Vsec is only applied to the rectifying circuit when the choke subsequently saturates, for example at the instant t2 or t3. In the diagram Vsec this is indicated by the lines 45 and 44 respectively.

The diagram Isec in Fig. 4A represents the current through the flywheel choke 14, for a constant load. The line 53 is obtained if the saturable choke 26 is in the saturated state immediately the switch 1 2 is closed. If the saturable choke only saturates after a time t2-t1, the current line 55 is obtained and if the choke only saturates after a time t3-t1 the current line 54 is obtained.

The diagram Isp in Fig. 4A represents the current through the power winding 27 of the saturable choke 26 for the last two of the above-mentioned cases, the current values between the instants t2-t4 and t3-t4 being equal to the corresponding currents shown in the diagram Isec. At the instant t4 the current Isp becomes zero and a demagnetizing current is produced in the control winding 28.

The various diagrams of Fig. 4B correspond to those of Fig. 4A, the instants t9, t10, t11 and t,2 indicated in Fig. 4B corresponding to the instants t1, t2, t3 and t4 respectively in Fig. 4A.

Because the magnetic bias is obtained in this case by current-source control from the control unit 29, demagnetization is possible only by means of a current in the power winding 27. At the instant t,2 the voltage Vsec is reversed. The flyback diode 20 is conducting and carries the currents 47 or 48 as indicated in the diagram Isec. The choke 26 now produces an induction voltage across the winding 27, so that the rectifying diode 1 3 is turned on and a demagnetizing current, designated 49 and 50 in diagram Isp, flows. While this occurs the voltage across winding 27 is substantially equal to the negative voltage Vsec (line 46).

The diagrams B of both Figures represent the variation of the induction or flux in the saturable choke.

During the time intervals t3-t4 or t"-t,2 the coil is saturated, i.e. the induction value is Bs.

The upward magnetization produced by the voltate Vsec takes place during the time interval t2-t, or t,o-tg if the initial magnetisation or bias value is B1 and during the time interval t2-t or t,,--t, if the initial magnetisation or bias value is B1 and during the time interval t3-t1 or t11-t9 if the initial bias or magnetisation value is B3.

Demagnetization by causing the control unit 29 to clamp the voltage across the control winding 28 at a fixed maximum value when said voltage has one sense (Fig. 4A) is effected during the time interval t,-t4, the final value B1 or B3 being reached at the instant t,.

Demagnetization of the core of the transformer 1 ceases at the instant tz because the voltage 46 then becomes zero, so that diode 1 3 is turned on and the voltage across the winding 27 also becomes zero. As a result of this the control voltage source in unit 29 no longer has any effect.The instantaneous field in the core of the choke 26 is now maintained by a current in the winding 27, as is indicated in the diagram Isp by the current values ib and id. When demagnetization is obtained by means of a current in the power winding 27 (Fig. 4B) the values B1 and B3 are reached at the instants t,3 and t,4, said values being maintained for the rest of the open period of switch 1 2 by means of the constant current produced in winding 28 by the control unit 29.

Figs. 5A and 5B show the initial current region between the instants t, and t3 and between the instants t9 It" respectively in the diagrams Isp of Figs. 4A and 4B respectively in detail. The straight rising portions of these curves arise because of the induction law (i = Vsec/L.t) where L is the self-induc- th tance of the winding 27 of choke 26 and t is time. It will be seen from the formula that the slopes of these portions are directly proportional to the voltage Vsec. The secondary rectangular voltage 43 appears at the instant t, or t9, substantially the whole of this voltage occurring across the power winding 27 of the choke 26.

Fig. 5A shows the variation of the current Isp for the instants tz, t1, t2 and t3. At the instant t7 at which the voltage Vsec returns to zero the current Isp through winding 27 assumes the value it or id which is associated with the induction value B1 or B3 via the field-strength and the B-H curve. At the instant t, the current increases by an amount ih corresponding to the change in field strength needed to traverse the flat portion of the B-H curve of Figs. 2 and 3 between the points x and y which are shown on the assumption that the initial value of the induction is B1.

The current then varies in accordance with the line 59 or 58, or in accordance with the line 60 or 61 if the input voltage and thus the secondary voltage is higher.

As is known the relationship between the field strength H and current i may be represented in simplified form by H.l = n.i, where 1 is the length of the magnetic circuit and n the number of turns of the winding through which the current iflows. The current eventually reaches a value is which gives rise to the field strength Hs corresponding to the saturation induction Bs. From this point in the diagram the induction only changes slightly, but the field strength and the current through the winding 27 increases appreciably. In Fig. 5A this is indicated by the arrows 56 and 57.

It should be noted that the coil 27 in the secondary circuit represents a high impedance when the current through it is less than is and represents a low impedance relative to the other electrical circuit parameters when the current through it is greater than is For the sake of clarity the initial current is shown enlarged in the diagram lsp of Figs. 4A and 4B relative to the current represented by the lines 54 and 55.

