EP2263425A1

EP2263425A1 - Circuit arrangement for operating hid discharge lamps

Info

Publication number: EP2263425A1
Application number: EP09724263A
Authority: EP
Inventors: Michael Zimmermann; Eduardo Pereira; Martin Huber; Frank Horn
Original assignee: Tridonic AG
Current assignee: Tridonic AG
Priority date: 2008-03-25
Filing date: 2009-03-19
Publication date: 2010-12-22
Also published as: WO2009118128A1; CN101982020B; DE102008017545A1; CN101982020A

Abstract

The invention relates to a circuit arrangement for operating a gas discharge lamp, with a full-bridge circuit to which a DC voltage (UO) is applied and which comprises four controllable switches (S1-S4), wherein a gas discharge lamp (EL) is intended to be arranged in a bridge branch which connects a node between the first switch (S1) and the second switch (S2) to a node between the third switch (S3) and the fourth switch (S4), and with a control circuit (1), which alternately activates one of the two bridge diagonals, each of which comprises a high-frequency-clocked switch and a low-frequency-clocked switch, wherein the control circuit (1) closes a high-frequency-clocked switch of a bridge diagonal whenever a measurement signal fulfils a reconnection condition, wherein the control circuit (1) also opens the low-frequency-clocked switch of the same bridge diagonal after a predetermined period of time (T2) after the opening of the high-frequency-clocked switch (Figure 2b) if until that point the reconnection condition for the high-frequency-clocked switch has not yet been fulfilled, in order to allow the bridge branch current to decrease more rapidly, wherein the closing of the low-frequency-clocked and/or the high-frequency-clocked switch takes place temporally after the onset of the reconnection condition of the high-frequency-clocked switch.

Description

Shutter arrangement for operating HID charge lamps

The present invention relates to a circuit arrangement for Bet re iben of Gasent charge lamps, in particular s other of high-pressure gas discharge lamps, which comes in electronic ballasts for corresponding gas discharge lamps used. High-pressure gas discharge lamps differ from low-pressure gas discharge lamps, among other things because they require higher ignition voltages and I change their color temperature with the respective supplied lamp power. The latter feature has the consequence that high-pressure gas discharge lamps are difficult or not dimmable. Rather, in order to maintain the color temperature of the high-pressure gas discharge lamp, the energy supplied to the j ewe iligen lamp must be kept constant by a corresponding regulation. An electronic ballast for high-pressure gas discharge lamps must therefore on the one hand generate a high ignition voltage and on the other hand offer the possibility to keep the power supplied to the lamp constant.

Known electronic ballasts for high pressure gas discharge lamps are based on a full bridge circuit comprising four controllable electronic switches. This principle sol l below with reference to FIG. 4 will be explained 25 wherein the circuit shown in Fig. 4, for example, from WO-A-86/04752 is known.

As has already been mentioned, this known circuit for driving a gas discharge lamp EL, in particular a high-pressure gas discharge lamp, a full bridge with four controllable switches Sl - S4, which are formed according to the aforementioned document in particular by bipolar transistors. In the bridge branch of this full bridge, a series resonant circuit consisting of a coil Ll and a capacitor Cl is connected, wherein the gas discharge lamp EL to be controlled is arranged parallel to the capacitor Cl. The full bridge is supplied with a DC voltage U ₀ . Free-wheeling diodes are connected in parallel with the switches or transistors S 1 -S 4, but they are not shown in FIG. 4 for the sake of simplicity. In order to operate the gas discharge lamp EL, it is proposed in WO-A-86/04752 to close the switch S4 during a first operating phase and to open the switches S2 and S3. Furthermore, during this first operating phase, the switch S1 is alternately switched on and off with a high clock frequency. During the switch-on duration of the switch S1, a direct current flows via the switch S1, the coil or inductor L1, the gas discharge lamp EL and the switch S4, which is always closed during this operating phase. By opening the transistor Sl of the current flow is interrupted and implemented in the coil Ll by the current flow previously constructed magnetic energy into electrical energy which provides a counter-voltage, which up to the next switch-on of the switch Sl the current flow through the gas discharge lamp EL in the same direction upright gets wherein the energy stored in the coil Ll is dissipated. By turning on the switch Sl again the circuit described above is closed again, so that the above-mentioned process is repeated. During this first operating phase, in which the switches S2 and S3 are permanently open and the switch S4 is permanently closed and the switch S1 is opened and closed at high frequency alternately, the gas discharge lamp EL is always flowed through in the same direction by the current. As a result, the gas discharge lamp EL flickers less during its operation and a higher light output is possible. During continuous operation with the DC voltage U ₀ , however, deposits can accumulate in the electrode region of the gas discharge lamp EL, which are caused by the constant flow of electrons in the same direction. In order to avoid these deposits, the gas discharge lamp EL is repeatedly reversed low-frequency. This happens because during a second phase of operation now the switches or transistors Sl and S4 are permanently opened and the switch S3 is permanently closed. Further, during this second phase of operation, the switch S2 is alternately turned on and off at high frequency, so that in principle the same operating mode as during the first operating phase described above, but during the second phase of operation, the current flow through the gas discharge lamp EL is reversed.

