RU2796945C1 - Class d amplifier channel pulse width modulator - Google Patents

Abstract

FIELD: amplifying technique.

SUBSTANCE: used in broadband power amplifiers of hydroacoustic radiating paths. To do this, in a single-channel class D amplifier containing a control circuit with width-modulated signal buses that provides alternate control of the transistors of the key amplifier, an output current sensor and a two-threshold control circuit are introduced that provide alternate control of transistors, taking into account the measured value of the output current, which makes it possible to eliminate current interruption in the inductor of the output filter with minimal turn-on delays of transistors.

EFFECT: improving the quality of the output signal while reducing dynamic energy losses.

1 cl, 5 dwg

Description

The invention relates to the field of amplifying technology and can be used in broadband power amplifiers of hydroacoustic transmitters.

Power amplifiers for excitation signals of hydroacoustic emitters are characterized by high power (from hundreds of VA to tens of KVA) and a wide frequency range (from hundreds of Hz to tens of kHz) in combination with a high quality of the output voltage when operating on a complex load with a very low active power factor. [1. Aleksandrov V.A. and others. Transmitting paths of low-frequency sonar / marine radio electronics No. 1 (67) 2019 p. 10-14]. As an effective direction for the development of power amplifiers with energy recovery from a complex load, key power amplifiers (KPA) with pulse-width modulation, defined according to the accepted classification as class D amplifiers, should be singled out [2. Artym A.D. Class D amplifiers and key generators in radio communications and broadcasting. M.: Communication. 1980, p. 207]. In such devices, the amplified signal is converted into a sequence of pulses that are amplified in power and fed to the load through a low-pass filter (LPF), the input inductive resistance of which limits the amplitude of the high-frequency current of the CCM.

The low-pass filter current in the switching amplification circuits is closed in turn through open key elements into the power supply and load buses, which minimizes static energy losses at the residual voltage of the amplifying devices during the regeneration of the reactive power component of the complex load. At the same time, there is a change in the conductivity of amplifying devices (transistors and diodes), which is associated with dynamic loss factors, mainly due to through currents transistor-diode and transistor-transistor. If the first factor relates to the fatal process of closing regenerative diodes, then the influence of the second factor can be eliminated by introducing transistor turn-on delays. Moreover, in the latter case, the turn-on delay should be greater than the maximum turn-off delay of the transistors, the value of which, as a rule, increases significantly (by a factor of 3-4) with an increase in the output current. The introduction of such turn-on delays of transistorized CCMs fundamentally affects the nonlinear distortions of the output voltage [3. Aleksandrov V.A. et al. Development of energy-efficient broadband hydroacoustic transmitters for underwater communication / Hydroacoustics 32(4). 2017 p. 56-64].

при таком управлении имеет место эффект обрыва тока дросселя, который приводит к весьма значительным нелинейным искажениям выходного сигнала [Александров В.А. и др. Искажения в двухтактных усилителях низкой частоты с широтно-импульсной модуляцией / Радиотехника №10. 1996 г. с. 36].To improve the quality of the output signals of the key amplification, made on the bridge final stage, in known technical solutions [US Pat. RF №2223529 publ. 02/10/2004; Pat. RF. No. 2309524 publ. 27.10.2007] it was proposed to ensure the control of transistors of only one diagonal with a given polarity of the modulating signal. This eliminates the simultaneous switching of transistors of one channel, made on a half-bridge rack, this eliminates the possibility of current flow transistor-transistor without introducing additional delays in the front of the control pulses. The selected circumstance makes it possible to reduce dynamic energy losses while reducing nonlinear distortions in the absence of a phase shift between the voltage and load current, which occurs only when the active power factor cosϕ _n =1. However, if this condition is violated, for the time of the phase shift between current and voltage

with such control, the effect of interruption of the inductor current takes place, which leads to very significant non-linear distortions of the output signal [Aleksandrov V.A. and other Distortions in push-pull amplifiers of low frequency with pulse-width modulation / Radio engineering №10. 1996 p. 36].

