WO1987002543A1

WO1987002543A1 - A method for pulse width regulation

Info

Publication number: WO1987002543A1
Application number: PCT/SE1986/000453
Authority: WO
Inventors: Holger Norlyk
Original assignee: Norlyk H
Priority date: 1985-10-17
Filing date: 1986-10-07
Publication date: 1987-04-23
Also published as: SE8504859L; SE455649B; SE8504859D0; DK307487D0; DK307487A

Abstract

A method of regulating the power effect in an electric load by pulse width regulation includes the steps of conducting a loading current, driving a load, through a transformer, and permitting such current to bring the transformer to a condition of magnetic saturation. The loading current is further conducted through a control circuit which receives a control current for switching the control circuit to the conductive, or live state from the transformer when the transformer is in the non-magnetically-saturated state. An external signal current of a certain duration is conducted through the transformer so that the control current is generated and the control circuit is switched to the conductive state at a first point in time. The signal current and loading current are conducted through the transformer until magnetic saturation occurs after the passage of a certain time interval, this time interval being shorter than the duration of the signal current. The signal current is of a magnitude which is sufficient to maintain the magnetically saturated state. For governing the power effect in the load, the relationship between the duration of the signal current and the time interval is modified.

Description

A METHOD FOR PULSE WIDTH REGULATION

TECHNICAL FIELD

The present invention relates to a method, by pulse width regu¬ lation, of governing the power effect in an electric load, the me¬ thod including the steps of permitting a loading current driving the load to bring^'a transformer to magnetic saturation, the loading 5 current being passed through a control circuit, and a control cur¬ rent for switching the control circuit to the conductive - or live - state is taken from the transformer in the non-magnetic saturated state thereof. I-ACXGRCUND ART

10 Swedish Patent Specification No. 7512624-3 illustrates and dis¬ closes a transistor circuit for making and breaking functions. This circuit is intended for operating different power consumers, for ex¬ ample fluorescent lighting tubes, and has been constructed so as to provide a high degree of opertational efficiency. According to this

15 Patent Specification, the current driving the load is passed through a winding on a transformer and thence further through a transistor which receives its operative current in the direction of conduction from another winding on the same transformer. Hereby, the operative current impressed on the transistor will be proportional to the

20 loading current in that the load, the transformer and other com¬ ponents are dimensioned in such a manner that the core of the trans¬ former is magnetically saturated when the loading current has at¬ tained a certain level. The control current of the transistor will will hereby disappear extremely rapidly, whereupon disruption of the loading ^'current takes place.

The apparatus according to the above-mentioned Patent Specifi¬ cation is of duplicate design in such a manner as to impress an ^* a.c. voltage to the load, in that one half of the circuit conducts ^■ the loading current when the other half is inactive, and vice versa. The transistor-circuit according to the above-mentioned Patent Specification displays many advantages, but does not generally per¬ mit that the power effect transmitted to the load can be controlled or regulated, apart from in the exceptional case when the power ef¬ fect generation is directly dependent on frequency. OBJECT OF THE INVENTION

The object of the present invention is to realise a method, by pulse width regulation, of varying the power effect in an electric load within a wide range. The invention also has for its object to design the method in such a manner that the requisite current may throughout be as small as possible without negatively affecting the operative degree of efficiency. Finally, a further object of the present invention is to realise such a method as attains feedback of energy from the load to the power source in the event of reactive components in the electric load, so that transients are thereby av¬ oided. SOLUTION

The above-outlined objects of the present invention are at- tained if the method^•intimated by way of introduction is character¬ ised in that a signal current of a certain duration is conducted through the transformer, so that the control current is generated ^"and the control circuit is switched to the conductive, or live, state at a first ti epoint; that the signal current and the loading current are conducted through the transforme^*- until magnetic satur¬ ation occurs after a time interval at a second timepoint, the time interval being shorter than the duration of the signal current, and this being of an order of magnitude for maintaining the magnetically saturated state; and that the relationship between the duration of the signal current and the time interval is modified for regulating the power effect. In one advantageous embodiment of the present invention, it is suitably applicable that the time interval is modified in that an outer magnetic field is superposed on the magnetic field caused in the transformer by the loading current and the signal current. 5 If, for example,-_eonstant power effect is desired in the elec¬ tric load, this may readily be attained if the present invention is further characterised in that a voltage across the load is sensed and that the outer magnetic field is rendered proportional to this voltage. 10 In one alternative embodiment of the present invention, it fur¬ ther applies that the time elapse interval is modified by a modi¬ fication of the magnitude of the^' signal current.