In summary, therefore, it can be stated that the full bridge shown in FIG. 4 is operated in principle with the DC voltage U ₀ , whereby, however, due to the low-frequency polarity reversal between the bridge diagonals Sl -. S4 or S2 - S3, that is supplied by the low-frequency switching between the two previously described first and second operating phases, the gas discharge lamp EL and the inductor Ll a low-frequency alternating current whose

Frequency of Umpolfrequenz corresponds. During the two phases of operation, either the switch Sl or the switch S2 high frequency alternately turned on and off. The size ratio between the clock frequency at which switches S1 and S2 are turned on and off alternately, and the much lower switching pole frequency, should be as large as possible, and may be, for example, 1000: 1. The larger this ratio, the smaller the throttle or coil Ll can be dimensioned. Due to the high-frequency switching of the switch Sl or S2, a correspondingly high-frequency current is generated, which flows through the throttle Ll. The throttle used to limit the lamp current can therefore be made smaller than in the case when it would be flowed through by a low-frequency current.

Ignition of the gas discharge lamp EL shown in Fig. 4 is carried out by means of the series resonant circuit formed by the inductor Ll and the capacitor Cl, wherein for ignition, an operation of the gas discharge lamp EL is required at a frequency which is close to the resonance frequency of the series resonant circuit. If this is the case, a voltage overshoot occurs at the gas discharge lamp EL, which leads to the ignition of the gas discharge lamp.

From EP-A2-0740492 is a circuit arrangement for igniting and operating a gas discharge lamp, in particular a high-pressure gas discharge lamp, known. For igniting or operating the gas discharge lamp is proposed in this document, with the help of a corresponding control circuit arranged in the bridge diagonal switches Sl, S4 and S2, S3 of the full bridge during a first phase of operation complementary to control with a relatively high frequency until the gas discharge lamp ignites. Subsequently, the control circuit switches over to a second operating phase (nominal operating phase) in which the control circuit activates the switches S 1 -S 4 of the full-bridge arrangement in a complementary manner with a relatively low frequency. In addition, a control device is used according to this document, which is coupled on the output side via a capacitance with the full bridge such that the full bridge is arranged parallel to the capacitance. The control device also serves to supply power to the full bridge and regulates in particular the power supplied to the gas discharge lamp. For this purpose, the voltage applied to the output terminals of the control device and the instantaneously flowing current are measured, the corresponding values are multiplied and the actual value formed is fed to the control device as the actual value of the lamp power. The aforementioned control circuit is connected to the controller and outputs the target value of the output power

Control device, wherein the control circuit, in particular during the first phase of operation described above

(Start-up operation phase) raises the setpoint so that the

Control device of the full bridge can supply a higher output power. The ignition of the gas discharge lamp can be effected by an ignition device which is coupled to the arranged in the bridge arm inductance Ll. Alternatively, you can the gas discharge lamp can be ignited by using the capacitance C 1 shown in FIG. 4 and connected in parallel with the gas discharge lamp EL, which together with the inductance L 1 forms a series resonant circuit.

Another circuit arrangement for igniting and operating gas discharge lamps, in particular metal halide high-pressure gas discharge lamps, which is known from GB-A-2319 677, is shown in FIG. This circuit arrangement also comprises four controllable switches S 1 -S 4 connected in a full bridge, which can be formed by bipolar transistors or field-effect transistors. In the bridge branch of this full bridge circuit is a gas discharge lamp EL and a series resonant circuit formed by an inductance Ll and a capacitance Cl. For starting, ie igniting, the gas discharge lamp EL, the full bridge is operated at a relatively high frequency with the aid of a corresponding control circuit, which can individually control the individual switches S 1 -S 4 via corresponding bridge drivers, which may be in the range 20-40 kHz. This high frequency is particularly chosen such that it is in the vicinity of the resonance frequency of the series resonant circuit consisting of the inductance Ll and the capacitance Cl, so that after a certain time the gas discharge lamp EL ignites. The ignition of the gas discharge lamp El can be detected, for example, by monitoring the lamp current or by monitoring the lamp brightness. Once the ignition of the gas discharge lamp EL has been detected, the full bridge is switched to a low operating frequency, which may be in particular in the range 50-200 Hz, switched to the Lamp to operate. As can be seen in Fig. 5, the circuit arrangement known from this document also comprises a designated as ignition or autotransformer transformer whose primary winding L2 is arranged in series with the capacitance Cl of the series resonant circuit, while the secondary winding is connected in series with the gas discharge lamp EL , This transformer with the inductors L2 and L3 serves to generate an increased voltage in the secondary coil L3, which is applied to the gas discharge lamp EL when a current flow through the capacitor Cl occurs (which is the case in particular when the high ignition frequency is applied). In this way, the ignition and the operation of the gas discharge lamp EL can be facilitated.

