DE10339478B4 - Circuit and method for processing a supply voltage with voltage spikes - Google Patents

Abstract

Circuit for processing a supply voltage (47a) with a voltage peak (60) to obtain an output voltage (U _a ) with reduced or eliminated voltage peaks, having the following features:
a first capacitance (43) between a first node (41) and a second node (42), wherein an input voltage (U _e ) between the first node (41) and the second node (42) due to the supply voltage (47a) can be applied ;
a second capacitance (46) between a third node (44) and a fourth node (45), wherein the output voltage (U _a ) between the third node (44) and the fourth node (45) can be tapped, wherein the first capacity ( 43) has a smaller capacitance value than the second capacitance (46);
a controllable resistor (48) between the first node (41) and the third node (44); and
a control device (49) for driving the controllable resistor, wherein the control device (49) is formed,
in a first case (52), in which the input voltage (U _e ) is less than one ...

Description

The The present invention relates to circuits for processing supply voltage with voltage peaks and in particular switching regulators, where the sieve capacity has such a circuit.

7 shows a known buck converter with a simple switch, as shown for example in "Semiconductor Circuitry" U. Tietze, CH Schenk, Springer-Verlag, 9th edition, 1989, Figure 18.37 on page 564. The down converter in 7 includes a ring-like interconnection with a coil 400 , a capacitor 402 and a diode 404 , The down converter or switching regulator in 7 further includes a charging switch 406 as well as an in 7 not shown control, which is designed to the charging switch 406 to control so that the output voltage of the switching regulator, the in 7 with U _SR is designated to keep at a defined level or to keep in an area around the defined level. The frequency with which the switch 406 is opened or closed depends on a load 408 from, ie of the power consumption of the load. When the power consumption of the load is high, the duty ratio, ie, the period in a cycle in which the switch is closed, will be high at the time in one cycle in which the switch is open. In contrast, in the case of a low-current load (I _SR ), the switch will only be closed in one cycle for a shorter period of time.

In the 7 The circuit shown comprises a certain number of nodes, which are set out below. At a first entrance node 410 the circuit is a pole of an input voltage source U _{0 is} connected, while at a second input node 412 another potential of the input source U _{0 is} connected. The second entrance node 412 is typically the ground node. A first exit node 414 is also referred to as a first output rail or positive output rail, while a second output node 416 Also referred to as the second output rail or negative output rail when the in 7 shown convention is used for the output voltage of the switching regulator U _SR . The desk 406 is on the one hand between the first entrance node 410 and a first intermediate node 418 connected. Further, the diode 404 so between the first intermediate node 418 and the second input node 412 connected that anode of the diode to the second input node 412 while the cathode of the diode is connected to the first intermediate node 418 connected is. Further, as it is in 7 shown is the capacitor 402 between the first exit nodes 414 and the second output node 416 connected. According to the in 7 shown configuration of the network of diode, coil and capacitor is the coil between the first intermediate node 418 and the first output node 414 connected.

The following will highlight the functionality of the in 7 shown circuit received. As long as the switch 406 is closed, U _{D is} equal to the negative input voltage U ₀ . When it opens, the inductor current I _{L maintains} its direction, and U _D decreases in magnitude until the diode becomes conductive, that is, approximately at 0 potential. The time course of the coil current results from the law of induction, according to which the voltage across the coil is equal to the product of the inductance L of the coil and the derivative of the coil current with respect to time. During the switch-on time, ie when the input voltage U ₀ at the diode 404 is applied to the throttle, the voltage U ₀ - U _SR . During the switch-off time t _from the switch 406 the voltage U _L = -U _SR is applied to the choke. This results in a current change ΔI _L , which is given as follows: .DELTA.I L = 1 / L · U SR · t out ,

From this balance again the output voltage can be calculated, which is defined as follows: U SR = t one / (T one + t out ) · U 0 = t one / T * U 0 = t · U 0 ,

In the above equation T = t _on + t _off = 1 / F, the oscillation period, and p = t _on / T is the so-called duty cycle. It can be seen that, as expected, the arithmetic mean of U _D is the output voltage. Typically, the inductance L of the coil 400 chosen so that a minimum current is not exceeded, as is known in the art. It is also known that as the clock frequency is increased, the inductance can be reduced. Furthermore, when the frequencies are too high, the cost of the switching transistor and the drive circuit increases. In addition, dynamic switching losses increase in proportion to the frequency. For these reasons, switching frequencies between 20 kHz and 200 kHz are preferred.

The capacitor 402 determines the ripple of the output voltage. The generation of the switching signal for switching the charging transistor 406 is usually done by a pulse width modulator and a regulator with voltage reference. In detail, a reference voltage, which supplies a desired value, is fed to a subtractor, to which the actual output voltage U _SR is also supplied as an actual value. The output of the subtractor is fed to a variable gain amplifier which feeds a comparator, on the other hand to the signal generated by a sawtooth generator to be led. The output of the comparator is the control signal for the switch 406 in 7 , Typically, the variable gain amplifier is a PI control amplifier. It increases its output signal until the difference at the output of the subtractor becomes 0, until the output voltage U _SR equals the desired output voltage. Typical dimensions for coil sizing are in the two-digit micro-Henry range, while typical values for capacitors are in the three-digit micro Farad range when switching frequencies in the range of 50 kHz are used.

In 7 Switching regulators shown to provide a suitable power supply to a subsequently connected circuit, such as an ASIC. The power supply usually consists of one or more constant DC voltages of, for example +5 V or ± 15 V. Often this is not from the outset in the desired form available and must first by, for example, a in 7 shown switching regulator, which can be supplemented by a downstream linear regulator to eliminate the ripple of the output voltage can be generated. At the entrance to the in 7 The switching regulator shown is usually an inverter, which generates the input voltage U ₀ from the AC or three-phase network (230 V or 400 V) of the power companies.

So deviate from the in 7 Switching controller shown, other controllers with a transformer, a rectifier, a smoothing capacitor and possibly a linear regulator for voltage stabilization. However, the transformer is expensive to produce and therefore expensive. He also needs a lot of space. Another disadvantage of the transformer is its frequency-dependent working range. This is z. B. restricted to the mains frequency of 50 Hz or 60 Hz. If the frequency deviates, then this also results in a deviation of the output voltage of the transformer. With a DC voltage at the input, the voltage transmission does not work.

Letting go of the transformer and using only rectifier, smoothing capacitor and a linear regulator, then much energy is lost in the form of heat. In addition, it must then be ensured sufficient cooling of the linear regulator, which in turn is very expensive and needs space. All this is bypassed if, as it is based on 7 has been shown, instead of the linear regulator uses a switching regulator. Due to the significantly better efficiency, little energy is lost in the form of heat, and consequently the cost of cooling is much lower. The switching regulator, as has been stated, requires a choke (the inductance 400 in 7 ), which is relatively expensive to manufacture. However, this has only one winding and is thus easier to manufacture than a transformer, the two windings be sitting. In addition, the choke can be reduced by choosing a higher working frequency.

Many well-known switching power supplies, as well as in 7 shown switching power supply are problematic in certain respects. Usually the input voltage range is limited to a ratio of U _{E, max} / U _{E, min} ≤ 5, which can be seen in the catalogs of different manufacturers. This range is too low for some applications and should be increased to a ratio of, for Eg 20: 1.

The Voltage supply of the controller itself takes place either via a separate voltage source or is generated from the input voltage, what an extra Voltage regulator and thus additional Effort means.

Further is for strived for a flexible use, the input voltage significantly to choose larger can, as the maximum allowable Operating voltage of the regulator itself, without requiring additional voltage regulator for the Generation of this operating voltage can be used.

Further should when applying the input voltage as fast as possible, controlled Starting the switching regulator guaranteed be. Especially for Time-critical applications should have this so-called startup delay like this small as possible be.