The closest to the proposed device is a pulse-width modulator of a bridge inverter with alternate control of key elements for their uniform loading, taking into account the polarity of the amplified signal, described in RF patent No. 2309524, published on 10/27/2007. an inverter made according to a half-bridge circuit with separate transmission of direct F (+) and inverse F (-) PWM signals, taking into account the state of two control commands. The present device, in terms of the number of common features in terms of a single-channel class D amplifier, is the closest analogue of the proposed device and can be taken as a prototype of the invention.

The block diagram of the prototype device, shown in Fig. 1, contains a control circuit 1 and a key amplifier 2, made according to known rules on a rack of key elements KE1 and KE2, included in a half-bridge circuit between power supply buses +E and -E. Key elements 2.1, 2.2. contain field-effect transistors with a reverse diode and drivers of pulse signals coming from the outputs of the control circuit 1. In turn, the first and second inputs of circuit 1 are connected to the buses direct F(+) and inverse F(-) PWM signals, and the first and second control inputs to the control buses Y ₁ and Y ₂ .

In accordance with the principle of operation of the prototype device, the synchronization signal Y ₁ is used as control signals, which represents a meander-type pulse sequence obtained by dividing the clock pulses that determine the formation of PWM signals, and the glued amplified signal Y ₂ that determines the polarity of the half-waves of the modulating effect:

It is the signal Y ₂ that determines the passage of pulses F (+) or F (-) to control the upper 2.1 or lower 2.2 key element of the key amplifier 2.

Pulse signals X ₁ and X ₂ coming to the control inputs of the key amplifier 2 are determined from the conditions:

In this case, there are no simultaneous switching of transistors, the change in conductivity of which is determined by the change in the polarity of the modulating effect of the formation of the output voltage of the key amplifier 2, connected according to known rules through the LPF choke. The load voltage is provided by the conductive transistor of this CE and the open reverse diode of another CE. For example, for U>0, the pulse signal F(+) is fed to the control of the transistor KE1, which fixes the voltage V ₁ =+E at the output of the key amplifier for the duration of the pulse τ _and (+). During the pause between pulses t _and (-) and t _and (+), the output voltage V is formed by closing the low-pass filter current through the regenerative diode KE2, V = -E. For U<0, the control of the transistor KE2 is provided, fixing during pulses of duration t _and (-) the output voltage V ₁ \u003d -E, and during the pause between pulses of durations t _and (+) and t _and (-) voltage V ₁ \u003d + E, is determined by the open diode KE1. Thus, for the output current t ₁ coinciding in phase with the signal U in the prototype device, an undistorted pulse voltage V is formed, coinciding with the PWM signal F in the absence of simultaneous switching of transistors KE1 and KE2 with a PWM clock frequency, which minimizes output signal distortion while reducing dynamic energy losses.

However, the principle of operation proposed in the prototype device leads to significant distortion of the output signal in the presence of significant phase shifts between the voltage and load current. Characteristic for broadband hydroacoustic radiators is the phase shift ϕ _n corresponding to the active power factor cos ϕ _n =0.11…0.5. Under these conditions, in the phase interval ϕ _n , a current of reverse polarity should be closed through the key amplifier 2 compared to the polarity of the modulating signal:

At the same time, in the prototype device, the controlled key element closes the voltage V through an open regenerative diode to the corresponding voltage bus E, and the choke current does not close through the uncontrolled key element, and the effect of current interruption is observed. There is no current in the inductor, and the load voltage determined by the filter capacitance is formed at the output of the key amplifier 2. As a result, the linearity of the modulation characteristic of the pulse-width modulator is violated, which leads to unacceptable signal distortions at the load, reaching 10-30%.

The selected effect is analyzed in detail in [4. Aleksandrov V.A. Distortion in push-pull low-frequency amplifiers with pulse-width modulation / Radio engineering 1986 No. 10. from 37-40].

With regard to the half-bridge class D amplifier circuit, the distorted amplitude characteristic has a similar character, taking into account the high-frequency current flow of the low-pass filter choke. This disadvantage leads to an increase in non-linear distortion in the prototype device and technical analogues, which prevents the use of known technical solutions aimed at reducing dynamic energy losses.

The objective of the invention is to improve the quality of the output signal while improving the energy characteristics of a single-channel class D amplifier.

