DC Chopper Converter

The present invention generally relates to DC chopper converters. More specifically, the invention relates to a DC chopper converter for controlling the amplitude and direction of a load current in a bidirectional control bridge of the type having four bridge arms, each arm having a switching element and a free-wheeling diode coupled in parallel, a load being coupled between the junctions of the arms such that an H-configuration is formed between two power supply terminals, switch elements that lie diagonally opposite each other across the bridge being periodically driven or operated so as to control the load current supplied to the load.

A DC chopper converter of the type described above is disclosed in European patent EP 0 024 300 and corresponding U.S. Pat. No. 4,314,325. The disclosures of these patents are fully incorporated herein by reference.

In the disclosed device, a circuit is provided for controlling the amplitude and direction of a load current in a bidirectional direct current control bridge of the type having four bridge arms, each arm having an electronic switching element electrically disposed therein, and a load, illustratively an ohmic-inductive load, connected to the junctures of the bridge arms. The load current is conducted to the load by applying energizing pulses to each of the switching elements in a pair of diagonally disposed bridge arms, the respective pulse energizing signals being out of phase with respect to one another. Load current is responsive to the duration of the interval during which the switching elements in the diagonal pair of bridge arms are simultaneously conducted.

Additionally, circuitry is provided for pulse energizing the electronic switching elements in the second diagonal pair of bridge arms, while inhibiting conduction of the pulse energizing signals to the first diagonal pair, in response to a change in polarity of a control signal. The amplitude of the control signal governs the pulse-width ratio of the pulse energizing signals, and therefore, the duration of the intervals of simultaneous conduction of the switching elements in a diagonal pair. Circuitry is provided for inhibiting the conduction of all pulse energizing signals for a predetermined interval of time in response to the change of polarity of the control signal.

U.S. Pat. No. 4,314,325 discloses other prior art devices for controlling load currents by the use of a four-arm bridge including a publication entitled "Siemens-Zeitschrift 43 (1969), No. 5 and U.S. Pat. No. 3,260,912.

The device stated to be disclosed in the Siemens-Zeitschrift publication comprises a basic four-arm bridge circuit containing electronic switching elements in which a first pair of electronic switching elements which are on diagonally disposed arms of the purchase circuit are periodically and simultaneously closed and opened while the remaining second pair of diagonally disposed switching elements remains open Reversal of the current through the load is achieved by simultaneously opening and closing the second pair of switching elements while the first pair remains opened. It is stated that this method of operating a four-arm bridge control circuit has the advantage of a linear relationship between the control voltage and the load voltage.

It is also stated that it is a disadvantage of the foregoing system that the polarity of the load voltage and the current which flows into the bridge changes during the opening intervals of the periodically operating electronic switching elements. This results because the inductive load component causes load current to remain flowing during intervals that the operated switching elements are open. Such current flows through free-wheeling diodes which are disposed and shunt across each such electronic switch. Such diode current flows back into the power supply, in a direction opposite to the current flow during the time that such switching elements are closed, load voltages are reversed concurrently with such bypass diode current. The effect o this operation is that the load will experience a current having a relatively large alternating current ripple component which produces additional heat loss in the load. Moreover, in situations where a motor is used as the load, the large alternating current component can create whining noises.

In U.S. Pat. No. 3,260,912, the teachings of which are fully incorporated herein, there is disclosed another DC chopper converter designed for reducing the amplitude of the alternating current component in the load current by the use of pulse-width control. In this system, a first pair of diagonally disposed electronic switching elements in the bridge are opened and closed during time intervals which are offset with respect to one another. Thus, during the open interval of each such electronic switching elements, the diagonally associated electronic switching elements remain closed. This offset driving arrangement provides an advantage over the above-described simultaneous driving arrangement because the load current which continues to flow after a particular switching element is opened, as a result of the inductive component of the load, does not flow back into the power supply, but circulates through the closed electronic switching element and a free-wheeling diode. This arrangement, therefore, produces after each cycle of switching elements closure which delivers to the load electrical energy from the power supply, a bypass phase which is distinguishable from the energy reversing backflow phase of the previously-discussed arrangement, which does not reverse the load voltage, but reduces it to zero. Thus, the alternating current ripple components in the motor load current have the same pulse frequency as in the first-mentioned arrangement, but only one-half of the magnitude. This arrangement, however, has the disadvantage of a non-linear relationship between the control voltage and the load voltage, particularly in the range of small control voltages. Thus, it is possible in situations where small control voltages are utilized, that the load voltage would approach zero before the control voltage approaches zero. This results from the fact that the conductive intervals of diagonally disposed electronic switching elements cannot overlap because a safety interval must be maintained to prevent electronic switching elements which are disposed on the same half of the bridge from being simultaneously conductive and causing shortcircuit conditions. Similarly, an overlap of the conductive intervals of diagonally opposite switching elements can occur when negative control voltages are utilized only when such control voltages exceed predetermined negative values.