If output voltage stabilisation is desired for the secondary circuit and the associated con tro31unit is constructed accordingly, it can be seen from Fig. 4A and Figs. 5A how such stabilisation is obtained. If the input voltage and thus the secondary voltage should in crease the line 59 is replaced by the line 60 and the choke saturates at the instant tx instead of the instant t2. The control unit detects that the converter output voltage is increasing and consequently clamps the (re verse) voltage across the control winding 28 at a higher value during the next off period of the switch 1 2. This results in increased de magnetization in the next period t2-t4 and, in Fig. 4A and in Fig. 5A, the upper curves are replaced by the lower ones.The result is that, referring to Fig. 5A, saturation is reached at the next instant ty via the line 61. As ty-t1 is greater than t2-t, the current in the flywheel choke 14 now has a smaller part of the remainder of the period in which switch 1 2 is closed in which to build up, this providing compensation for the higher voltage.

Fig. 5B shows the initial current region of the diagram Isp of Fig. 4B. A field corre sponding for example to the induction value B, or B3, already exists at the instant t9 because of the steady current through the control winding 28. From this instant the current Isp varies in accordance with the line 51 or, at a higher secondary voltage, in accordance with line 52, c.f. the line pairs 59 and 60 or 58 and 61 respectively in Fig. 5A.

Depending on the value of the field originally present and on the value of the secondary voltage saturation of the core of choke 26 occurs at the instant tx, tio ty or t11. Current ia is equal to I/n. (Hs-H1) and the current ic is equal to 1 /n (Hs-H3), where H 1 and H3 are the field strengths corresponding to the induc tion values B1 and B3 respectively, and Hs is the field strength corresponding to the satura tion induction value.If the output voltage is stabilised in a similar manner to that de scribed with reference to Fig. 5A, the control unit will reduce the magnetic bias from for example B1 to B3 if the secondary voltage and hence the output voltage increases (by reducing the control current through the wind ing 28) so that instead of saturation occurring with a power winding current ia (at the instant tx) it will occur at a power winding current i0 (at the instant ty).

Fig. 6 shows a possible construction for part of the control unit 29 if the output voltage of the corresponding secondary circuit is taken as being the quantity to be con trolled. For this purpose a connection 31 is made between input 30 of the control circuit and output terminal 32 of the secondary circuit and the common point of the control circuit is connected to terminal 33 of the secondary circuit. The non-inverting input 63 of an operational amplifier 62 is connected to a voltage divider 64, 65. The inverting input 66 of the amplifier is connected to a zener diode 67, which is biased by means of a resistor 68. The amplifier and the zener diode are energized with a supply voltage which may, for example, be obtained by rectification of Vsec across the winding 2, or in the manner shown in Fig. 6.At the output 69-70 of the amplifier 62 a voltage is obtained which is proportional to the difference between the zener reference voltage and the output voltage across terminals 32 and 33 which has been attenuated by the divider 64, 65. The amplifier output is coupled to the control winding 28 in one of the ways which will be described with reference to Figs. 8, 9 and 10, or in the way shown for the similar amplifier 62 in Fig. 7.

Fig. 7 shows a possible construction for the control unit 29 if the output current of the corresponding secondary circuit is taken as being the quantity to be controlled and the saturable choke 26 is controlled by means of a steady current through control winding 28.

Said output current is passed through a sensing resistor 71 via connection 31 and input 30. The voltage across said resistor is compared with the zener voltage of a zener diode 72. By means of identical dividers 73, 74 the voltages are respectively applied to input 63 and input 66 of operational amplifier 62.

Output 69-70 of the amplifier supplies the amplified difference voltage to the control input of a controllable current source circuit, comprising a transistor 75 and an emitter resistor 76. The transistor collector current is passed through the relevant control winding 28 so as to obtain a bias.

The circuits of Figs. 8, 9 and 10 show how the output voltage of the amplifier 62 in Fig.

6 or Fig. 7 itself may be used to clamp the voltage across the control winding during the off periods of the switch 12, in order to obtain magnetic bias. The control winding 28 is connected in series with a diode 77, which may be the base emitter junction of a transistor 78. The series combination is connected to the output of the amplifier 62 in such a way that the demagnetizing current 79 which flows when switch 1 2 is open flows through the diode 77 or the collector-emitter path of the transistor, the induction voltage across the winding being clamped at a value equal to the control voltage across the terminals 69, 70 plus the diode voltage drop.The demagnetisation (Bs-B) obtained satisfies the equation Vst.t = ns.A (Bs-B), where Vst is the value at which the voltage across the control winding 28 is clamped, Q is the number of turns of the control winding, and t is the time for which the demagnetising current flows. Fig.

4A, diagram B shows the resulting induction variation in the time interval between t4 and tz.