In yet a further embodiment of the present invention, it ap¬ plies that the duration of the signal current is modified while the 15 time interval is kept constant.

BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWINGS

The. nature of the₍ present invention, and its aspects will be more readily understood from the following brief description of the accompanying Drawings, and discussion relating thereto. 20 In the accompanying Drawings:

Fig. 1 schematically illustrates a first embodiment of a circuit for implementing the present invention; Fig. 2 is a voltage-time diagram for certain of the compo¬ nents included in Fig. 1; . 25 Fig. 3 shows an arrangement between an outer magnetic core and a transformer core included in the ernbodiment of Fig. 1; and Fig.4 Illustrates an alternative embodiment of a circuit for implementing the present invention. 30 - •^•

DI-SCR-OPTION OF PREFERRED EMBODIMENT

Referring to the Drawings, the circuit shown in Fig. 1 may be used for power supply in general, the driven load being impressed with an a.c. voltage in the form of a square wave whose real effect 35 is varied by pulse-width regulation. If desired, the square wave may naturally be readily modified to sinal form by the employment of an LC Coupling. As one particularly advantageous field of application for the circuit, mention might be made of the driving of, for in¬ stance, fluorescent tube lighting, primarily in so-called emergency lighting, but, of course, also of the driving of other loads with both reactive components and considerable inertia, such as electric .motors.

In Fig. 1, the reference numeral TRi relates to a transformer with a core K^, a secondary winding L_ and two primary windings lη and L8. The load L proper is connected across the secondary winding Lg. Energy to this load is supplied to the circuit by the inter- mediary of the connections +a and -a.

A resistor 3 is further connected across the load L, and also a rectifier bridge LB whose output is connected to a light emitting diode LD.

The circuit according to Fig. 1 further includes a control cir- cuit whose major components are, on the one hand, a transformer -I-V? with the core K2 and windings ]_, L2, L3 and L4, and, on the other hand, the two transistors Ti and T2. ^"The transformer TR2 preferably has its core K2 designed as an annular core and is of the so-called saturable type,,, which implies that the core K2 moves to magnetic saturation on normal currents occurring in any one of the windings. Both of the primary windings L7 and s of the transformer TR]_ are connected in series with the two windings L3 and L4, respec¬ tively on the transformer TR2. The opposite ends of the windings L3 and L4 are further connected, via diodes, to the collector on the two transistors Tι_ and T2, respectively, and the emitters of these two transistors are connected to a common lead in communication with the connection -a. It will be appreciated that a loading current,

-depending upon whether the transistors T]_ and 2 are in the live or dead state, may flow from the connection ^"+a through either one of the windings L7 and LQ , either one of the windings L3 and L4 and further through either of the transistors Tτ_ and 2 to the con¬ nection -a.

The two windings Li and L2 on the transformer TR2 have a common centre point, which, via an adjustable resistor R, is connected to the conductor coupled to the connection -a. The opposite ends of the windings L]_ and L2 are in communication with the bases of the two transistors T]_ and T2, and also further by the intermediary of the resistors R]_ and R2_/ respectively, with the .collector on two trans¬ istors Ta and Tb, respectively. The emitter of these two transistors is in communication with a conductor, with the connection -b, while the connector +b coincides with the connector -a. The bases of the two transistors Ta abnd Tb are connected to a pulse generator which triggers the entire circuit.

The voltages, ^uBE _a and ^uBE_tb which are impressed between the bases of the transistors Ta and Tb and the common connection -b have a time duration which is apparent from Fig. 2. The device which em- its these voltages does not form part of the present invention, but may be a flip-flop, a crystal-controlled oscillator, the mains vol¬ tage etc.

