DE19635355C2 - Circuit for controlling clocked energy converters with a switching frequency that varies depending on the network - Google Patents

Abstract

A clocked energy converter (2), which supplies a consumer (3) from an AC voltage source (1), is controlled by a tax rate (4). The clock frequency is specified by an oscillator (5) integrated in the control unit (4), the frequency of which depends on an input variable. The voltage of the source (1) is fed to the input of the oscillator () via an electrical network (6), which causes a phase shift between the voltage at its input (7) and the voltage or the current at its output (8). 5), so that there is a dependence of the instantaneous switching frequency of the energy converter (2) on the instantaneous value of the output voltage of the source (1) at any moment and the maximum switching frequency change is approximately reached in the maximum of the input current of the energy converter (2) and maximum and The minimum value of the switching frequency occurs approximately at the minimum of the current at the input of the energy converter.

Description

The invention relates to a circuit arrangement for controlling clocked energy converters, for example switching power supplies, self-commutated converters and lamp ballasts, according to the preamble of claim 1. Such is known from US 5,146,398.

An oscillator is required to generate the switching frequency in such energy converters outputs a signal of the switching frequency (or a multiple thereof). As for example in "Unitrode Product & Applications Handbook 1995-1996 "and others in Application Note U-93 (S 10-9 ff.) described, a pulse width modulator is usually used, in which the output signal a controlled system is compared with a switching frequency sawtooth or triangle voltage. This frequency-constant delta voltage is generated, for example, by a capacitor an (adjustable) current source is charged up to an upper threshold; when reaching the upper threshold, another, stronger current source is activated, which discharges the capacitor, until a lower threshold is reached, at which the second-mentioned power source is deactivated again.

The disadvantage of this solution is that the spectrum of the Radio interference voltages at the sound frequency and their integer multiples high peaks which must be filtered to comply with legal EMC rules.

In "An AC-DC converter with low input distortion and near unity power factor" by M. J. Willers u. a. (European Power Electronics Conference Proceedings '93 Vol. 4. p. 1 ff) is u. a. a solution described to reduce the radio interference voltages overall by modulating the switching frequency. The modulation takes place here by means of an additional oscillator, which in each case at the zero crossing of the Mains voltage is set to a defined value. There is both a triangular course of the Switching frequency (with reversal points at the current minimum and at the current maximum) as well as sawtooth-shaped course (with resetting in each case at the zero voltage crossing), where better results were obtained with a sawtooth-shaped course (presumably because here in contrast each current / switching frequency combination occurs only once per half-wave for the triangular course). There is no further influence on the switching frequency by the mains voltage.

A similar arrangement is also in DE 35 41 307 C1 "circuit arrangement for generating a DC voltage from a sinusoidal input voltage " Modulation by means of an additional triangular oscillator, each at the zero crossing of the Mains voltage is triggered. This additional oscillator, which is initially a grid-synchronous Triangle voltage generated here consists of the zero voltage detector, a monoflop and one Integrator. The output voltage of the integrator is the switching frequency determining element,  a VCO, which generates a frequency proportional to the input voltage. Between Two zero crossings of the mains voltage is the curve of the triangular voltage at the output of the Integrators and thus also the switching frequency regardless of the course of the mains voltage.

In the document DE 35 41 307 C1 it is also stated that the steepness of the Switching frequency in the range of the minimum switching frequency (which in the solution presented here in Current maximum occurs) is advantageous because here to achieve a certain damping factor particularly large filter elements are required (It should be noted, however, that this is caused by Standards are relativized, since higher interference voltages are permissible at low frequencies). In principle, a concave frequency curve with a reversal point is therefore proposed Current maximum; however, a technical approach to realizing this functionality will be not described.