It becomes apparent, therefore, that a bidirectional direct current control bridge, which is operated in accordance with the latter method described above, has a region of insensitivity in which the load voltage is zero for small control voltages. This operational characteristic is a significant disadvantage for most applications. Electronic switching elements of the second diagonal, which are conductive during the non-conductive interval of the first pair of electronic switching elements, do not carry any current. The current is conducted through the free-wheeling diodes which are poled for conduction in a direction which is opposite to the forward conduction of the respectively associated switching elements. This can cause, in some applications, damage to the switching elements.

In all of the arrangements discussed above, the zero crossings of the source current are followed to determine when to switch over from one driving bridge diagonal to the other. Zero crossings in each direction are detected by a pair of detectors that produce an output signal only when an input signal thereto exceeds a predetermined bias signal level. It can be appreciated that a gap can exist between the two bias voltages. As a result, there is a region of insensitivity to zero crossings about the zero crossing axis and this can lead to non-linearities in the load current during switch-over between bridge diagonals. Therefore, the low current can only be precisely set within a limited dynamic range. Certain applications, however, such as in the drive gradient coils for nuclear magnetic resonance tomography, require a greater dynamic range, for example from 2 mA through 200 A, in which the load currents must be precisely controlled.

The present invention provides an improved DC chopper converter circuit and/or method for operating a DC chopper circuit. An aspect of the invention is a DC chopper converter circuit of the type described above wherein load current is precisely controlled within the zero crossing region, given reversals of the load current directions.

To this end, in an embodiment, the invention provides that switching elements coupled in series in one-half of the bridge circuit are also clocked or driven in offset phase. Since all switching elements, preferably transistors, are clocked in every direction of the load current, only displacement of the pulse duty factor is needed to accommodate transition from one current direction to the other, so that a continuous transition is provided, i.e., one without a kink or ripple in the current curve.

An aspect of the invention is that proceeding from no load current, and given a pulse duty factor of 50% for all switching elements in the bridge circuit as well as the out of phase driving of the switching elements lying on opposite diagonals, the load current can be increased by increasing the pulse duty factor of a pair of switching elements lying on a bridge diagonal. An especially great dynamic range can thus be achieved.

Another aspect of the invention is an embodiment wherein a safety margin time period is taken into consideration and accommodated between on-times of two switching elements coupled in series on one-half of a bridge circuit. This arrangement prevents overlapping closure of two switching elements coupled in series between two power supply terminals so that a short circuit does not occur across the supply terminals.

In one advantageous embodiment of the invention, only two switching elements lying on opposite bridge diagonals are clocked, whereas the respective switching element lying in series with the load remains opened when the load current exceeds a predetermined value. Since this measure is only undertaken for high load currents, it does not interfere with the continuous transition between load current direction changes. Moreover, the safety margin time periods can be relatively short since they need not be provided for higher load currents, which require longer switching times for the switching elements.

In yet another advantageous embodiment of the invention, each pair of serially coupled bridge arms are connected together via an inductor. The free-wheeling diodes are coupled in parallel to the inductor and a switching element terminal rather than only a switching element. The load is connected to taps of the two inductors. In case of a short-circuit arising when two switching elements in two serially connected bridge arms are closed at the same time, the current rise is limited by these inductors. Yet, these inductors hardly limit the rise time of the load current.

These and other features and aspects of the invention are set forth in greater detail below in the detailed description of the presently preferred embodiments and the accompanying drawings.

FIG. 1 is a schematic illustration of a DC chopper converter circuit.

FIGS. 2-5 waveform diagrams of the drive currents for the switching elements 1-4 of FIG. 1 for a load current of zero.

FIGS. 6 and 7 are waveform diagrams of the drive currents for switch elements 1 and 4 of FIG. 1 for a low level load current.

FIGS. 8 and 9 are waveform diagrams of the drive currents for switch elements 1 and 4 of FIG. 1 for mid-level load current.