The circuit illustrated in Fig. 1 functions as follows. At the time point t^, the voltage ^uBEta is impressed between base and emit- ter on the transistor Ta, with the result that this is switched to the conductive, or live state. Hereupon, a current will flow from the connection +b via the variable resistor R, the winding L_, the resistor R]_, the transistor Ta and back to the connection -b. The voltage across the winding L]_, as is apparent from Fig. 2 and as al- so lies across base-emitter on the transistor T, is directed in a such a way as to hold the_. transistor i in the blocked state so that no current can flow from collector to emitter.

The current flowing though the winding Li gives rise to a mag¬ netic field in the core K2 the magnetic field in the winding L2_/in- ducing a voltage which forces a current in the conduction direction of the transistor T2 via base-emitter therein. The transistor T2 will thereby be switched to the conductive or live state, with the result that a loading current begins to flow from +a via the winding Lg, the winding L4, the transistor T2 and back to the connection -a. The current flowing through the winding Lg gives rise, in the secondary winding Lg, to a current which, in its turn gives a cur¬ rent through the load L. The above-described sequence of events has occurred in principle instantaneously at the time point t_.

"When current begins to flow through the winding L4, the mag- netic field in the core K2 increases, such that the core K2 duly at¬ tains the magnetically saturated state. This occurs after the time interval Δt or at the time point t2 and has as a consequence that ,.control current in the forward conduction direction of the trans¬ istor T2 can no longer be induced in the winding L2. Thereby, the transistor T2 switches,-at_the time point 2_/ to the non-conductive, or dead state, which has as a consequence that the loading current 5 from the connection +a ceases to flow.

The current flowing through the winding L]_ and the transistor Ta is still on and is dimensioned such •that the. core K2 is still held in the saturated-state. If the load includes reactive com¬ ponents, a feedback of energy will also take place during this per-

10 iod of time. This feedback is directed such that the transistors T]_ and 2 are held blocked, in that the core K2 is already held in the saturated state by the feedback current, for which reason this blocked state will be maintained even if the current through the transistor Ta were to disappear. In extreme cases, for example if

15 the load L is a motor of high mechanical inertia, the feedback may, thus, take place during such lengthy periods of time that the trans¬ istors Ta and Tb will have time to switch (see further below) one or more times. The conditions are wholly analoguous when the transistor Tb has given the impulse.

20 When the voltage ^uBEt_a fades after one half period of the square wave at time point t3, the transistor Tb will be instead im¬ pressed with the voltage ^uBEy- between base and emitter so that this transistor is switched to the conductive, or iive state. This re¬ sults in a current beginning to flow from +b via the resistor R, the

25 winding 2,- the resistor R2,- the transistor Tb and back to the con¬ nection -b. In this instance, the voltage across the winding L2 will be counter-directed to the control current which would be required to switch the transistor T2 to the conductive state, for which rea¬ son this transistor is maintained in the non-conductive, or dead 0 state. On the other hand, the current through the winding L2 gives rise to magnetic field in the core K2 and this magnetic field generates, in its turn (assuming that the feedback from the load L has ceased) a control current in the winding L]_ in the conduction direction of the transistor T]_, such that this is switched to the 5 conductive, or live state. This switching of the function of the transistors Tj_ and T2 occurs at time point t3. The fact that the transistor T]_ now begins, at time point .3, to be conductive entails that current will flow from the connection +a through the winding L7, the winding L3,- the transistor T]_ and back to the connection -a. Here, current will be taken out via the transformer TRi to the load L and the loading current will flow through the winding L3 so that the magnetic field in the core K2 will thereby increase in proportion with the loading current. After a certain time interval, Δ-t, the core K2 will once again have at¬ tained magnetic saturation, which occurs at time point t4. This en- tails that control current can no longer be generated in the winding L]_, for which reason the transistor T]_ will once again be switched to the non-conductive, or dead state and the loading current between the connections +a and -a will disappear. During the remainder of the square wave ^uBEt_D, the current flowing through the transistor Tb^' and the winding L2 will maintain the core K2 in the saturated state, for which reason no loading current can flow through ^"the circuit. That time during which this takes place corresponds to the time dif¬ ference ts - 4. Feedback of energy stored in the load can also take place here. The point in time when, in most cases, feedback of energy from the load L ceases falls within the interval t2 to t3, t4 - ts and so on. However, as has been intimated above, there is nothing to pre¬ vent feedback from continuing for longer periods so that the tran¬ sistors Ta or Tb have time to switch during the feedback time. If this occurs, the transistors Ta and Tb will have no temporal control over the cycle as long as feedback takes place, even though their switchings are progressing and are controlled by the voltage source which emits the voltages ^uBΞt_a and ^uBEy-. On the other hand, when feeback has ceased, the transistors Ta and Tb will immediately re- assume their controlling function.