The technical result from the use of the claimed invention consists in supplementing the mode of alternate control of the key elements of the half-bridge circuit with the mode of joint control with a minimum delay of the pulse front at different polarities of the voltage and current of the load, which minimizes linear distortion while reducing dynamic energy losses.

The technical result is achieved by the fact that in a well-known pulse-width modulator of a class D amplifier, containing a control circuit, the first and second inputs of which are connected, respectively, to the tires of the first and second pulse-width modulated pulse signals, the first and second control inputs, respectively, to the first and the second control buses, and the first and second outputs are connected to the first and second inputs of the key amplifier, connected by power supply outputs to the positive and negative voltage buses, also containing a load bus, it is proposed to introduce a two-threshold comparison circuit and a current sensor connected between the output of the key amplifier and the load bus , wherein the current sensor output is connected to the input of a two-threshold comparison circuit, the first and second outputs of which are connected to the first and second control buses, and the control circuit includes the first and second pulse front delay nodes, as well as the first and second coincidence circuits, the outputs of which are connected to the first and second outputs of the control circuit, and the first inputs - with the first and second control inputs of the control circuit, the first and second inputs of which are connected through the first and second node of the pulse edge delay with the second inputs, respectively, of the first and second coincidence circuits.

The introduction of new features makes it possible to reduce non-linear distortions caused by a break in the output current of a key amplifier when operating on a complex load with an increase in energy efficiency by eliminating the through current transistor-transistor and ensuring switching of only one transistor when the polarity of the output low-frequency (LF) voltage and current coincide.

The essence of the invention is illustrated in Fig. 1, fig. 2, fig. 3, fig. 4a, 4b, which shows a block diagram of the prototype device (Fig. 1) and the claimed device (Fig. 2), as well as timing diagrams of signals (Fig. 3 and Fig. 4a), explaining the features of the proposed pulse-width modulator of the amplifier channel class D when compared with the prototype device (Fig. 4b). In FIG. 3 and FIG. 4a and 4b, the following designations are adopted:

u is the voltage of the modulating signal;

U _n - reference sawtooth voltage for the formation of PWM signals;

F ₁ and F ₂ - direct and inverse PWM signals on the output buses of the device;

V and U - output pulse voltage and low-frequency component of the amplified and modulating signal;

U _i and U _in - the measured values of the output current of the key amplifier and the low-frequency component of the output current of the load;

+U ₀ and - U ₀ - threshold values of the measured value of the output current of the key amplifier, set in the two-threshold comparison circuit; (Y ₁ and Y ₂ - signals on the first and second control buses;

X ₁ and X ₂ - output signals of the control circuit to the inputs of the key amplifier;

ϕ - phase shift between the low-frequency component of the output voltage and the low-frequency component of the load current;

U' - distorted value of the low-frequency component of the output voltage;

ΔU - deviation of the value U' from the value U, due to the effect of current interruption.

The proposed pulse-width modulator of a single-channel class D amplifier (Fig. 2) contains a control circuit 1, including delay nodes 1.3, 1.4 of the pulse front and coincidence circuits 1.1, 1.2, with PWM signals F _{1shim and} F _2shim and control buses Y ₁ and Y ₂ , as well as with output buses X ₁ and X ₂ connected to the inputs of the key amplifier 2, with power supply voltage buses -1-E and - E, between which the key elements KE1 and KE2 are connected, and the output bus connected through the current sensor 3 with bus load, and two-threshold circuit 4 comparison, including two key circuits 4.1 and 4.2.

The composition of the proposed device includes blocks made according to known rules in accordance with the declared implementation features, the combined use of which ensures the achievement of a technical result.

The control circuit 1 is designed to transmit sequences of inverse pulses E _1PWM and E _2PWM from the buses of width-modulated signals to the control inputs of the key amplifier 2, taking into account the enabling commands Y ₁ and Y ₂ generated on the first and second control buses. As part of the control circuit, the pulse edge delay nodes 1.3 and 1.4 and the coincidence circuits 1.1 and 1.2 are used. In this case, to form a pulse front delay in nodes 1.3, 1.4, a resistive-capacitive circuit with a forced diode capacitance discharge can be used, provided that logic elements are used as a coincidence circuit and with outputs on Schmitt triggers (with amplitude hysteresis). In the proposed device, a minimum sufficient delay time of no more than 0.1 μs can be provided.