FIG. 10 is a waveform diagram of the output of two bridge switching transistors coupled in series.

FIGS. 11-14 are waveform diagrams of the drive currents for the switch elements of FIG. 1 with safety time margins taken into consideration.

FIGS. 15-18 are waveform diagrams of the drive currents for the switch elements of FIG. 1 for a high level load current.

FIG. 19 is a schematic view of another DC chopper converter circuit.

In FIG. 1 there is illustrated a DC chopper converter circuit in typical four-arm or H-bridge configuration. As illustrated, two pairs of serially connected switching elements are coupled between two supply voltage terminals 11 and 12 of a voltage U.sub.v. To this end, serially connected switching elements 1 and 2 are coupled between the two supply voltage terminals 11 and 12 and form one leg of the H configuration. Two other serially connected switching elements 3 and 4 are also coupled between the supply voltage terminals 11 and 12 to form the other leg of the H configuration. A load 9 provides the remainder of the H configuration and includes one terminal portion coupled between switching elements 1 and 2 and another terminal portion coupled between switching elements 3 and 4 to thereby form the bridge of the H-configuration.

For the purposes of this specification and accompanying claims, switchings 1 and 4 are considered to be disposed along one diagonal of the H bridge configuration while switching elements 2 and 3 are considered to be disposed along another diagonal of the H bridge configuration. Thus, switching elements 1 and 4 constitute one diagonal pair of switching elements, while switching elements 2 and 3 constitute another diagonal pair of switching elements at the same time, switching elements 1 and 2 are considered to be vertically disposed and to comprise one pair of vertically disposed switching elements. Similarly, switching elements 3 and 4 comprise another pair of vertically disposed switching elements.

Each switching element has a free-wheeling diode connected parallel to it, whereby all the free-wheeling diodes are poled for conduction opposite the polarity of the supply voltage. A free-wheeling diode 5 is coupled parallel to switch 1, while free-wheeling diode 6 is coupled parallel to switch 2, free-wheeling diode 7 is coupled parallel to switch 3, and free-wheeling diode 8 is coupled parallel switch 4.

As further illustrated, a load current I.sub.A flows through the load 9 and is captured by a transducer element 10. The captured current is supplied to a comparator 13 from the transducer 10 as actual value I.sub.A actual. The actual value I.sub.Aist is compared to a rated value I.sub.Asoll in the substracting amplifier 13. An output of the subtracting amplifier 13 is supplied to a regulator 14 which, in turn, supplies its output to a control circuit 15. The control circuit 15 includes four drive signal outputs D1, D2, D3 and D4 that are supplied to the switching elements 1, 2, 3, and 4, respectively. Thus, the switching elements 1, 2, 3, and 4 are driven in view of the difference signal appearing at the output of subtracting amplifier 13.

The regulator 14 preferably comprises a proportional-plus-integral (PI) unit. The output of the regulator 14 thus preferably represents the signal difference times a proportionality constant plus the integral of the difference.

The switching elements 1-4 preferably comprise transistors, for example FET (field effect) power transistors. The load 9 preferably is an inductive load (e.g., a motor) so that a continuous load current usually arises given adequate switching frequency of the switching elements 1-4. As long as the load 9 is separated from the supply voltage terminals 11 and 12, the current driven by the conductivity of the load 9 flows across corresponding free-wheeling diodes 5-8 as well as across switching elements 1-4 that may still be closed.

The load current I.sub.A can be arbitrarily controlled and can assume positive and negative values. When, for example, the DC chopper converter is utilized in a nuclear magnetic resonance apparatus for the power supply of gradient coils, a slight deviation from a prescribed rated curve of the load current is especially important. In particular, extremely small currents must be precisely set. For example, the required dynamic range lies between 2mA and 200 A. As the following consideration illustrates, this dynamic range is not possible merely by simultaneous switching of two switching elements lying diagonally opposite one another, for example the simultaneous switching of switching elements 1 and 4, or by permanently turning on one of these two switching elements and clocking the other switch.

A required pulse width t.sub.p is proportional to a desired, stationary current I as illustrated in the following equation: t.sub.p =k.times.I; whereby k is a proportionality constant.

The proportionality constant k is equal to the ohmic resistance of the load 9 divided by the switching frequency multiplied by the supply voltage. Accordingly, the following equation is valid:

where R is the ohmic resistance of the load 9, f is the switching frequency, and UV is the supply voltage.