When the transis__r T]_ is in the conductive, or live state, the current will flow, as has been mentioned above, from +a through the windings L7 and L3 and saturation will occur in the core K2, whereupon the current from the connection +a will disappear and feedback from the load L may possibly take effect. If the e.m.f.n. caused by the energy stored in the winding Lg is greater than the voltage between +a and -a, feedback will take place and current will flow from the winding L3 to +a, to -a, through the diode D2_/ through

- ^' the winding L4 and back to the winding Lg. It should here be noted that the .direction of current flow through the winding L4 is the same as when the transistor-T]_ was conductive prior to saturation of the core K2.

To prevent current from flowing, on feedback, from -a, through the resistor R, through the winding L2_/ through the transistor T2, through the winding L4 and to the winding LQ , use is made of the di¬ ode D4 (in the transistor _;^•_, the diode D3). In the second half-period, i.e. when the transistor T2 is con¬ ductive and breaks on saturation, the conditions are wholly analo- ^"guous and, for example, it is the winding L7 which realises the feedback.

Once the time point _5 has been reached, voltage will once again be impressed on the base-emitter of the transistor Ta, where¬ after the sequence is repeated.

According to the invention, it is possible to vary the timeΔ , in othder words the time taken to attain magnetic saturation in the core K2, for. example by variation of the adjustable resistor R. Since the magnetic flux which the core K2 can conduct at maxi¬ mum depends upon the material properties of the core, its size and configuration, it will readily be perceived that if an- outer mag¬ netic field is superposed on the magnetic field generated in the core K2 when current flows through the transformer TR?, the time in- terval At may thereby also be affected. If, thus, a magnetic field is superposed on the core K2 in the same direction as the magnetic field which strives to saturate the core, the point of saturation will be reached earlier, for which reasonΔt will be shorter. If, in

-the opposite case, the outer magnetic field is counter-directed, Δt will, naturally, be longer.^~

If a constant voltage is to be maintained across the load L, this may be realised in that a rectifier device, preferably a recti¬ fier bridge, is coupled across the load in series with the resistor R3. The output from this rectifier device LB drives a light emitting diode LD which is optically coupled to a photoresistor which, with its one end, is connected to the connection +b and, with its other end, is connected to the base of the transistor T3. Further, the base is connected, via the resistor R4, to the connection -b. The winding L5 is wound on a core K3 which is magnetically coupled to the core K2- ^In this instance, the practical embodiment may be as shown in Fig. 3, in which the core K3 is of U-shaped configuration, while the core K2 is an annulus, as intimated above. The gaps be¬ tween the annular core K2 and the U- or C-shaped core K3 are as nar¬ row as possible so that no air gap may, to any appreciable degree, affect the magnetic flux.

If the voltage across the load L tends to vary, the output vol- tage from the rectifier device LB will vary proportionately and the light flow emitted by the light emitting diode LD will fluctuate at the same rate. This entails that the photoresistor will realise a more or less powerful current via base-emitter on the transistor T3 so that the conductive capacity of this transistor will increase when the light flow emitted from the light emitting diode LD in¬ creases as a result of an increasing voltage across the load L. Na¬ turally, in this instance, the current flowing through the tran¬ sistor T3 and the winding L5 discharged via the connections +b and -b will vary at the same rate, and will give rise to a corres- pondingly variable magnetic field in the core K3. An increase of the voltage across the load L will, hence, give rise to an increased magnetic field in the core K3, this magnetic field being superposed on the magnetic field in the core 2. The result of this will be, as has been intimated above, that the time elapsed to attain magnetic saturation in the core K2 will reduce, in other words the time inte¬ rval t will be reduced, for which reason the pulse width of the loading current will reduce through the transformer TR]_.