The DE 199 46 025 A1 the DE 197 00 100 C2 , the DE 195 07 553 A1 or the DE 197 06 491 A1 reveal switching regulator based on the in 7 shown schematic diagram work, depending on the embodiment, the coil 400 either between the first intermediate node 418 and the first output node 414 , as in 7 is shown, or alternatively between the second input node 412 and the second output node 416 is arranged. The control voltage for the switch 406 , which can be implemented as a transistor, is controlled by the regulator, which controls the timing of the switch 406 determined, or generated externally. This leads to additional circuit complexity, which means additional costs in terms of design, testing and production. Additional costs are disadvantageous, in particular, since switching regulators, in particular if they are provided inside luminaires or are also designed to be fully integrated with an integrated circuit to be supplied, increase the price of the end product and, in particular, in chip applications in which the chip area is a criterion. due to the increased chip area consumption arise.

In particular, when starting up a switched-mode power supply, that is to say when the input voltage applied to the load switch and the diode is initially switched on, high voltage peaks may occur. This is because when a current flows through a coil, and the current is turned off or converted to another current value, a high voltage spike occurs with a direction opposite to the previous current direction. After switching power supplies or such switching regulator, as shown in 7 are based on the fact that constantly a current through a coil (and a capacitor connected in series) is turned on or off, this may already arise voltage spikes. However, in particular in the run-up process, where the states are not defined in such a way, or where initial controls must be made, which may set a current through a coil to 0, such spikes may arise. If the output voltage containing such a voltage spike used to z. B. in a switching regulator to supply the control of the switching regulator itself, so it may be that such a spike in the output signal that should provide the control, can lead to destruction of the control or of input components of the controller. This could be counteracted by designing the input stages of the controller to be very voltage-proof. However, this oversizing required in terms of withstand voltage and normal operation results in additional costs. This approach is becoming less feasible when very high voltage spikes and relatively low output operating voltage or output voltage setpoint are needed. Then the "mismatch" between the voltage design of the controller on the one hand and the actual voltage to be handled by the controller on the other hand, always larger.

The DE 4007593 A1 discloses an input current peaking switching power supply. To avoid input current peaks when plugging in or switching on a switched-mode power supply or when recharging the switching power supply after a failure of the AC input voltage or after a break during the operation of the input capacitor of a fast-clocked voltage regulator is eliminated and instead an output capacitor is routed to the output of the switching power supply. At the same time, the fast clocked voltage regulator itself is equipped with an active current limiter to protect it from destruction. The fast clocked voltage regulator is powered by a rectifier, which in turn is coupled ge via an input filter to the AC line voltage itself.

adversely at these concepts of active current limitation is the fact that the active current limitation implemented by circuitry must be, especially in integrated designs to an additional Test effort and to a potentially increased committee share of finished leads integrated circuits, because the number of components compared to a case without active Current limit significantly increased Has.

The 6,335,654 B1 discloses an inrush current control circuit having an input terminal, which is connected to a DC power supply, and with a Output terminal connected to a load capacitor. While When booting up a system, the power that charges the load capacity is limited. For this purpose, the control circuit generates a voltage ramp on the load capacitance, instead an abrupt application of a DC voltage. The voltage ramp results in a constant low current to charge the load capacity. For this purpose, an input-side voltage divider is provided, the by a potential at the middle tap point a serially connected transistor controls. Between the gate of the transistor and the output node is an RC series circuit arranged.

The U 5, 592, 072 A discloses a voltage regulator with a switching regulator and a downstream regulator controller, which is a linear regulator controller and a driver for comprises a serially connected transistor. Input side and output side in terms of of the transistor are two capacitors.

The Object of the present invention is an economical Concept for processing a supply voltage with voltage spikes and especially a cheaper Switching regulator, which is protected against voltage spikes to create.

These The object is achieved by a circuit for processing a supply voltage according to claim 1, a method for processing a supply voltage with voltage spikes according to claim 14 or a switching regulator for generating a regulated switching regulator output voltage Claim 15 solved.

The present invention is based on the finding that a voltage spiked supply voltage is processed so that the voltage peak is not simply cut off, but that the energy contained in the voltage spike is used to drive an output voltage to the desired output voltage level faster than if the voltage spike were simply cut off, d. H. the energy of the same would not be exploited. For this purpose, two capacitors are coupled via a controllable resistor. A controller for driving the controllable resistor is configured to set the controllable resistor to a high resistance value when the input voltage is below an input voltage setpoint. If the input voltage then exceeds the setpoint, ie if the voltage applied to the first capacitance equaling the input voltage becomes greater than the predetermined value, the resistance of the controllable resistor is reduced such that the input voltage remains constant or slower than that Supply voltage increases. Reducing the resistance of the controllable resistor causes charge to flow into the second capacitor. An energy of the voltage spike is thus not discarded, but is used to charge the second capacitor. Once the voltage on the second capacitor has reached a predetermined output voltage setpoint, the controllable resistor is controlled to a small resistance value, such that the two capacitors are substantially in parallel, thus acting as the resulting capacitance with a capacitance equal to the sum the capacitance values of the two single capacitors. By controlling the resistance when the input voltage has become greater than the input voltage set point, it is achieved that the charge acceptance of the first capacitor is reduced due to the supply voltage and due to the already present in the supply voltage peak voltage, so that more charge in the second capacity flows as in the case where the controllable resistor had a high resistance value.

at a preferred embodiment of present invention is the first capacitance at which the input voltage is applied, a relatively small capacity, while the second capacity at the the output voltage is applied, is a relatively large capacity. This is achieved that the input voltage quickly changes the input voltage setpoint achieved, which is preferably used to control to supply a switching regulator. By controlling the invention of the resistance, in such a way that it becomes ever smaller, becomes the Voltage rise of the input voltage limited and preferably even kept constant, allowing the controller to supply the input voltage gets no voltage spike and thus not dimensioned accordingly voltage-resistant got to. As soon as the controllable resistance is preferably monotonically dependent on the current input voltage is opened, the flows in the voltage peak energy contained in the large second Capacitor and is thus used to this second capacitor to charge quickly. Such a fast charging of the second capacitor is especially for the fast start-up of switching regulators, because the load, the switching regulator supplied, the output voltage receives, ie parallel to the second Capacitor is connected.

going the switching regulator then in the stationary mode over, so the controllable resistance becomes a small resistance value brought such that the two capacitors are connected in parallel and as a single capacity with the sum of the two capacity values the two single capacitors act.

Preferably is used as a controllable resistor, a transistor to Beginning, so if the input voltage is less than the input voltage setpoint is, is locked. When the input voltage reference is set by the Input voltage exceeded the transistor goes into the triode region, then when the output voltage has reached the output voltage setpoint, in the low impedance to be brought into a conductive state.

In the preferred embodiment of the present invention, the transition of the transistor from the stopband to the conducting region, that is, passing through the triode region of the transistor is defined and performed depending on the current input voltage and / or the current output voltage, such that the input voltage due to the "slow "Turning the transistor off from the value where the transistor was completely off does not significantly decrease. This ensures that the input voltage due to the small first capacitance quickly reaches its target value, which, when the circuit according to the invention is used in a switching regulator, is such that a control of the switching regulator already with a voltage greater than or equal to the input voltage Set point is, can work. For a load supply operation, the switching regulator is ready when the transistor is fully turned on and the two capacitances act together as a single capacitance. In a conventional switching regulator, therefore, the filter capacitor according to the invention is replaced by the circuit for processing a supply voltage with voltage peaks. The dimensioning of the individual capacitances to one another is selected such that the first capacitance is made so small that it is charged to the value required for the control in a predetermined delay after the switch-on, ie after the application of the input voltage to the charging circuit and diode, while the second capacitance is then set to equal the difference between a predetermined one Capacitance value for the filter capacitor, which is predefined due to the ripple of the output voltage, and the capacitance of the first capacitor is dimensioned.

preferred embodiments The present invention will be described below with reference to FIG the accompanying drawings explained in detail. Show it:

1 a schematic block diagram of the switching regulator according to the invention;

2 A preferred embodiment of the switching regulator according to the invention with a self-locking transistor as a charging switch;

3 an alternative preferred embodiment of the present invention with a normally-on transistor as a charging switch; and

4 a block diagram of the circuit according to the invention for processing a supply voltage with voltage spikes;

5 a flowchart for explaining the operation of the control device in 4 ;

6a a time course of the supply voltage with a voltage spike;

6b a time course of the input voltage without voltage spike;

6c a time course of the output voltage without voltage spike;

6d a time course of the resistance value of the controllable resistor; and

7 a block diagram of a known switching regulator, which is also known as a buck converter.