Given typical values of a resistance R=100 mOhms, a switching frequency f of 25 kHZ and a load current I.sub.A =200 amps, a switching time t.sub.p of 2.5 .mu.s results. When the load current I.sub.A is then to be set to 2 mA, a desired pulse width of the respective switching element 1-4 of t.sub.p =26.6 ps results. However, such a short on-time cannot be realized using only a transistor as a switch element.

Since this short on-time cannot be realized with a transistor, in accordance with the invention, the on-time preferably is generated by the overlap of the on-times of two switching elements 1 and 4 or 2 and 3, respectively, arranged along a bridge diagonal. This is set forth in greater detail below with reference to FIGS. 2-9.

In FIGS. 2-5, there are illustrated waveform diagrams for the drive currents for the switch elements 1-4. Accordingly, these waveform diagrams illustrate the switching status of the switching elements 1-4, the case wherein the load 9 is not to be supplied with a supply voltage.

As illustrated, the switching elements 1 and 4 (FIGS. 2 and 5), just like the switching elements 2 and 3 (FIGS. 3 and 4) are operated in a push-pull action so that no overlap between the on-times thereof occurs, thus, the load 9 is not coupled to the supply voltage U.sub.v. The switching elements 2 (FIG. 3) and 4 (FIG. 5) are each respectively switching elements in a push-pull action relative to the switching elements 1 (FIG. 2) and 3 (FIG. 4), respectively, in order to avoid a short circuit between the supply terminals 11 and 12. The pulse-duty factor, i.e, the ratio of on-time TE to off-time TA, is equal to one for all of the switching elements 1-4, for this situation.

When one wishes to increase the load current I.sub.A, then the pulse-duty factor of two switching elements lying on the same bridge diagonal is increased, i.e., the on-time TE is increased relative to the off-time TA. This is accomplished by shortening the off-time TA. The selection of the switch pair driven in this manner depends on the desired direction of the change in the current I.sub.A. In the two remaining switching elements, that do not carry current, the on-time simultaneously is shortened to reduce the pulse-duty factor.

In FIGS. 6-9, there are illustrated waveform diagrams for the drive circuits of the circuits 1-4 wherein the supply voltage U.sub.V is to be applied to the load 9. Accordingly, the waveform diagrams illustrated the switching status of the switching elements 1-4.

With reference to FIG. 6 (switching status of switch 1) and FIG. 7 (switching status of switch 4), it should be appreciated that an overlap of the on-times occurs in the regions depicted with hatch marks. During this period, the load 9 is supplied with the supply voltage U.sub.V so that a corresponding change in the current I.sub.A occurs.

In accordance with the invention, the pulse-duty factors, i.e. the on-times of the switching elements 1 and 4 can be further increased as illustrated in FIG. 8 (switching status of the switch 1) and FIG. 9 (switching status of the switch 4), so that the hatched overlapped times become even greater. Accordingly, a greater change in the current I.sub.A in the load occurs due to the longer application of the supply voltage U.sub.V thereto.

In the drive situations illustrated in FIGS. 6-9, the switching elements 2 and 3 that are connected in series with the switching elements 1 and 4, respectively, are driven out of phase, i.e., so that no short across the terminals 11 and 12 of the supply voltage U.sub.V occurs. For the sake of clarity and ease of understanding, the switching diagrams for the switching elements 1 and 4 have been omitted in FIGS. 6-9.

In view of the foregoing, it can be seen that overlap times, and thus, the resulting stationary current can be made arbitrarily small, at least theoretically. However, the switching elements 1-4 in fact do not enjoy sharp turn-ons and turn-offs. Accordingly, the switching elements do not have arbitrarily steep switching element edges.

As is illustrated in FIG. 10, however, the overlap need not be formed by rectangular pulses. A curve A in FIG. 10 illustrates the typical turn-off behavior of a transistor while a curve B illustrates the typical turn-on behavior of a transistor. The overlap time that is critical for control purposes is illustrated by a hatched region C.

The illustrated drive control has the special advantage that a steady transition from the positive load current I.sub.A to a negative load-I.sub.A is possible.