The intended function of these solutions in the variants with a triangular course Modulation voltage is dependent on component tolerances, drift effects and Fluctuations in the frequency of the input voltage; even small deviations from the ideal state lead to operation at a limit of the integrator with the consequence of a temporarily constant Switching frequency in the current minimum or maximum and corresponding peaks in Interference voltage curve. There is no automatic adaptation to a changed frequency of the Input voltage, as is the case with worldwide use of devices (with 50 or 60 Hz mains frequency) is required. The variant with a sawtooth shape behaves both with respect to the Dependence on component tolerances as well as on the dependence on the input frequency cheaper; however, the frequency deviation becomes smaller at a 60 Hz mains frequency.

In "An AC-DC converter with low input distortion and near unity power factor" by M. J. Willers u. a. (European Power Electronics Conference Proceedings '93 Vol. 4. p. 1 ff) is next to that already above solution described with an input voltage-triggered wobble generator also theoretically described a solution in which the modulation as a function of the mains input voltage he follows. It is described as a disadvantage that due to the minimal rate of change Switching frequency at the current maximum, higher interference voltages occur at this operating point.

A technical implementation of this approach is from US 5,146,398 "Power factor correction device provided with a frequency and amplitude modulated boost converter " Power factor correction circuit whose current control operates according to the current mode principle is described; To compensate for distortions in the mains current, a input voltage-dependent current to the charging current of the frequency-determining Capacitor modulates the switching frequency so that a sinusoidal curve with the frequency Maximum in the zero point of the mains voltage and the minimum in the voltage maximum arises.  

Another similar solution can be found in SGS-Thomson's “Frequency Modulation on L4981B »(AN833 / 0795) known. In this document a control module for a Power factor correction circuit described u. a. a circuit part for frequency modulation contains. Here the rectified mains voltage is divided by the effective value with a divider divided; the output current of the divider reduces the charging current via a subtracting circuit of the frequency-determining capacitor. So that is the amplitude of the mains voltage Switching frequency minimal with a flat course.

The invention as specified in the preamble of claim 1, as well as those described above Solutions must be differentiated from energy converter concepts with load control via the Variation of the switching frequency, as described, for example, in the «Unitrode Product & Applications Handbook 1995-1996 »and a. are described in Application Note U-138 (S10-363 ff.).

The object of the invention is one as specified in the preamble of claim 1 to provide clocked energy converters with the most cost-effective circuit possible Controls energy converter so that the spectrum of radio interference voltages at the network input as possible runs evenly without peaks.

This object is achieved according to the characterizing part of claim 1.

The subjects of the main claim and the associated subclaims offer the following advantages:

Regarding claim 1: The radio interference voltages are reduced by modulating the switching frequency, without a complex additional oscillator for generating the modulation frequency is required; it is rather the variation of a supply voltage during a network period used to change the switching frequency. Since the interference energy from switching to switching current depends, it makes sense to run the network so that the change speed of the switching frequency in the maximum current amount and in the range of the minimum current amount is minimal and a frequency jump occurs in the minimum current amount. This can be done by a phase ver shift preferably by about 90 ° of the switching frequency-determining voltage or the switching frequency-determining current, compared to the voltage of the source. Has a favorable effect also from that due to the steadily falling or increasing course of the half wave Switching frequency each current / frequency combination occurs only once per half-wave. Through the bond The switching frequency to the input voltage is an automatic adaptation to changes in the Given input voltage frequency, the frequency deviation can change.  

To claim 2: A VCO sets an input voltage, or an input current, in one of the Frequency dependent on the input variable, here also designs with variation of the Output frequency within an upper and lower bound band are known.

Regarding claim 3: It is a simple form of a matching circuit for controlling a VCO

To claim 4: The rectification makes a rippled DC voltage double Grid frequency generated and thus a symmetrical with respect to the network half-waves modulation causes.

To claim 5: By using RC elements, a phase shift is caused.

To claim 6: By applying an additional DC voltage, for example negative DC voltage at the output caused by shifting the input voltage be prevented.

Regarding claim 7: an amplifier can be used for adaptation; can continue through a suitable wiring of the amplifier, for example as an integrator for the AC voltage part or as a differentiator, a phase shift between the output voltage of the network and oscillator can be effected.

Regarding claim 8: By using the divider, the influence becomes one with respect to the effective value variable source voltage eliminated.