Before going into detail on the circuit for processing a supply voltage according to the invention, will be described below with reference to the 1 to 3 a switching regulator is shown in which the circuit according to the invention for processing a supply voltage instead of the sieve capacitor ( 102 in the 2 and 3 ) can be used. It should be noted, however, that the circuit according to the invention for processing a supply voltage not only with the in the 2 and 3 illustrated switching regulator concept but also in the 7 shown switching regulator concept can be used. It should also be noted that the circuit according to the invention can be used for processing any supply voltage which has voltage peaks which may not be passed on to an output side, or only reduced, since circuits connected on the output side are not specified for such voltage peaks.

1 shows a switching regulator according to the invention for generating a regulated output voltage U _SR using an input voltage U ₀ , the sake of clarity as a voltage source U ₀ 10, in series with an internal resistance R _i 12 is shown is shown. If an on / off switch 14 is actuated, lies between a first input node 110 and a second input node 112 a tension. Between the first entrance node 110 and the second input node 112 are to some extent serially to each other a controllable switch 106 and a network 101 which is a typical switching regulator network with a coil, a capacitor and a diode. Typically, such switching regulator networks have 101 also a shunt resistor to create a current path with a defined ohmic resistance.

The output voltage of the in 1 is controlled between a first output rail, which is also referred to as a positive output rail and the first output node 114 is identical, and a second output rail, which is also referred to as negative output rail and with the second output node 116 is identical to. At the in 1 embodiment shown is the positive output rail 114 can be brought to a first (positive) potential, while the negative output rail 116 can be brought to a second potential which is smaller than the first potential.

The network 101 comprising a coil, a capacitor and a diode, in particular, has the diode switched to be coupled to the positive output rail when the coil is coupled to the negative output rail or to be coupled to the negative output rail when the coil is coupled to the positive output rail. Typically, the diode is thus with an output rail, either the positive output rail 114 or the negative output rail 116 be coupled.

The desk 106 who in 1 is also referred to as a charging switch since it charges the capacitor in the network 101 is provided. In particular, it is connected in series with the diode. The charging switch further includes a control input 107 over which the charging switch 106 can be closed, which means that the input voltage does not drop at the charging switch, but at the network 101 , If the switch is open, the input voltage drops over the switch off, the network 101 is thus (apart from transient states) not charged with the input voltage.

About the control input 107 the charging switch can thus be opened or closed. The network 101 further comprises a capacitor, which is also referred to as a filter capacitor and is connected such that the regulated output voltage can be tapped at the capacitor. An essential part of the network 101 is also the inductance, which is also referred to below as a coil which is coupled on the one hand with the diode and on the other hand with the capacitor.

According to the invention comprises in 1 Switching controller shown further a switching device 109 which is controllable in response to a switching control signal from a controller 111 is provided, either the first outgoing rail 114 or the second output rail 116 via a coupling device 113 with the control input 107 to couple the charging switch. In particular, the charging switch 106 is formed such that it is closed due to a potential at an output rail, and is opened due to a potential at the other output rail. In the embodiments set out below is the charging switch 106 , as later still on the basis of 2 and 3 is shown as a self-conducting or self-blocking NMOS transistor. In this case, the charging switch 106 closed by the potential at the positive output rail (made conductive) while being opened by the potential at the negative output rail (is brought to idle). For professionals, it is clear that when the switch 106 is designed as a PMOS transistor, either corresponding potential transformations in the coupling device 113 can be performed, or that the potential at the negative output rail is used to close the switch, while the potential at the positive output rail 114 is used to switch 106 to open, so bring to idle mode.

Usually, the controller 111 comprise a regulator which may be of any desired design as long as it outputs a signal which results in opening and closing of the charging switch 106 the output voltage U _{SR has} a defined desired time characteristic. Usually, the controller 111 work in such a way that it produces an output voltage with an average value at the output 114 . 116 of the network 101 ensures regardless of which load is turned on or which current draws a load. The time course of the voltage U _SR will typically be a voltage waveform with some ripple around an average. The ripple can be brought into predetermined tolerance ranges by dimensioning of coil and capacitor, in many cases, a signal with a ripple around an average already sufficient, especially if the requirements for the switching regulator output signal, ie the voltage U _SR , are not as high as, for example, a lamp or something similar. On the other hand, in the case where the switching regulator as shown in FIG 1 is integrated, together with an integrated circuit to be supplied by it, on a substrate, the same through one of the output 114 . 116 The linear regulator connected downstream of the switching regulator must be supplemented in order to meet requirements for a voltage U _SR which corresponds to the possibilities of the in 1 exceed switching regulator shown to deliver at reasonable cost an output signal U _SR with appropriate specification.

The on / off control is made such that when the controller 111 determines that the charging switch should be opened, the switch 109 to the opening of the charging switch 106 assigned output rail, in 1 for example 116 , is connected while then when the charging switch 106 should be closed, because the capacitor in the network 101 should be recharged, the controller 111 the switch 109 activated to now the potential of the upper output rail 114 to the control input 107 of the charging switch 106 to pair.

2 shows a preferred embodiment of the present invention, in which the coil 100 with the diode 104 and the filter capacitor 102 is interconnected. When comparing 2 and 7 will be apparent that the coil 100 not anymore, like in 7 , with the cathode of the diode 104 is connected, but with the anode of the diode 104 connected is. Further, from the comparison of 2 and 7 seen that now the positive output rail 114 with an intermediate node 118 over which the charging switch 106 with the diode 104 is connected to the output node 114 , ie the first output rail, coincides. Furthermore, in 2 a shunt resistor R _Sh 101 plotted between the coil 100 and the negative output rail 116 is switched. A knot 103 between the shunt resistor 101 and the coil 100 is also called a lower intermediate node.

In particular, the charging switch 106 at the in 2 embodiment shown as a self-blocking NMOS transistor whose drain D with the first input node 110 is connected, and whose source S is first shorted to a bulk terminal B of the transistor, and its source S further directly to the positive output node 114 , So the positive output rail is coupled. The coupling device tung 113 from 1 includes at the in 2 shown embodiment, a zener diode D _Z 113a , a parallel-connected capacitor C _Z 113b and a ballast resistor R _Z 113c , The ballast resistor serves to connect the capacitor, which forms a voltage dropping device, in parallel 113b and the diode 113a over the first entrance node 110 to supply electricity, as will be explained later.

For the operation of in 2 shown switching regulator or even of switching regulators, the transient Cha characteristics of the memory elements coil L and capacitor C _{S are} important because the switching regulator, as in 2 is shown, a regulation of the output voltage by constantly turning on / off the charging switch 106 reached.

A Coil is characterized by the fact that the voltage across the Coil falls off, equal to the time derivative of the current passing through the coil flows. Accordingly, a capacitor is characterized in that the Current flowing through the capacitor, proportional to the time derivative is the voltage across the capacitor.

Regarding the Coil is also great Meaning that the voltage applied to the coil can jump, but that the current through the coil can not jump. Becomes Therefore, a coil is turned on, so a DC voltage (via a Source internal resistance) applied to the coil, so the voltage increases on the coil abruptly on the value of the applied DC voltage on and off then exponentially. At the same time, the current starts through the coil from its original value 0 to slowly increase, until at some point has reached a value equal to the quotient of the applied Voltage and the internal resistance of the source is. If this stationary time is reached, is also the over the coil has fallen voltage to 0.

Corresponding the reverse is the case with the capacitor. If the capacitor is switched on, so the voltage on the capacitor rises slowly, while the Power can jump through the capacitor.