The arrangement of the cores K2 and K3 illustrated in Fig. 3 enjoys the advantage that it. is not dependant upon the direction of the magnetic field in the core K3, since it will be sufficient with an increase of the magnetic field in the core K2 in its G.._.half in order that the "highest magnetically loaded" half of the core reach saturation, for which reason fluctuations in the magnetic field can no longer take place and, thus, neither can the transformer transmit any current. DESCRIPTICN OF AL.3__3NATIVE EM-3COIME -.S

Fig. 4 illustrates an alternative circuit arrangement for car¬ rying out the method^{^}according to the present invention. The circuit according to Fig. 4 includes a transformer with the core K2and the six windings Lχχ₇ Lχ2, ^13, I_.4_/ as well as L]_S and L2S. The circuit further includes four transistors, namely Tχχ₇ χ2, 13 and χ4. As in the above-disclosed embodiment, the load is referred to by re¬ ference numeral L and the connections for the current source driving the load L are the connections +a and -a. Moreover, there is one further connection pair, namely +b and -b for the supply of current for the control circuit proper, the connections -a and +b being com¬ mon. The circuit further includes the two transistors Ta and Tb, which, as has been described above, are connected to some form of oscillator which emits a square wave according to Fig. 2. Otherwise, the circuit is designed and connected in the manner as will be ap¬ parent from Figs. 1 and 4.

If, at time point t]_, the transistor Ta ^'is impressed with the voltage ^uBEt_a between base-emitter, a current will flow from +b, through the winding Lχι, the resistor Rπ, the transistor Ta and back to the connection -b. In this instance, the voltage across the winding LT is directed in such a manner that the transistor Tχι is blocked, in other words no current can flow from collector to emit¬ ter.

The current flowing through the winding Lτj_ gives rise to a magnetic field in the core K2, which generates currents in both of the windings Lχ2 and Lχ4. These currents are directed in such a man¬ ner that the transistors 2 ^an(3 ^τ14 respectively, are switched to conductive, or live state. This entails that a current can begin to flow (at time point _χ) from the connection +a, through the trans- istor Tχ4 (via collector-emitter), the winding LχS_; the load L, the winding L2S, the transistor T 2 (via collector-emitter) and back to the connection -a. The current through the two windings ≤ and L2S increases the magnetic field in the core K2 so that this, at time point t2, in other words after the interval Δt, attains the magne- tically saturated state. This further entails that current can no longer be generated in the windings L 2 and Lχ4, for which reason the transistors χ2 and T 4, respectively, are switched to the dead state and thereby the current through the load L is broken. However, the voltage ^uBEt_a remains across base-emitter on the transistor Ta so that, thereby, current from the time point t2 to the time point t3 may still flow through the winding xx and the core K2 will here- by also be maintained in the magnetically saturated state.

During the time when the current flowed through the load L, an energy (in reactive loads) was built up therein which gives rise to an e.m.f. which strives to return the energy to the driving cir¬ cuit. If, in this instance, e.m.f.n. is greater than the voltage be- tween +a and -a, the e.m.f. of the load will force a current through L2S, through the diode D5 and to +a, and from -a through the diode D5, through the winding ≤ and back to the load.

In energy feedback during the second half-period, the current flows in the opposite direction through the load L, and the windings L ≤ and L2S and further through the diodes Dχo and θχχ. The diodes D7, D8 and D9 are intended to prevent current in the base-collector transition in the transistors Tχχ_r Tχ2, Tχ3, and Tχ4.