Regarding claim 9: The resistor causes a DC voltage to be applied accordingly Claim 7, wherein the voltage on the capacitor is used as a DC voltage source. Herewith the effect of a variable rms value of the source is also eliminated for the connection.

Regarding claim 10: A current-dependent modulation can be used to adapt the modulation to different load conditions occur and thus the influence of a variable amplitude of the Mains voltage can be eliminated.

Regarding claim 11: The mostly low output level of one, especially with direct shunt measurement Current detection can be increased via an amplifier circuit.

To claim 12: The oscillator circuit used in conventional shaft regulator IC behaves as VCO if you put the capacitor in parallel from the charger of the IC and the  charging modulation voltage. The modulation is done here Change an input current.

Regarding claim 13: The range of the switching frequency is determined by a current-limiting element set.

Regarding claim 14: The use of a voltage-dependent current source brings about a linearization of the Switching frequency depending on the output voltage of the network.

Regarding claim 15: by superimposing several variables, for example mains current and Mains voltage, changes depending on load and mains voltage can be compensated.

Regarding claim 16: The application is also possible when using several energy converters, for example with PFC with a downstream DC / DC converter.

Regarding claim 17: Interferences are avoided by synchronized control.

Regarding claim 18: Switching frequencies become switching frequencies due to phase-shifted switching times Harmonics reduced (superposition causes reduction)

Regarding claim 19: As a rule, the uniformity is optimal when controlling several energy converters Phase shift.

Regarding claim 20: To achieve symmetrical modulation in n-phase networks, a modulation voltage with a frequency of 2 × m × f _{N is} generated.

In the following, the invention is described in more detail by way of example with reference to eight figures

Show it

Fig. 1: Block diagram (as a block diagram)

Fig. 2: embodiment of a network between network and oscillator

Fig. 3: embodiment with additional amplifier

Fig. 4: embodiment with divider

FIG. 5 embodiment with use of the network-side current as a frequency-dependent network size

Fig. 6: embodiment with connection of the network to a voltage controlled oscillator

Fig. 7: embodiment for the connection of several energy converters

Fig. 8: embodiment for three-phase connection

Fig. 1 is a basic illustration of the principle. A clocked energy converter ( 2 ), for example a DC / DC converter or an actuator for generating an input current with a predetermined current curve shape, feeds a load ( 3 ). The converter ( 2 ) is supplied from a network ( 1 ) with an input voltage that varies in the millisecond range, for example an AC network. For control purposes, a control circuit ( 4 ) is used, which contains, among other things, an oscillator ( 5 ). The frequency of the oscillator ( 5 ) depends on a voltage or a current at the input; This is implemented, for example, by a capacitor which is charged depending on the input and is discharged to another value when a certain threshold value is reached. In addition to an internal voltage of the control circuit ( 4 ), the output voltage ( 8 ) of a matching network 6 , the input ( 7 ) of which is connected to the network ( 1 ), is connected to the input of the oscillator ( 5 ). The matching network ( 6 ), for example by rectification, resistance division, amplification or phase shift via capacitors or chokes and resistors (passive components), converts the input voltage into a different, adapted voltage which depends on the input voltage over the entire period and changes continuously during a period. Since the frequency of the oscillator ( 4 ) depends on the instantaneous value of the voltage at the output ( 8 ), there is a continuous change in the clock frequency of the energy converter ( 2 ) with a modulation frequency depending on the voltage at the output ( 8 ) of the network ( 6 ).