The Elements coil and capacitor thus differ in that that the current can jump through the capacitor while the Tension on the coil can jump. The voltage on the capacitor on the other hand can not jump. The same applies to the electricity through the coil, he also can not jump. This leads to, since at the time of switching on, ie at the time t = 0, when at the time t = 0 the switching takes place, the coil as Idle affects while the capacitor acts as a short circuit.

following is referred to the switching off of the elements. Will be on a DC voltage charged capacitor from the DC voltage source separated, nothing happens. He holds the cargo when he is no over can discharge a resistor. Finds the switching off of the capacitor however over a short circuit, so leads This causes a very high current to flow, at which time the current and voltage arrow on the capacitor in opposite Show direction. So the capacitor works when switched off is, as a generator.

Becomes a coil through which a current flows is off again Of importance, how the coil is turned off. Will a coil of A current flows through it, so there exists a magnetic field in which energy is stored. When a coil is disconnected from the source, that between source and coil is an idle, so would the current actually brought abruptly to 0 through the coil. The in the However, magnetic energy stored in the coil must be able to drain away. Therefore introduces Turn off a coil by creating an open circuit between the source and the coil to a high voltage spike, which causes that on the switch forms an arc over which the stored in the magnetic field Drain energy to the source can. Because of this, a coil becomes theoretically infinite high voltage spikes that would be destructive, too avoid over discharge a resistor, a diode or a capacitor. This leads to a spike with reduced height. If a coil is switched off, Thus, the voltage arrow and the stream arrow again show different Directions. The coil thus works as a generator.

Is how it is in 2 (or 3 ), the coil is connected in series with a capacitor, so at the turn-on time, that is, when a voltage of 0 is changed to a voltage having a certain DC value, the capacitor is a short circuit and the coil is an open circuit. Thus, the entire voltage is initially applied to the coil and then decreases with increasing current through the series resonant circuit of coil and capacitor. At the time of turning off a voltage on the coil, the coil still generates a current that continues to flow for some time until the energy stored in the coil (or the energy stored in the capacitor) has drained away.

The functionality of the switching regulator in the steady state will be described below with reference to FIG 2 described. At the time when the switch 106 closed, is the full span U _Gl 10 . 12 at the diode 104 at. About the switch 106 a current flows to the capacitor 102 invites. The potential at the positive output rail 114 thus increases. This simultaneously leads to a current through the coil 100 ,

On average, the current is through the coil 100 equal to the current flowing to the load between the positive output rail 114 and the negative output rail 116 is turned on, is delivered.

After at the time of turning on the switch 106 the full voltage at the diode 104 is located, and after, as has been stated above, the voltage across the capacitor can not jump, but can jump on the coil, the potential of the negative output rail is simultaneously at the switch-on 116 raised. Immediately at the switch-on time of the switch 106 therefore increases the potential of the negative output rail, while the potential at the positive output rail 114 still in the previous state remains. Thus, at the time the switch is turned on, the voltage U _SR initially decreases slightly until the positive potential 114 through the charging current, via the charging switch 106 flows, has increased.

Eventually, the controller will 111 the switch 109 Press, so that he no longer with the positive output rail 114 but with the negative output rail 116 connected is. This causes the transistor 106 locks, and that no more power from the source 10 . 12 into the network of coil, diode and capacitor flows. This causes the coil to act as a generator, in that the potential at the node 103 drops and the coil emits a current through the diode 104 the capacitor 102 loads, so that after switching off the switch, the voltage U _SR still somewhat increases. This increase is slowed down by the current flowing across the load and becomes when the current drawn by the load becomes smaller than the current that the coil 100 at the node 114 supplies, to a voltage drop. This voltage drop is getting bigger, because the coil eventually no longer supplies enough electricity. Before the voltage becomes too small, the controller intervenes again and connects the switch 109 again with the positive output rail, which causes the switch 106 closed again and supplies power. Due to the above-described relationships, the output voltage U _SR decreases after switching on a little further until the positive node 114 of the capacitor has "retightened" and the voltage U _SR rises again. Eventually, the controller will again switch 109 Press that he is using the negative output rail 116 is connected so that the transistor 107 is locked, which in turn causes the voltage to rise somewhat further, due to the energy stored in the coil (and capacitor) leading to a current across the diode 104 and in the positive nodes 114 leads.

The in 2 shown switches 106 is designed as a self-blocking NMOS transistor in this embodiment. NMOS transistors then conduct when the voltage between the gate, so the control terminal 107 and the source is positive. NMOS transistors, which are self-conducting, already conduct at a voltage U _GS > 0. At the in 2 On the other hand, the self-blocking NMOS transistor shown must have the voltage between the gate 107 and the source (node 118 in 2 ) greater than the threshold voltage U _{th of} the transistor 106 be. For this reason, the coupling device comprises 113 from 1 in the 2 shown elements ballast resistor 113c , Parallel capacitor 113b and zener diode 113a ,

The zener diode 113a is configured to act as a short circuit from a certain breakdown voltage, which is fixedly configured, which means in other words, that at the parallel circuit of capacitor 113b and diode 113a always the breakdown voltage U _Z determined by the zener diode drops. Is the switch 109 with the positive track 114 connected, the voltage drop across the Zener diode voltage U _{Z is applied} directly between the gate and the source of the transistor. The control input 107 of the transistor is thus coupled to the positive output rail via the parallel circuit of capacitor and zener diode, to the effect that the transistor becomes conductive.

The only constraint for this is that the breakdown voltage determined by the Zener diode is greater than the threshold voltage of the transistor. However, this requirement is readily met because transistor threshold voltages are typically not very large and Zener diodes are present at any defined breakdown voltages. Furthermore, the Zener diode only has to be tuned very loosely to the transistor since the value of the breakdown voltage of the Zener diode only has to be greater than the threshold voltage U _{th of} the transistor. Whether she is much taller or not initially does not matter. Thus, the requirements for the threshold voltage of the transistor 106 and to the breakdown voltage of the diode 113a in terms of the required tolerance is very low, such that low-cost transistors on the one hand and low-cost diodes on the other hand can be used, which on the one hand greatly facilitates the integrability and on the other hand also plays a role in price. This is due in particular to the fact that in the production of integrated circuits quite on a wafer deviations with regard to the Threshold voltage of the transistor and with respect to the breakdown voltage of the diode may occur.

After this the requirements for the relationship between threshold voltage However, the transistor and breakdown voltage of the diode very loose are, can very high deviations can be accepted on a wafer, without a circuit becoming a committee. The inventive concept is thus particularly advantageous in that the reject rate and so that the cost of the final product can be kept low.

If the controller detects that the switch 106 is to be switched off again, it controls the changeover switch 109 such that it is connected to the negative output rail. This causes the potential at the gate 107 of the transistor to U _{SR is} smaller than in the on state. This is the voltage at the source of the transistor 118 , which is initially not affected by the switching, much greater than the potential at the gate terminal 107 , which immediately results in U _GS becoming negative and the transistor is turned off, ie brought into an idle state.

A special advantage of in 2 The inventive circuit shown is that a defined start-up of the circuit is ensured. For the defined start it is especially irrelevant whether the changeover switch 109 at the time of switching on with the positive output rail 114 or the negative output rail 116 connected is. If the switch 109 In particular, as a multiplexer of transistors or as an inverter is realized, it is undefined whether the switch 109 with the upper rail 114 or the lower rail 116 is connected, if it is assumed that before turning on the entire in 2 circuit shown in a de-energized state was that all potential in the circuit are at the value 0.

From this output state, in which all potentials are equal to 0, it is assumed below to a start-up of the circuit in 2 display. After all potentials are equal to 0, so is the potential between the gate 107 and the source 118 of the switch is 0. Since the transistor is a self-locking transistor, the switch is initially disabled. Becomes a positive tension 10 . 12 between the first entrance node 110 and the second input node 112 which will typically be the ground node, first the capacitor C _z 113b charged via the series resistor R _Z until the voltage across the capacitor C _Z 113b and the parallel Zener diode D _Z, the threshold voltage U _{th of} the transistor switch 106 reached. For this purpose, as it has been stated, it is irrelevant whether the switch 109 with the node 114 or the node 116 connected is.