Key amplifier 2 as part of a class D amplifier is implemented according to a typical half-bridge circuit [1. Aleksandrov V.A. and others. Transmitting paths of low-frequency sonar / marine radio electronics No. 1 (67) 2019 p. 10-14] key elements 2.1, 2.2 on field-effect transistors connected in series between the positive and negative power supply buses. To match the control of field-effect transistors with the outputs of the control circuit, pulse signal drivers are included in the key elements. The output pulse voltage of the key amplifier 2 is transmitted, as a rule, through the output bus and the LPF choke to the load, which is used in hydroacoustic radiating installations as piezoelectric transducers with a pronounced capacitive component of conductivity. In this case, the low-frequency regeneration current from the load capacitance to the power supply buses should be closed in the interval of the phase shift between the low-frequency component U of the output voltage V and the low-pass filter current.

Measurement of the load current provides a current sensor, which can be performed on a current transformer with an allowable frequency that is a multiple of the minimum frequency of the signal being amplified.

The two-threshold comparison circuit should highlight the levels of instantaneous values of the measured value of the output current to form a zone of simultaneous anti-phase control of the transistors of the key amplifier 2 by generating control commands Y ₁ and Y ₂ :

where U ₀ is the minimum allowable level of the absolute value of the measured inductor current U _i , at which only one transistor of the key amplifier 2 can be controlled:

определяется исходными сигналами F_1шим и F_2шим с задержкой фронта импульсов.Where

is determined by the original signals F _1shim and F _2shim with a delay in the front of the pulses.

The proposed pulse-width modulator of a single-channel class D amplifier operates as follows.

например, в результате сравнения выходного сигнала U с опорным пилообразным напряжением U_п (фиг. 3), поступают на первую и вторую шины ШИМ сигналов:Pulse sequences F ₁ and F ₂ formed out of phase

for example, as a result of comparing the output signal U with the reference sawtooth voltage U _p (Fig. 3), they arrive at the first and second PWM signal buses:

Further, taking into account the level of commands Y ₁ and Y ₂ received on the first and second control bus, pulse signals F ₁ and F ₂ with a front delay τ _f are transmitted in the form of pulses X ₁ and X ₂ to control the first and second key element KE2.1 and KE2 .2 key amplifier 2:

The time interval τ _f , formed by the delay nodes, is very small, does not take values of 0.05-0.1 μs, and is determined from the condition of exclusion of the through current transformer-transformer at the value of the output current i _L corresponding to the measured value \ |U _i |< U ₀ .

As a result, taking into account the expression (4) for the measured value |U _i |> 0 in the proposed device, only the transistor of one CE is switched when the current is closed between pulses through the diode of another CE:

For U _i >U ₀ the conductivity of the transistor KE1 at X ₁ =1 and the conductivity of the diode KE2 at X ₂ =0;

For U _i <-U ₀ , the conductivity of the transistor KE2 is ensured at X ₂ - 1 and the conductivity of the diode KE1 at X ₂ =0.

In this case, the pulse voltage V is practically not distorted, and in the amplitude characteristic of the key amplifier there is a slight shift by the calculated value [4]:

where T is the switching period.

For the measured value |Ui|<U ₀ in the proposed device, a transition zone is implemented, which involves switching the transistors of the key amplifier 2 in turn. In this case, as illustrated in FIG. 3, when comparing the measured values Hi of the inductor current, containing components expressed in 4, with the threshold values + U ₀ and - U ₀ , pulses Y ₁ and Y ₂ are formed allowing alternate switching of transistors KE1 and KE2, controlled by signals X ₁ and X ₂ .

However, even in this mode, there is no simultaneous switching of transistors during the signal decay (U _i ) to the boundary value U ₀ . The exception is the change in the signals F ₁ and F ₂ during equality |U _i | = _U0 . It is in this boundary zone that the possibility of through current transistor-transistor is eliminated by the minimum delay of the front of the pulses, the values of which can be significantly less (at least three times) than the typical turn-on delay, which guarantees the elimination of through current transistor-transistor in the entire range of change in the current of the inductor used in technical counterparts [1. Aleksandrov V.A. and others. Transmitting paths of low-frequency sonar / marine radio electronics No. 1 (67) 2019 p. 10-14, 2. Artym A.D. Class D amplifiers and key generators in radio communications and broadcasting. M.: Communication. 1980, p. 207, 3. Aleksandrov VA et al. Development of energy-efficient broadband hydroacoustic transmitters for underwater communication / Hydroacoustics 32(4). 2017 p. 56-64, 4. Aleksandrov V.A. Distortion in push-pull low-frequency amplifiers with pulse-width modulation / Radio engineering 1986 No. 10. from 37-40].