Fig. 2 shows an embodiment for the network ( 6 ). The voltage at the input ( 7 ) is first rectified via the rectifier bridge ( 10 ), consisting of 4 diodes ( 11 , 12 , 13 , 14 ), the diodes ( 13 , 14 ) being able to be part of the rectifier for the power section at the same time, which results in a wavy DC voltage with twice the mains frequency. This ensures symmetry for the network half-waves. The impedances ( 21 , 22 ) set the u. U. high voltage at the input ( 7 ) in a low voltage for the electronics around; the impedance ( 23 ) limits the current at the output ( 8 ). In principle, the impedance ( 22 ) can be dispensed with; the impedance ( 23 ) can be short-circuited. In the simplest case, only the impedance ( 21 ) is present, designed for example as a capacitor, the current of which is dependent on the rate of change of the input voltage and thus causes the phase shift. The impedances ( 21 , 22 ) can be implemented as a parallel connection of a capacitor with a resistor; this also causes a phase shift between the voltage at the input ( 7 ) and the voltage at the output ( 8 ), thus achieving the desired high rate of change in the voltage maximum. The parallel resistor is useful for discharging the capacitor of the impedances ( 21 , 22 ). The diode bridge ( 11 , 12 , 13 , 14 ) can also perform the function of rectification for supplying the power section ( 2 ) in the case of a power factor correction circuit. In addition, the connection of a DC voltage ( 25 ) via a resistor ( 24 ) is shown here; this source supplies a current which at least compensates for a possible current flowing into the output ( 8 ), which can occur, for example, when a capacitor is discharged. This prevents a negative voltage or a negative current at the input ( 5 ).

If only resistances are provided as impedances ( 21 , 22 , 23 ), the switching frequency is modulated without a phase shift between the voltage of the source ( 1 ) and output ( 8 ) of the network ( 6 ); the rate of change of the switching frequency is minimal in this case in the current maximum.

Fig. 3 shows the expansion of the circuit of FIG. 2 by a negative feedback amplifier. With appropriate wiring by appropriately determining the impedances ( 32 , 33 ) in accordance with the known proposals for wiring operational amplifiers, for example as an integrator for the AC voltage component or as a differentiator, this amplifier can be used not only to adapt the voltage but also to phase shift the voltage at the input of the amplifier ( 34 ) and voltage or current at the input of the oscillator ( 5 ).

In the solutions described above, the switching frequency range changes when the amplitude of the voltage of the network ( 1 ) changes; when the voltage drops, the frequency swing becomes smaller.

Fig. 4 shows how the influence of the change in the amplitude of the voltage of the network ( 1 ) can be compensated by a divider ( 41 ). The rectified voltage at the input ( 7 ) is smoothed, for example by a low-pass filter with voltage reduction, consisting of an ohmic divider ( 47 , 48 or 21 , 22 , 23 , 45 , 48 ) and a capacitor ( 46 ). The time constant is large compared to the frequency of the network ( 1 ); the voltage at the input ( 43 ) of the divider is therefore proportional to the amplitude of the voltage of the network ( 1 ) and changes only slightly during a network period. The voltage at the output ( 44 ) is that it is the quotient of the voltage at the input ( 42 ) divided by the voltage at the input ( 43 ); it therefore has the curve shape of the input voltage, but the amplitude is always almost constant. A negative current at the input ( 42 ) of the divider can also be compensated for here in a voltage-dependent manner by using the voltage at the capacitor ( 46 ), which is an image of the effective value of the mains voltage, as the voltage source ( 25 ); with a fixed DC voltage ( 25 ) the switching frequency would depend on the switching frequency.

Another way to compensate for the influence of the change in the mains voltage is to supply a voltage dependent on the mains current to the input of the oscillator ( 5 ). An exemplary embodiment is shown in FIG. 5; the (rectified) current of the network ( 1 ) flows through a shunt ( 53 ) and causes a current-proportional voltage drop here. The voltage at the shunt ( 53 ) has twice the mains frequency; further rectification is therefore unnecessary. With small currents, the range of variation of the frequency decreases; this is acceptable since experience has shown that the radio interference voltages also decrease with falling currents. In view of the low level of the voltage at the shunt ( 53 ), amplification is advisable; alternatively, a current transformer in the mains supply line (with a correspondingly higher level) with rectification or a direct current transformer can be used.