Is the switch 109 with the node 114 connected, the voltage U _Z falls on the capacitor 113b anyway directly between the gate 107 and the source 118 of the transistor.

In contrast, the switch 109 with the negative rail 116 connected, so the capacitor charges 113b also on the resistor R _Z on. The potential at the node 118 , Which determines the source potential of the transistor, however, is not initially brought out of its 0 value, since there is no charge of the capacitor 102 through which the potential 118 could be raised, because the switch 109 with the negative rail 116 connected is.

In both cases of the switch 109 Thus, the gate-source voltage increases from a value of 0 at the time of switching on the source 10 . 12 by pressing the switch 14 in 1 to a value equal to the threshold voltage of the transistor. Once this is the case, the drain-source path of this switch 106 conductive, and the capacitor C _S is (regardless of the position of the switch 109 ) charged. The charging current for the capacitor 102 (C _S ) flows simultaneously through the coil 100 to the mass 112 , This leads directly to the fact that the output voltage U _SR due to the increase of the potential at the positive output rail 114 opposite the potential at the negative output rail 116 increases. This charging process continues until the voltage on the capacitor C _{S is} the voltage across the Zener diode minus the threshold voltage of the switch 106 reached.

Would the switch 109 with the positive output rail 114 be connected, the charging process would continue to continue, as the transistor 106 is open. Is the switch 109 against it with the negative rail 116 connected, the switch is then closed again when the potential between the gate and source is equal to the threshold voltage. Assuming that the potential at the source is equal to U _SR , and assuming that the potential at the gate of the transistor is equal to the voltage drop across the Zener diode U _Z , then there will be a value U _SR at the output of the switching regulator in which the transistor blocks again, as the difference between the voltage at the Zener diode and the threshold voltage. If nothing else were done, the transistor would close again and the output voltage would remain at U _SR = U _Z - U _th .

By dimensioning the zener diode 113a is this "persistence value" of the output voltage U _SR relatively freely selectable. In a preferred embodiment of the invention, the Zener diode is dimensioned so that U _SR becomes so large (if in the "unfavorable case" of the changeover switch 109 to start up with the negative rail 116 connected), that the control 111 , which is preferably powered by U _SR , can already work. The voltage U _SR is thus made so large by dimensioning the Zener diode (and the threshold voltage) that when this voltage is applied to the controller, the states at the nodes in the controller are already defined.

The controller will therefore detect a value U _SR and compare it to a limit. Once the controller determines that U _SR has reached the predetermined state at power-up, the controller is 111 effective to ensure that the switch 109 with the positive track 114 connected is. If this was already the case during startup, then the controller performs 111 here no change of the switch 109 by. However, this was not the case, so at the time of startup, the switch was 109 , as in 2 shown with the negative rail 116 connected, the controller is the switch 109 in such a way that now no longer the negative rail 116 to the control terminal 107 is coupled, but that the positive rail 114 to the control input 107 is coupled. This is the potential of the positive output rail at the anode of the Zener diode. As a result, the voltage at the filter capacitor C _{S increases} 102 due to the charge switch 106 flowing charging current continues until a desired output voltage at C _S is applied. Now follows the actual voltage regulation by the changeover switch 109 through the controller 111 is switched back and forth to turn the switch T on and off, and thereby the output voltage U _SR according to voltage and current demand of a load (in 2 not shown).

In the 2 The circuit shown is thus particularly advantageous in that it starts up in a defined manner, wherein it is irrelevant at the time of startup, in which position the switch 109 stands. Further, it is preferable to control 111 dimensioned so that it already works defined when the output voltage equal to the zener diode voltage U _Z minus the threshold voltage of the transistor 107 is. Thus, without own supply for the control 111 the startup process continues to be defined. This leads in particular to a low-cost circuit, since no special start-up measures must be taken, except to ensure that the changeover switch 109 with the positive track 114 connected is. After no special measures are required and in particular voltage checks with regard to the control 111 etc., the startup process takes place very quickly.

It should also be noted that the voltage increase by appropriate dimensioning of the resistor 113c and the capacitor 113b can be dimensioned very quickly. So it is anyway preferred, the resistance 113c relatively small dimensions, so that the power loss generated by it does not come in significant sizes. Furthermore, it is preferred that the capacitor C _Z , anyway only for stabilizing the Zener diode 113a or to the (smaller) junction capacitance, also small to dimension so that it is quickly charged to the voltage U _Z. The startup thus takes place in such a way that no time constants would have to be taken into account, which would significantly slow down the startup of the switching regulator.

3 shows an alternative embodiment, different from 2 differs in that the transistor T 106 is now a self-conducting NMOS transistor, and that the coupling device 113 from 1 , in the 2 through the elements 113a . 113b . 113c was realized in 3 through a simple coupling 113d is realized. While in 2 the potential at the first output rail or the second output rail via the coupling device 113 with a voltage drop is applied in 3 the potential at the first output rail 114 or the second output rail 116 by a simple connection directly to the control terminal 107 of the switch 106 coupled. The transistor 106 is designed as a normally-on n-channel MOS-FET or n-channel JFET. The wiring of the transistor, the charging switch 106 with respect to drain and source is identical to the case of 2 ,

The voltage source or Spannungsabfalleinrichtung from the Zener diode with the series resistor and the parallel capacitor of 2 by contrast 3 for an immediate potential coupling from an output rail to the control terminal 107 and the switch to reach. Preferably, the threshold voltage of the transistor UTh is dimensioned to be equal to the Z voltage of the Zener diode minus the threshold voltage of the normally-off MOSFET in the first case, so that the output voltage U _SR reaches such a value in the startup process, through the control 111 is already operable to the switch at a certain time of booting 109 to be able to steer in such a way that it is defined with the positive rail 114 connected is.

Threshold voltages in normally-on NMOS transistors or N-JFET transistors are defined as defining a negative voltage between the gate and source of the transistor at which the transistor 106 just locks. Voltages that are higher than the negative threshold voltage will then cause the transistor to conduct, while voltages even more negative than the threshold voltage will cause the transistor to turn off.

In the following, the booting process of in 3 illustrated circuit explained. Again it is assumed that in the initial state all potentials are equal to 0. Thus, the switch conducts (it is self-conducting), which leads to the potential at the output rail 114 is raised. It should be noted that at the time of switching on, both the node 114 as well as the knot 116 be raised to the DC potential, but that the potential difference between tween the node is equal to zero. The two nodes 114 and 116 are abruptly raised because the capacitor, as has been stated, acts as a short circuit at the time of turning on a DC voltage. Only then, when charge on the knot 114 over the charging switch 106 is delivered, a potential difference arises between the node 114 and the node 116 which causes the output voltage U _{SR to increase} from 0 V to values greater than 0 V. After switching on the voltage thus becomes the capacitor 102 charged via the transistor, which is self-conducting.

Would the changeover switch 109 , which may again be made such that its initial state is undefined, with the positive rail 114 be connected, the charging process would continue to be continued, since the gate and source of the transistor are short-circuited, such that U _{GS is} equal to 0 V, which always means a short circuit between the drain and source in the normally-on transistor. Would, however, the switch 109 be connected to the negative rail, the state of charge would eventually stop because the source potential, so the potential of the positive rail 114 continues to increase. If the source potential is greater than the threshold voltage, the transistor is closed and the output voltage U _SR no longer increases. Therefore, as has been stated, the threshold voltage of the transistor 106 chosen such that an output voltage U _SR then applied to the output is already sufficiently high to the control 111 , which is supplied with the voltage U _SR to let work defined, so that then the switch 109 from the negative rail 116 separates and with the positive track 114 connects, so that the transistor is re-opened, so that the charging process of the capacitor 102 can continue until the controller 111 goes into normal operation and makes a switch control due to an actual desired output voltage U _SR .