At time point t3, the voltage ^uBE _a ceases, irrespective of wh¬ ether any possible energy feedback from the load has ceased or not, and instead the voltage ^uBEy_ is impressed on the transistor Tb, so that this becomes live and conductive. Here, a current will flow from the connection +b through the winding Lχ2, the resistor Rχ2, the transistor Tb and back to the connection -b. The voltage across the winding Lχ2 is directed such that the transistor T 2 is held in the blocked state, i.e. no current can flow from collector to emit¬ ter.

The current flowing through the winding Lχ2 strives to create a magnetic field in the core K2, which, on condition that the energy feedback from the load L has had time to cease and, therewith, the saturated state in the core K2, in its turn generates current in the windings xx and Lχ3 so that the two transistors T x and χ3, re¬ spectively, are switched to the conductive state. Hereupon, a load¬ ing current will flow from the connection +a, through the transistor χ3, the winding 2S, the load L, the winding LχS_r the transistor Txx and back to the connection -a. The current flowing through the two windings Lχ≤ and L2S increases the magnetic field in the core K2 so that this, after a time interval Δt, and at the time point T4, arrives at the magnetically saturated state, which has as a result that the control current in the conduction direction of both of the transistors T^*jT and T^*|3 disappears. The current still flowing through the winding L12 is here sufficient to maintain the core K2 in the magnetically saturated state, such that, thereby, the two transistors T-^*** and T13 are,still held blocked and, as a result, no current can flow through the load L during the time t up to ts, when the voltage. ^uBE _a is once again impressed on base-emitter of the transistor Ta and the sequence is repeated. Correspondingly to that described with reference to Fig. 1, it is possible to couple, across the load L, a circuit which senses the voltage and which transmits this sensed voltage to a transistor cir¬ cuit corresponding to the circuit T3, L5, 3 and R4 in Fig. 1. The function of such a circuit and its magnetic coupling to the core K2 correspond wholly to that described above.

In the above-described circuit arrangements, it has been pre¬ supposed that the period lengths of the voltages ^uBE a -"d ^uBEt_ were to be held constant. This entails a simple solution, but is, of course, not the sole solution, it being conceivable that the time interval t instead be held constant and that the period length of the above-mentioned voltages be varied so that the relationship be¬ tween t and the period 'length will thereby vary, with a con¬ sequential power effect regulation in the load L.

The present invention should not be considered as restricted to that described above and shown on the Drawings, many modifications being conceivable without departing from the spirit and scope of the appended Claims.

Claims

13CLALMS

1. A method of .regulating the power effect in an electric load, comprising the steps of permitting a loading current driving the load to bring a transformer (TR2) to magnetic saturation, and pas¬ sing said loading current through a control circuit (Tχ_r T2; Tχ T 2, Tχ3, Tχ4), and taking a control current for switching the con¬ trol circuit to the conductive, or live, state from the transformer in the non-magnetically-saturated state thereof, characterised in that a signal current of a certain duration (t3 - x) is passed through the transformer (TR2) such that the control current is gen- erated and the control circuit (Tχ 2; Iχχ, Tχ2, Tχ3, χ4), is switched to the conductive state at a first time point (tχ),^« that - the signal current and the loading current are led through the tran¬ sformer until such time as magnetic saturation occurs after a time interval (Δt) at a second time point (t2), the time interval being shorter than the duration of the signal current and this being given a magnitude for maintaining the magnetically saturated state; and that the relationship between the duration of the signal current and the time interval is modified for regulation of the power effect in the load L.

2. The method as claimed in claim 1, characterised in that en¬ ergy stored in the load is fed back when the transformer is in the magnetically saturated state; and that the feedback current is di¬ rected for maintaining the transformer in the saturated state.

3. The method as claimed in claim 1 or 2, characterised in that the time interval is modified by the superposition of an outer magnetic field on the magnetic field generated in the transformer (TR2) by the loading current and the signal current.

4. The method as claimed in claim 3, characterised in that a voltage across th load (L) is sensed; and that the outer magnetic field is rendered proportional to this voltage.

5. The method as claimed in any one of claims 1 to 3, charac¬ terised in that the time interval (Δt) is modified by a modification of the magnitude of the signal current.

6. The method as claimed in any one of claims 1 to 5, charac- terised in that the duration of the signal current is held constant.