The field strength H and the associated cur rent ivary in substantially the same way, assuming that the relative permeability of the core material of the choke is virtually constant. Magnetization and demagnetization quantities can be given a suitable relationship to each other by suitably choosing the turns ratio between the power winding and the control winding.

In Figs. 8, 9 and 10 the output of the operational amplifier (terminals 69 and 70) is shown floating. However, if desired, terminal 70 may be connected to the common terminal 33 or terminal 69 may be connected to the supply line terminal 32.

Fig. 11 shows a detailed possible construction for one secondary output circuit of Fig. 1, the control unit 29 being constructed in a similar manner to that described with reference to Fig. 6 and Fig. 8 or 9. Output terminal 70 of amplifier 62 (not shown in Fig.

11) is in fact connected to the common terminal 33 and output terminal 69 supplies a control voltage to an emitter-follower 80 having an emitter resistor 81. Diode 77 may be dispensed with if the maximum allowable base-emitter reverse voltage of transistor 80 is sufficiently high. The control voltage for winding 28 is effectively Vout-V69, where Vout represents the output voltage across terminals 32 and 33 and V69 the output voltage of the amplifier 62. As the voltage V69 has a negative sign in the subtraction just mentioned, the functions of input terminals 63 and 66 have been reversed in comparison with Fig. 6.

If, for example, the voltage across terminals 32 and 33 tends to increase, (because the supply voltage to the converter increases or because the load current is reduced) the voltage on input 66 increases, the difference voltage between inputs 63 and 66 decreases and the voltage on terminal 69 decreases.

During the off-phase of the converter period, (switch 1 2 open), the voltage across winding 28 will consequently be higher. This means that in this off-phase a faster demagnetization takes place, resulting in a lower induction value being reached, so that during the next on-phase (switch 1 2 closed) the delay produced by the choke 26 is larger even if the transformer secondary voltage has increased, because now a greater induction or flux change has to occur before the choke 26 saturates. The effective on-time consequently decreases, i.e. the transformer supplies power for a shorter time, and this counteracts the original voltage increase. The compensation obtained depends on the control loop gain, which is mainly determined by the gain factor of the amplifier 62. Control is very fast, because the magnetic bias change has its effect during the next period of the control oscillator to that in which the output voltage change occurred.

Claims (7)

1. A switched mode power supply circuit arrangement of the forward converter type, comprising a series combination of a controllable switching element and an input winding of a transformer, which series combination is connected across a pair of direct-voltage input terminals, first and second output circuits each having output terminals for connection to a load, and a first control circuit having its output coupled to a control input of said switching element so that said control circuit will render said switching element periodically and alternatively conductive and non-conductive in operation, each said output circuit comprising an output winding of the transformer, a rectifying diode, and a flywheel circuit comprising a diode and a choke, the first said output circuit being provided with means constructed to adjust and stabilise an electrical quantity at the output terminals of said first output circuit, characterised in that the adjusting and stabilising means comprise a saturable second choke having a power winding and a control winding, and a second control circuit having an input for said electrical quantity and an output connected to said control winding, said power winding being connected in series with the output winding and the rectifying diode of said first output circuit and said second control circuit being constructed to apply a control signal to said control winding so as to control, in accordance with the value of said electrical quantity, the initial value which the magnetisation of the core material of said second choke has each time said switching element closes.

2. An arrangement as claimed in Claim 1, characterised in that said material has a remanence which is low relative to the saturation magnetisation of said material.

3. An arrangement as claimed in Claim 1 or Claim 2, characterised in that the second control circuit comprises a controllable voltage source a control input of which constitutes the input for the electrical quantity and the output of which is connected across said control winding via a series diode in such manner that when said switching element is closed said series diode will be reverse-biassed by the output voltage of said voltage source and when said switching element opens said series diode will become forward-biassed by the voltage then induced across the control winding.

4. An arrangement as claimed in Claim 3, characterised in that the controllable voltage source comprises an amplifier having inverting and non-inverting inputs, a reference voltage source having its output connected to one input of the amplifier, and means for supplying a signal representative of said electrical quantity to the other input of the amplifier.

5. An arrangement as claimed in Claim 3 or Claim 4, characterised in that said series diode is constituted by the base-emitter junc tion of a transistor, the emitter and collector of said transistor being connected to respec tive-ends of said control winding.

6. An arrangement as claimed in any preceeding claim, characterised in that the first control circuit is a pulse-width-controllable pulse train generator, and in that the arrangement includes a comparator to a first input of which is coupled the output of said second output circuit, to a second input of which is coupled the output of a reference voltage source, and the output of which is coupled to a pulse width control input of said generator so as to adjust the width of the output pulses of said generator in such a sense as to maintain the voltage at the output of the second output circuit substantially constant.

7. A switched mode power supply circuit arrangement substantially is described herein with reference to Figs. 1, 6 and 8, 9 or 10 of the drawings, to Figs. 1 and 7 of the draw ings, or to Figs. 1 and 11 of the drawings.