Fig. 6 shows the functional relationship of the invention with an oscillator (5) of a typical switching regulator IC (described inter alia in "Unitrode Product & Applications Handbook 1995-1996" in Application Note U-93, S10-9 et seq., For the UC3846 ). A capacitor ( 61 ) is charged by a current source ( 62 ). An amplifier ( 64 ) compares the voltage of the capacitor ( 61 ) with a reference voltage ( 66 ); when the upper threshold value is reached, the switch ( 65 ) is closed and the capacitor ( 61 ) is discharged via the current source ( 67 ), which is considerably stronger than the current source ( 62 ). When the lower threshold value predetermined by the hysteresis of the amplifier ( 64 ) is reached, the switch ( 65 ) is opened again. The output ( 69 ) of the amplifier ( 64 ) is fed to internal components of the control ( 4 ); a new switching cycle of the energy converter ( 2 ) is started during the pulse-like discharge of the capacitor ( 61 ). In some control concepts, the approximately triangular voltage of the capacitor ( 61 ) is compared with a control voltage for controlling the energy converter ( 2 ); in Fig. 6. the corresponding connection ( 68 ) is led out.

In this example, the voltage at the output ( 8 ) of the network ( 6 ) is equal to the voltage at the capacitor ( 61 ); depending on the impedances ( 21 , 22 , 23 ) and the voltage of the network ( 1 ), a current flows through the resistor ( 23 ), which additionally charges the capacitor ( 61 ). The capacitor ( 61 ) therefore reaches the threshold value faster than specified by the current source ( 62 ); the switching frequency increases. The higher the current through the resistor ( 8 ), the higher the switching frequency.

Fig. 7 shows an application example of the terminal displays a plurality of clocked power converter, for example, the series circuit of a power factor correction circuit with DC / DC converters. The output ( 69 ), which has a multiple (with the number of energy converters as multiplier) of the switching frequency of the downstream energy converters ( 75 , 76 , 77 ), of a solution as described above, consisting of a matching network ( 6 ) and an oscillator ( 5 ), is first turned on guided a frequency divider, which leads the discharge pulses of the output ( 69 ) in succession to the control circuits ( 72 , 73 , 74 ) of the converters ( 75 , 76 , 77 ).

Fig. 8 shows an embodiment for a three-phase network; the 6-pulse rectification via the diodes ( 11 , 12 , 13 , 14 , 15 , 16 ) results in a voltage at the output ( 8 ) with six times the mains frequency, which leads to symmetry for each leading phase of the diodes.

Claims (20)

1. Circuit for driving clocked electronic energy converters ( 2 ), the input current of which has a profile which has a frequency and phase position which corresponds to the profile of the output voltage of the source ( 1 , 53 ), and a circuit part ( 5 ) which determines the switching frequency is connected via an electrical network ( 6 ) to a source ( 1 , 53 ) with a periodically time-varying output voltage that results in a dependence of the instantaneous switching frequency of the energy converter ( 2 ) on the instantaneous value of the output voltage of the source ( 1 , 53 ) at any moment characterized in that

• 1. the electrical network ( 6 ) causes a phase shift between the voltage at its input ( 7 ) and the voltage or the current at its output ( 8 ), so that the maximum switching frequency change in approximately the maximum amount of the input current of the energy converter ( 2 ) is reached becomes
• 2. and the maximum and minimum values of the switching frequency occur at the minimum amount of current at the input of the energy converter

2. Circuit according to claim 1, characterized in that the circuit frequency determining the sheep frequency ( 5 ) is designed as a voltage (VCO) or current-controlled oscillator and the output ( 8 ) of the network ( 6 ) is connected to the input of the oscillator. 3. A circuit according to claim 1 or 2, characterized in that the network ( 6 ) has a series connection of passive components, which indirectly ( Fig. 2, 3, 4, 6, 8) or directly ( Fig. 5) with the source ( 1 , 53 ) is connected, a further passive component ( 23 ) being connected to the tap of the series circuit ( 21 , 22 ) and at least one component being a capacitor or an inductor. 4. A circuit according to claim 3, characterized in that

• 1. the source ( 1 ) emits a rippled DC voltage or
• 2. is an AC voltage source which is connected to the input of the network ( 6 ) via a rectifier circuit ( 10 , 11 , 12 , 13 , 14 , 15 , 16 ).