In summary, the boot process is in 3 thus such that first the switch T is conducting. After applying a positive input voltage U _Gl to the input of the circuit, the capacitor 102 charged via the switch T and the inductor L until the voltage across the capacitor C _{S reaches} the threshold voltage of the switch T (when the switch 109 with the negative rail 116 was connected). Then the voltage at the output of the circuit U _{SR is} already so large that a controlled operation of the regulator is guaranteed. This now switches the changeover switch s _w so that the positive rail of the output voltage U _SR is present at the gate of the switch. As a result, the voltage continues to increase at C _S until the desired output voltage is applied to C _S. Now follows the actual voltage regulation. By switching the changeover switch s _w , by the control circuit 111 the switch T is switched on and off, whereby the output voltage U _{SR is} regulated.

Regarding the dimensioning of the threshold voltage of the transistor 106 in case of 3 or the threshold voltage of the transistor 106 and the Zener diode voltage U _Z of 2 It should be noted that these values are dimensioned so that the controller 111 in the automatically without intervention in the circuit and in the assumed position of the changeover switch 109 on the negative track 116 reached maximum output voltage value U _SR , which results automatically without any intervention, is already so large that the controller can be supplied with U _SR .

Will the controller 111 or the threshold voltage or the threshold voltage and the Zener diode voltage are dimensioned such that the output voltage U _SR , which is "automatically" reached, is already slightly higher than the voltage at which the controller 111 defined works, so can the controller 111 also be configured to immediately, if it can work defined, so if the input voltage is sufficiently large, the switch 109 to the top, so to put on the positive output rail. Thus, the switching time is not limited to the fact that always the maximum automatically achievable output voltage must be present in order to switch the switch. Instead, it is necessary that at some point in the startup process it is ensured that the changeover switch with the positive output rail 114 connected is.

If, on the other hand, a controller were used that works independently of U _SR , it could from the beginning, so already z. B. at the time of switching on the DC voltage U _Gl be ensured that the switch 109 with the positive track 114 connected is. Due to the simplicity of the circuit, however, it is preferred that the controller 111 is supplied from the output voltage U _SR , in which case in the dimensioning of the transistor and possibly the diode is rather gone to the limit, to the effect that z. B. 90% of the maximum adjusting output voltage U _SR until the changeover of the changeover switch 109 be used so that the requirements for the control as low as possible, typi cally circuits that are already operational with lower voltages, with the other voltage levels of in 2 and 3 shown circuits are rather incompatible, so that as high as possible automatically adjusting output voltage U _{SR is} preferred.

In the in the 2 and 3 the embodiments shown, the coil is not, as in 7 , coupled to the cathode of the diode, but to the anode. This has the advantage that when the switch 109 with the positive track 114 is connected, the potential generated between the gate and source of the transistor is not affected by the transient characteristics of the coil. However, if sufficient security and sufficient strength of the transistor has been used, in another embodiment the coil may also be coupled to the diode on the cathode side, such that the switching device 109 if you have the positive rail with the transistor gate 107 connects, so to speak, the coil to the gate of the transistor either directly coupled or short-circuited or indirectly coupled via a voltage drop device.

One Arranging the coil on the anode side to the diode also has the advantage that then the input voltage of the circuit only by the dielectric strength limited by the three elements switch, diode and choke. Further supplied itself, the switching regulator and runs up defined. The jumping the common mode potential of the output voltage between the positive and the negative rail of the input voltage depending on the state of the Switch T is for the load is not primary Meaning, because she does not notice this jumping because of the load only the potential difference between the upper output rail and the lower output rail experiences, but not the "absolute" potential the positive output rail or the negative output rail for themselves.

As already stated, the startup time itself is determined by the sizing of R _Z and C _Z in the case of 2 shown circuit variant or by the maximum drain current I _D , the inductance of the inductor L and the capacitance of the sieve capacitor C _S in the case of the circuit variant of 3 Are defined. An additional start-up delay to ensure the voltage supply of the regulator is not needed.

It should also be pointed out that, according to the invention, the input voltage range, which is usually limited to an input voltage ratio of U _{E, max} / U _{E, min} , can be increased to at least 50: 1 according to the invention, as long as the dielectric strength of the diode , The coil and the switch is carried accordingly, since these elements are acted upon by the input voltage. In contrast, neither the capacitor nor the switch nor the controller are subjected to such high voltages, so that for these elements, the maximum input voltage need not be taken into account, which in turn leads to a cheaper and more flexible Schaltreglerkonzept according to the present invention.

4 shows a circuit according to the invention with a first input node 41 and a second input node 42 between which a first capacity 43 is switched. In the 4 The circuit shown also has a first output node 44 and a second output node 45 between which a second capacity 46 is switched. In between the entrance nodes 41 . 42 a supply voltage with a voltage peak is applied, which in 4 only model as a voltage source 47a with an internal resistance 47b outlined. Between the first entrance node 41 and the first output node 44 is also a controllable resistor 48 switched by a control device 49 is controllable. The control device 49 preferably receives an input voltage setpoint U _esoll and an output voltage setpoint U _asoll . Furthermore, the control device receives 49 in a preferred embodiment of the present invention, the current value of the input voltage U _e and the current value of the output voltage U _a . It should be noted that the setpoint values in the control device 49 can be permanently programmed. It should also be pointed out that the two setpoint values U _setpoint and U _{setpoint may be the} same or different. Should they be different, it is preferred that the output voltage setpoint U _{asoll is} greater than the input voltage setpoint U _desired .

Is the approximate time course of a voltage peak of the supply voltage 47a previously known, which may be in certain applications, such as in a switching regulator, so comes the controller 49 without the current input voltage value U _e or U _a off. Instead, it can have a corresponding state measure machine or some other device to resist 48 in accordance with a predetermined timing from a high resistance preferably monotonically decreasing to a low resistance value durchzusteuern. If the time profile of the voltage peak is not immediately known, then it is preferred to carry out a nominal value / actual value comparison of the input voltage and / or the output voltage.

Regardless of whether actually in the control device 49 a setpoint / actual value comparison or a preset time sequence for controlling the resistance, etc. is performed, the control device is designed to, in a first case in which the input voltage U _{e is} less than the predetermined input voltage setpoint U _e , the controllable resistance to drive so that the controllable resistor has a first high resistance value R ₁ , as in a block 52 in 5 is shown. The comparison of the actual input voltage U _e and the input voltage setpoint U _esoll is by a block 51 in 5 presented, which is then addressed when in a starting block 50 the control device 49 from 4 is activated and has provided the appropriate setpoints. The controller is further configured to be monotonic in a second case in which the input voltage is greater than the predetermined input voltage setpoint and in which the input voltage is less than the output voltage setpoint, which will normally be the case rising to the voltage peak, the controllable resistance 48 from 4 to drive so that the controllable resistor has a second lower resistance, as in a block 54 is shown. Whether the second case exists or not is determined by a decision block 53 certainly. In the second case, the "closing" of the controllable resistor 48 acting on a mean resistance value such that a charge absorption of the first capacity 43 at least reduced, and that more charge in the second capacity 46 flows as in the case where the controllable resistor has the high resistance, that is, in the first case.

The control device 49 is further adapted, in a third case, by a decision block 55 is determined, and that is that the output voltage is greater than or equal to the output voltage setpoint, the controllable resistance 48 to drive it to have a third low resistance so that the first and second capacitors are connected in parallel across the third low resistance value, as in a block 56 in 5 is shown. Will, as has already been stated, as controllable resistance 48 a transistor T _{1 is} used as it is in 4 is shown, the high resistance value R ₁ from block 52 in 5 is achieved by blocking the transistor, while the low resistance R ₃ is ensured by turning on the transistor T ₁ using a corresponding gate voltage. The average resistance is achieved by operating the transistor in the triode region.