It should be noted that the technical effect of using the proposed technical solution can also be achieved in two-channel and multi-channel circuits of key amplification by summing the output signals through low-pass filters.

The main advantage of the proposed technical solution in comparison with the prototype device (RF patent No. 2309524) manifests itself when a complex load is excited under conditions of a significant shift between the load current and voltage, which is typical for the operation of the amplifier on a broadband hydroacoustic radiator, the capacitive conductivity of which can significantly exceed the active conductivity.

In FIG. 4.a and 4.b illustrate the timing diagrams of signals in the proposed (Fig. 4.a) device and the prototype device (Fig. 4.b) in the phase interval cp of the shift between the phase of the low-frequency voltage U and the load current.

In the proposed device, the principle of operation is preserved with minimal distortion of the shape of the pulsed voltage. The LPF inductor current is alternately closed through transistors and diodes KE1 and KE2, which ensures the closure of the regeneration current from the load capacitance to the power supply buses. At the same time, non-linear distortions in the rated power mode at a switching frequency of 100 kHz do not exceed 0.5% and may increase slightly (up to 1%) at low levels of load current corresponding to the boundary values of the measured value U _i.

In the prototype device, which assumes the ability to control transistors only when the polarity of the modulating signal is changed and coincides with the low-frequency component and the ideal pulse voltage V, there is no possibility of recuperation of the low-frequency load current, which leads to a current interruption through the regenerative diode and distortion of the pulse voltage shape. In FIG. 4.a shows that the measured value of the inductor current U _i , for a positive value U, penetrating through the diode KE2 drops to zero, the diode closes and the load voltage is induced at the output of the key amplifier 2 until the next turn on of the transistor KE 1.

As a result, there is a significant change in the value of U with a deviation of ΔU from the given shape, which causes nonlinear distortions of up to 10-15%) at the nominal output power level with a possible increase of up to 20-30%) at a low level of low-frequency current.

Thus, in the claimed device, an improvement in the quality of the output signal is achieved by eliminating the interruption of the inductor current while minimizing the delay in the front of the pulses in terms of improving the energy characteristics of the corresponding separate operation of the transistors of key elements in accordance with the polarity of the output current. At the same time, in comparison with the prototype, the non-linear distortions associated with current interruption are reduced from 10-15% to 0.5-1.5%, in terms of reducing dynamic losses by more than two times compared to technical analogues with simultaneous switching of key elements .

An experimental model of a broadband amplifying device is being manufactured at the enterprise based on the proposed technical solution for excitation of hydroacoustic emitters, the test results of which confirmed the achievement of the claimed technical result and the validity of introducing the present invention into a pulse-width modulator of a class D single-channel amplifier.

Claims (1)

A pulse-width modulator of a single-channel class D amplifier, containing a control circuit, the first and second inputs of which are connected, respectively, by buses of the first and second pulse-width-modulated pulse signals, while the first and second control inputs, respectively, with the first and second control buses, and the first and second outputs are connected to the first and second inputs of the key amplifier, the power supply outputs of which are connected to the buses of positive and negative power supply voltage, also containing a load bus, characterized in that it includes a two-threshold comparison circuit and a current sensor connected between the output of the key amplifier and a load bus, wherein the output of the current sensor is connected to the input of a two-threshold comparison circuit, the first and second outputs of which are connected to the first and second control buses, and the control circuit includes the first and second node of the pulse edge delay, as well as the first and second coincidence circuit, outputs of which are connected to the first and second outputs of the control circuit, and the first inputs - to the first and second control inputs of the control circuit, the first and second inputs of which are connected through the first and second node of the pulse front delay with the second inputs, respectively, of the first and second coincidence circuits.

Images