5. Circuit according to one of claims 1 to 4, characterized in that one of the components ( 21 ; 22 ) is designed as a parallel connection of a capacitor and a resistor and the other components ( 22 , 23 ; 21 , 23 ) are resistors. 6. Circuit according to one of claims 1 to 5, characterized in that the output ( 8 ) of the network ( 6 ) is connected via a resistor ( 24 ) to a DC voltage source ( 25 ). 7. Circuit according to one of claims 1 to 6, characterized in that an amplifier ( 31 ) with an input impedance ( 32 ) and a feedback impedance ( 33 ) between the output ( 8 ) of the network ( 6 ) and the frequency-determining circuit part ( 5th ) is arranged and that amplifier ( 31 ) and the associated circuitry ( 32 , 33 ) cause a phase shift between the voltage at the output ( 8 ) and the voltage or the current at the frequency-determining circuit part ( 5 ). 8. Circuit according to one of claims 1 to 7, characterized in that the output ( 8 ) of the network ( 6 ) to the counter input ( 42 ) of a divider ( 41 ) is guided, at the denominator input ( 43 ) one to the voltage of the source ( 1 , 53 ) proportional DC voltage with low ripple, and its output ( 44 ) is connected to the input of the frequency-determining circuit part ( 5 ). 9. Circuit according to one of claims 6 or 8, characterized in that the inputs of the divider ( 42 , 43 ) are connected to one another via a resistor ( 47 ). 10. Circuit according to one of claims 1 to 9, characterized in that an output signal of a current actual value detection ( 53 ) is used as a network-dependent source. 11. The circuit according to claim 10, characterized in that an amplifier circuit is arranged between the current actual value detection ( 53 ) and the network ( 7 ). 12. Circuit according to one of claims 2 to 11, characterized in that the voltage controlled oscillator ( 5 ) consists of a capacitor ( 61 ), a charging device ( 62 ) which charges this capacitor, and a discharge device ( 63 ), and the output ( 8 ) of the network ( 6 ) is connected to one pole of the capacitor ( 61 ). 13. The circuit according to claim 12, characterized in that a current-limiting element is inserted between the output ( 8 ) of the network ( 6 ) and the capacitor ( 61 ). 14. Circuit according to one of claims 1 to 13, characterized in that a circuit is inserted between the output ( 8 ) of the network ( 7 ) and the frequency-determining circuit part ( 5 ) capacitor, which the voltage at the output ( 8 ) of the network ( 6 ) converted into a current proportional to this. 15. Circuit according to one of claims 1 to 14, characterized in that several of the networks ( 6 ) described are used in parallel, the outputs ( 8 ) of which are connected to one another and are routed together to the frequency-determining circuit part ( 5 ) of an energy converter ( 4 ) . 16. Circuit according to one of claims 1 to 15, characterized in that the outputs ( 8 ) of several of the networks connected to the source ( 6 ) are connected to the frequency-determining circuit parts ( 5 ) of a plurality of control circuits ( 4 ) for a plurality of clocked energy converters ( 2 ) are. 17. Circuit according to one of claims 1 to 15, characterized in that on the frequency-determining circuit part ( 5 ) to which the or the networks ( 6 ) are guided, via a coupling circuit ( 71 ) control circuits ( 72 , 73 , 74 ) for one or more energy converters ( 75 , 76 , 77 ) are connected so that all connected energy converters are operated at the same clock frequency. 18. Circuit according to claim 17, characterized in that the coupling circuit ( 71 ) causes a time shift between the switching times of the control circuits ( 72 , 73 , 74 ). 19. A circuit according to claim 18, characterized in that the coupling circuit (71), a frequency divider 1 / n for n energy converter (75, 76, 77) is used and at each pulse of the oscillator (5) one of the n power converters (75, 76 , 77 ) is switched. 20. Circuit according to one of claims 1 to 19, characterized in that the source ( 1 ) is m-phase and the rectifier ( 10 ) as a bridge with 2 × m diodes ( 11 , 12 , 13 , 14 , 15 , 16 ) is executed.