The turn-on and turn-off time of transistors cannot be neglected, however, and must be taken into consideration in the drive. It must be assured that the switching elements 1 and 2 or 3 and 4, respectively, lying in series in serially coupled bridge arms are never turned on simultaneously. Since the power supply source U.sub.V should have an optimally low inductance, the simultaneous turn-on would lead to a short circuit having an extremely rapid rise and, thus, would lead to a destruction of the switching elements 1-4. In order to prevent this from occurring, the duration of a turn-on command or instruction for a transistor (t.sub.on) in a bridge arm must be delayed by a safety margin time t.sub.tot in comparison to duration of transistor turn-off (t.sub.off) for the transistor lying in the same bridge half. Thus, the following equation is employed:

FIGS. 11-14 illustrate switching waveforms for the switching elements 1-4, whereby the safety margin times t.sub.tot are schematically illustrated. The solid lines illustrate the actual driving current waveforms taking a safety margin time t.sub.tot into consideration. The broken lines illustrate the theoretical turn-on times for the switching elements. In comparison to the actual waveform, the safety margin time t.sub.tot is illustrated in greatly exaggerated form for the sake of clarity and understanding.

It can be appreciated that the turn-off time of a transistor is highly dependent on the current that the transistor carries just before the turn-off time. Without undertaking further measures, the safety margin time t.sub.tot would have to be designed for a maximum current, for example, 200 amps. Accordingly, the safety margin time t.sub.tot would have to be made relatively long.

A shortening of the safety time t.sub.tot becomes possible when the two switching elements in the diagonal that does not carry current are entirely shut-off as soon as the load current I.sub.A exceeds a defined value (for example, 10 amps). A corresponding switching diagram for the switching elements 1-4 is shown in FIGS. 15-18 for the direction of current. In this case, only the switching elements 1 and 4 are clocked, so that a positive load current I.sub.A flows, whose amount is defined and controlled by the overlap time of the on-times of the switching elements 1 and 4. The switching elements 2 and 3 (FIGS. 16 and 17) remain shut-off since the load current I.sub.A has exceeded a defined value.

Such an arrangement preserves the advantage of the continuous zero-axis crossing of the load current I.sub.A since all switching elements 1-4 are clocked given low currents and, thus, a steady transition can be made from one direction of the current to the other. On the one hand, however, the safety time t.sub.tot only has to be designed to the turn-off times at a relatively low current and can thus be kept short.

A further solution of the short-circuit problem that can arise due to overlap of the on-times of two switching elements 1 and 2 or, respectively, 3 and 4 lying in serially coupled bridge arms is illustrated in FIG. 19. As illustrated, the circuit of FIG. 1 is modified such that inductors 16 and 17 are coupled between serially connected bridge arms. Accordingly, inductor 16 is coupled between switching elements 1 and 2 and inductor 17 is coupled between switching elements 3 and 4. Thus, if switching elements 1 and 2 are closed at the same time, the short circuit current must also pass through the inductor 16. Similarly, the current would pass through inductor 17 if switching elements 3 and 4 were simultaneously closed.

As further illustrated, the free-wheeling diodes 5-8 are repositioned such that they are no longer coupled parallel only with their respective switching elements 1-4. Instead, each diode 5-8 is coupled parallel to its respective switching element 1-4 and the associated inductor 16 or 17. Thus, each free-wheeling diode is still coupled in series with the switching element coupled in series with the switching element previously associated with the diode. However, the two diodes of a bridge half previously coupled in series now include an inductor coupled therebetween.

Each inductor 16 and 17 is provided with a tap to which the load 9 is coupled. Thus, the load 9 is coupled between portions of inductors 16 and 17.

In this arrangement, should a short occur between terminals 1 and 12 by virtue of the simultaneous closure of switching elements 1 and 2 or 3 and 4, current rise is limited by the inductor 16 or, respectively, 17. At the same time, however, because the load 9 is connected only to taps of the inductors 16 and 17, the inductors do not adversely affect the load current rise.

Moreover, a problem that specifically arises given the use of MOSFET transistors as the switching elements is also solved. MOSFET transistors have an inherent free-wheeling diode that is depicted by the broken lines illustrated in FIG. 19. This free-wheeling diode exhibits the unpleasant property that it normally switches elements more slowly than the allocated switching path of the MOSFET. However, due to the presence of the inductors 16 and 17, the rise of a short-circuit current is retarded and the significantly faster diodes 5-8 then remain effective as the free-wheeling diodes in this circuit.

While preferred embodiments have been shown, modifications and changes may become apparent to those skilled in the art which shall fall within the spirit and scope of the invention. It is intended that such modifications and changes be covered by the attached claims.