Hereinafter, referring to the 6a to 6d on the functionality of in 4 indicated circuit shown. It should be emphasized that the time courses in 6a to 6d are merely sketchy to explain the functionality of the circuit according to the invention. Thus, actual circuit patterns can be greatly different from those in the 6a to 6d deviate from shown time courses. In 6a is an example of a supply voltage time course with a voltage spike 60 shown. Furthermore, in 6a the input voltage setpoint U _esoll and the output voltage setpoint U _{asoll are plotted} . Referring to 6b follows the input voltage in the period between t = 0 to t = t _{1 of} the supply voltage, since the resistance of the controllable resistor 48 who in 6d with R ₁ is very high. Then, at time t ₁ , the resistance value is reduced as it is by a falling edge at 61 in 6d is shown. This means that charge that is moved due to the supply voltage, no longer exclusively in the first capacitor 43 flows, but via the controlled resistor in the second capacity 46 , This causes the output voltage at the second capacitor 46 increases between the time t ₁ to a time t ₂ , as more and more charge in the second capacity 46 flows. Due to the opening of the controllable resistor 48 between t ₁ and t ₂ , the input voltage U _e no longer rises in exactly the same way as the voltage spike but considerably slower. With a corresponding control by the control device 49 It may also be effected that the input _voltage remains substantially at the input voltage setpoint U _desired .

In any case, due to the 6b and 6c to see that the voltage spike 60 in 6a neither through the input voltage U _e nor to the output voltage U _a , but that in the voltage peak 60 stored energy completely for fast charging of the usually large second capacitor 46 is used advantageously.

Depending on the control of the controllable resistor 48 The timing of the input voltage and the output voltage can be influenced. If the resistance between time t ₁ and time t _{2 is} not monotonic with respect to a time-varying second resistance value, but only at time t ₁ to a resistance value 62 on which the controllable resistor would then remain until time t ₂ , this would result in the input voltage dropping slightly at time t ₁ compared to the input voltage setpoint, while the output voltage would make a jump at time t ₁ . The input voltage would then increase again, exceed the input voltage setpoint, and eventually reach a value determined by the value of R ₂ 62 in 6d is predetermined.

According to the invention, the dimensions of R ₂ 62 or, if a monotonically decreasing resistance value profile is desired, the shape and slope of the edge 61 chosen such that the input voltage is preferably not more than a tolerance range of z. B. 20% below the input voltage setpoint increases, since, as will be explained later, the input voltage U _{e is} used to the controller 111 of the 1 to 3 to supply with voltage, so that a starting process of in the 1 to 3 shown switching regulator and runs fast. For this, it is not critical whether the input voltage rises steeply or flatly between time t ₁ and time t ₂ , as long as the voltage requirements of the controller 111 , which is supplied with the input voltage U _e , are met. Of course, the flank 61 in 6d also be replaced by a step function or by any other example square or cubic continuously or stepwise shaped function, depending on the existence of the circumstances. Always be sure that the voltage spike 60 does not fully penetrate both the input voltage and the output voltage.

It is further preferred to dimension the two capacitors so that they are capable of all the energy of the voltage peak 60 take. If a particularly high voltage peak comes with a lot of energy, and the capacitors are not sufficiently dimensioned, they nevertheless already achieve a mitigation of the voltage peak, although a complete elimination of the voltage peak is preferred. When the in 4 shown circuit according to the invention as a capacitor 102 in the 2 This is typically not a problem, since the capacitance value generated by the parallel connection of the two capacitors 43 and 46 is realized, anyway assumes considerable magnitudes, which are readily sufficient to "absorb" even very high voltage spikes.

That in the 2 and 3 shown circuit concept is particularly problematic and can be particularly well due to in 4 circuit shown to be supplemented, especially for large input voltages 10 . 12 in 2 or 3 at the time when the switch 109 during startup is driven to the positive output rail 114 Being connected, this causes the charging current for the capacitor 102 that equals the current through the coil 100 is, as it were, turned off. In case of 2 In addition, due to the fact that the Zener diode is operated in breakdown, the diode breakdown current is also through the coil 100 to the mass 112 flows. So is the switch during the startup process after reaching the maximum automatically adjusting voltage U _SR 109 it is because of in the coil 100 stored energy can be generated a voltage spike, which is the node 103 at the time of the peak makes strong negative, this additional difference is only compensated by the fact that the current from the coil via the then operated in the flow direction diode 104 the capacitor 102 invites. This voltage spike in U _SR could lead to the control without any further precautions 111 if destroyed by U _SR , it will be destroyed.

To prevent this, the capacitor becomes 102 in the 2 and 3 according to the invention by the in 4 shown arrangement of two capacities including controllable resistance and velvet control device replaced. The capacitor 43 is referred to as auxiliary capacitor C _H , while the actual sieve capacitor is now referred to as C ' _s . The transistor 48 is disabled as long as the voltage U _{SR is to be} smaller than the output voltage U _a . The capacitor C _H is chosen small, so that it is charged quickly to the target voltage. When its voltage U _{SR reaches} the desired value, that is to say the input voltage setpoint, the transistor opens and conducts the excess energy from the inductor L to the significantly larger filter capacitor C ' _s . Is, as it is preferred, in 4 shown circuit in a switching regulator of the configuration of 2 and 3 used, the controller is 111 powered by U _SR (and not U _a ), so can the controller 111 be very quickly functional, since the capacitor C _{H is} chosen to be small. This, ie by controlling the switch, ensures that C ' _{s will} continue to be charged until the voltage U _{a reaches} the output voltage setpoint.

At the in 4 shown embodiment for the input voltage setpoint U _e, a voltage value usefully chosen, the of the configuration of the controller 111 depends. It is preferably chosen so that it is at least as large that the controller 111 can operate with a voltage equal to the input voltage reference. The output voltage setpoint is passed through be given the general specification of the switching regulator, since the output side, a load is arranged, the output voltage at the capacitor 46 "Sees". Preferably, the transistor is thus only fully turned on and used for parallel connection of the two capacitors, when the output voltage U _{a is} exactly as large as the input voltage U _e .

As stated above, the first capacitor becomes 43 much smaller dimensioned than the second capacitor 46 , At the time when the controllable resistance 48 is not at the low resistance, thus acting in comparison to the specified sieve capacity 102 from 2 or 3 For example, reduced sieve capacity, which initially, if the controllable resistance 48 has a high resistance equal to the value of C _H and then corresponding to the reduction of the resistance value of the controllable resistor 48 becomes larger and larger until finally it is equal to the sum of C _H and C ' _s . In the booting process of the switching regulator, so if the in 1 shown source 10 over the switch 14 is turned on, acts as a sieve capacity of the switching regulator only the small capacity C _H. The large ripple of output voltage actually associated with a small Siebkapazität plays here, however, no role, since the switching regulator is already ready to power a load when the output voltage has the output voltage setpoint, in which case the transistor T _{1 is} fully turned on and thus the both capacitances C _H and C ' _s are connected in parallel.

However, it is advantageous that the capacitance C _H , because it is smaller than the capacitance C ' _s , quickly comes to a voltage level at which the controller 111 of the 2 and 3 that like it in the 2 and 3 is shown with U _SR , ie U _e in 4 , is supplied.

As for the 2 and 3 has been executed, it is thus preferable to make the input voltage setpoint Ue- soll so large that the controller 111 of the 2 and 3 when booting, when it is operational, it ensures that the changeover switch 109 with the positive output rail 114 connected is. Furthermore, it has already been noted that this voltage level at which the controller 111 can work, less than the voltage applied to the zener diode 113b in 2 decreases, less the threshold voltage of the transistor must be. At the in 3 The embodiment shown must have the voltage level at which the controller 111 can work greater than the threshold voltage U _{th of} the transistor 106 be when the controller with U _SR , so at 4 to work with U _e .

Preferably, the control device is also 49 in 4 that the controllable resistance 48 controls, designed to be supplied with U _e . This means that the control device 49 already defined can work if their supply voltage U _e or U _{SR has reached} a certain value. In this embodiment, the controller would 49 be designed to, when it operates defined, immediately the resistance value of the controllable resistor 48 then it is inherently clear that the input voltage U _{e has reached} an input voltage threshold.

As stated above, the in 4 shown advantageous in that it limits the voltage U _SR , that produces a voltage U _e and U _SR , which has no voltage spikes, and that at the same time the charging of the capacitor C ' _s at the output of the circuit is accelerated, in that the Energy of the voltage peak for charging the (large) filter capacitor 46 is exploited. Thus, at the same time the startup of a switching regulator, the in 4 shown circuit instead of the filter capacitor has accelerated, since the necessary for the control of the switching regulator supply voltage is reached quickly and is not subjected to voltage spikes.

Claims (19)

Circuit for processing a supply voltage ( 47a ) with a voltage peak ( 60 ) to obtain an output voltage (U _a ) with reduced or eliminated voltage peaks, comprising: a first capacitor ( 43 ) between a first node ( 41 ) and a second node ( 42 ), wherein an input voltage (U _e ) between the first node ( 41 ) and the second node ( 42 ) due to the supply voltage ( 47a ) can be applied; a second capacity ( 46 ) between a third node ( 44 ) and a fourth node ( 45 ), wherein the output voltage (U _a ) between the third node ( 44 ) and the fourth node ( 45 ), the first capacity ( 43 ) has a smaller capacity value than the second capacity ( 46 ); a controllable resistor ( 48 ) between the first node ( 41 ) and the third node ( 44 ); and a control device ( 49 ) for driving the controllable resistor, wherein the control device ( 49 ) is designed in a first case ( 52 ), in which the input voltage (U _e ) is less than a predetermined input voltage setpoint (U _esoll ), the controllable resistance ( 48 ) in such a way that the controllable resistor has a first high resistance value, in order in a second case ( 54 ), in which the input voltage (U _e ) is equal to or greater than the predetermined input voltage setpoint (U _esoll ), and in which the output voltage (U _a ) is less than the predetermined output voltage setpoint (U _asoll ) to control the controllable resistor such that the controllable resistance has a second lower resistance value, so that a charge acceptance of the first capacitance ( 43 ) is at least reduced and more charge in the second capacity ( 46 ) flows as in the case where the controllable resistance ( 48 ) has the high resistance, and in a third case ( 56 ), in which the output voltage (U _a ) at the second capacitor ( 46 ) is equal to or greater than the predetermined output voltage setpoint (U _asoll ), the controllable resistance ( 48 ) so that it has a third low resistance so that the first capacitance ( 43 ) and the second capacity ( 46 ) are substantially in parallel, wherein the first resistance value (R ₁ ) is higher than the second resistance value (R ₂ ), and wherein the second resistance value (R ₂ ) is higher than the third resistance value (R ₃ ). The circuit of claim 1, wherein the first resistance value represents essentially an open circuit, and the third resistance value essentially represents a short circuit. Circuit according to Claim 1 or 2, in which the control device ( 49 ) is configured to reduce the second resistance monotonically decreasing from the first resistance value to the third resistance value in the second case. Circuit according to one of the preceding claims, in which the control device ( 49 ) is configured to monotonically decrease the second resistance in the second case such that the input voltage (U _e ) does not move out of a tolerance range that extends around the predetermined input voltage setpoint. A circuit according to claim 4, wherein the tolerance range chosen is to increase by plus or minus 20% by the predetermined input voltage setpoint to extend around. Circuit according to one of the preceding claims, in which the first capacitor ( 43 ) is dimensioned to be smaller than a limit capacitance, which determines an upper limit for a period of time how fast the input voltage (U _e ) reaches the predetermined input voltage setpoint when a supply voltage having a predetermined rise rate is used. Circuit according to one of the preceding claims, in which the first capacitor ( 43 ) is dimensioned to have a capacitance at least as large that at the predetermined input voltage setpoint, a predetermined amount of charge on the first capacitance ( 43 ) is stored. A circuit according to claim 7, wherein the predetermined amount of charge is so large that, by the input voltage (U _e ), the controller ( 49 ) for controlling the resistance value or a controller ( 111 ) is operable to control the input voltage. Circuit according to one of the preceding claims, in which the second capacitor ( 46 ) is at least so large that it can store a charge that is due to a voltage spike ( 60 ) of the supply voltage in the first node ( 41 ) without the output voltage generating a margin around the output voltage setpoint. A circuit according to claim 9, wherein the tolerance range extends by plus or minus 20% around the output voltage set point. Circuit according to one of the preceding claims, in the output voltage setpoint is equal to the input voltage setpoint is. Circuit according to one of the preceding claims, in which the control device is designed to determine, in order to determine whether the first, the second or the third case, the input voltage (U _e ) with that in the control device ( 49 ) and the output voltage (U _a ) with the in the Steuerungsein direction ( 49 ) to compare stored output voltage setpoint. Circuit according to one of Claims 1 to 11, in which the control device ( 49 ) is adapted to transition from the first case to the second case on the basis of a temporal state sequence at a time and to transition from the second case to the third case at a later time, wherein a time period between the second time and the first Time in the control device ( 49 ) is programmed. Method for processing a supply voltage ( 47a ) with a voltage peak ( 60 ) to obtain an output voltage (U _a ) with reduced or eliminated voltage spikes, comprising the steps of: charging a first capacitance (C _H ) between a first node ( 41 ) and a second node ( 42 ), wherein an input voltage between the first node ( 41 ) and the second node ( 42 ) due to the supply voltage ( 47a ) is produced; Loading a second capacity ( 46 ) between a third node ( 44 ) and a fourth node ( 45 ), wherein the output voltage (U _a ) between the third node ( 44 ) and the fourth node ( 45 ), the first capacity ( 43 ) has a smaller capacity value than the second capacity ( 46 ); Controlling a resistance between the first node ( 41 ) and the third node ( 44 ) in such a way that in a first case in which the input voltage (U _e ) is less than a predetermined input voltage setpoint (U _esoll ), the controllable resistor has a first high resistance value (R ₁ ), that in a second case, in where the input voltage (U _e ) is equal to or greater than the predetermined input voltage setpoint (U _esoll ), and in which the output voltage (U _a ) is less than the predetermined output voltage setpoint (U _asoll ), the controllable resistance is a second lower one Resistance value, so that charging the first capacity ( 43 ) is at least reduced, and that loading the second capacity ( 46 ), compared to the case where the controllable resistance ( 48 ) has the high resistance, and that in a third case, in which the output voltage at the second capacitor ( 46 ) is equal to or greater than the predetermined output voltage setpoint (U _setpoint ), the controllable resistor has a third low resistance value (R ₃ ), such that the first and second capacitors (U 43 . 46 ) are connected substantially in parallel, wherein the first resistance value (R ₁ ) is higher than the second resistance value (R ₂ ), and wherein the second resistance value is greater than the third resistance value (R ₃ ). Switching regulator for generating a regulated switching regulator output voltage (U _a ) using an input voltage ( 10 ), comprising: a controllable switch ( 106 ); a network ( 101 ) with a diode ( 104 ), an inductance ( 100 ) and a screening device ( 102 ); a switching regulator control ( 111 ) for regulating the output voltage by periodically actuating the controllable switch ( 106 ) to control the output voltage, and wherein the screening device ( 102 ) has a circuit according to one of claims 1 to 13. Switching regulator according to claim 15, in which the supply voltage from a supply voltage generating device to the first capacitor ( 43 ), wherein the supply voltage generating device is the coil ( 100 ) and the diode ( 104 ) of the network, and wherein the voltage at the second capacitance ( 46 ) represents the output voltage of the switching regulator. Switching regulator according to Claim 15 or 16, in which the switching regulator control ( 111 ) is adapted to the controllable switch based on the input voltage (U _e ) at the first capacitance after switching on the switching regulator input voltage (U _e ) 10 ) to control. Switching regulator according to one of Claims 15 to 17, in which the switching regulator control device ( 49 ) from the input voltage (U _e ) at the first capacitor ( 43 ) is available. Switching regulator according to one of Claims 15 to 18, in which the switching regulator control ( 111 ) is adapted to control a change-over switch, wherein the change-over switch with a voltage coupling device ( 113a . 113b . 113c ) is connected to either a potential at a positive output rail ( 114 ) or a potential at a negative output rail ( 116 ) with a control input of the controllable switch ( 106 ), the controllable switch ( 106 ) is formed as a transistor, and wherein the control input ( 107 ) of the controllable switch ( 106 ) a gate terminal of the transistor ( 106 ).