However, the circuit shown in Fig. 5 which uses an autotransformer whose primary winding L2 is connected in series with the series resonant circuit capacitance Cl and the secondary winding L3 thereof in series with the gas discharge lamp EL has the disadvantage that a rib current flowing through the full bridge is also present is transformed and accordingly affects the lamp current negative. Although the circuit arrangement known from EP-A2-0740 492, which has likewise been discussed previously, makes it possible to regulate or maintain the power supplied to the full-bridge, it requires a relatively large number of components for this, so that the circuit arrangement is relatively complex and expensive is.

DE 199 16 879 A1 discloses a circuit arrangement for operating high-pressure gas discharge lamps (HID lamps). known, as is known from Figures 1-3 of the present application. In this case, the switches S1, S2, S3 and S4 (see FIG. 1) are respectively clocked so that one of the bridge diagonals S1, S4 or S3, S2 is activated alternately. Each bridge diagonal consists of a high-frequency clocked switch and a low-frequency clocked switch.

The high-frequency clocked switch is in each case switched on again when the current flowing in the bridge branch current I _L 2 has reached a lower reversal point, ie a minimum value.

Furthermore, it is known from DE 199 16 878 A1 that the low-frequency clocked switch of an activated bridge diagonal is opened in addition to an open high-frequency clocked switch if the bridge branch current has not yet reached its minimum value after a predetermined period of time.

The invention now relies on this idea of additionally opening the low-frequency clocked switch and further develops this principle to the effect that the requirement for modern circuit technology is satisfied.

The above object is achieved according to the present invention by the features of the independent claims. The subclaims describe each of the preferred and preferred embodiments of the present invention.

According to a first aspect, the invention proposes one Circuit arrangement for operating a gas discharge lamp, with a full bridge circuit to which a DC voltage

(Uo) is applied and the four controllable switch (Sl - S4), wherein a gas discharge lamp (EL) in a bridge branch, which is a node between the first switch (Sl) and the second switch (S2) with a node between the third Switch (S3) and the fourth switch (S4) connects, and is arranged with a control circuit (1) which alternately activates one of the two bridge diagonal, each consisting of a high-frequency clocked switch and a low-frequency clocked switch, wherein the control circuit high-frequency clocked switches a bridge diagonal always closes when a measurement signal meets a reclosing condition, the control circuit and the low-frequency clocked switch the same

Bridge diagonal opens after a predetermined period (T2) after opening the high-frequency switch

(Figure 2b), if until then the reconnection condition for the high-frequency clocked switch is not yet met to let the bridge branch current fall faster, the closing of the low-frequency and / or hochfreqent clocked switch delayed in time after the occurrence of the reclosing condition of the high-frequency clocked Switch takes place.

According to a further aspect of the invention, a circuit arrangement is proposed for operating a gas discharge lamp, in particular a high-pressure gas discharge lamp, having a full-bridge circuit to which a direct current voltage (Uo) is applied and which comprises four controllable switches (Sl-S4), a gas discharge lamp ( EL) in a bridge branch connecting a node between the first switch (Sl) and the second switch (S2) to a node between the third switch (S3) and the fourth switch (S4), and with a control circuit that alternately one of the two bridge diagonals, each consisting of a high-frequency clocked switch and a low-frequency clocked switch activated, wherein the control circuit (1) always closes a high frequency clocked switch a bridge diagonal when a measurement signal meets a reclosing condition, the control circuit also clocked the low-frequency Switches the same bridge diagonal after a predetermined period (T2) after the opening of the high-frequency clocked switch opens, if until then the reconnection condition for the high-frequency clocked switch is not met to let the bridge branch current fall faster.

The timing of the additional opening and / or the time of the subsequent closure of the low frequency clocked switch may be externally adjustable and / or adaptively, i. be event-dependent adjustable.

The time of closure of the low-frequency and / or high-frequency clocked switch may in particular be in a period which begins with the occurrence of the reclosing condition of the high-frequency clocked switch.

The period may preferably be no later than 3uS, preferably 30OnS to 2.5uS, after the entry of the Restart condition of the high frequency clocked switch ends.

The timing of closing the low frequency clocked switch may be after the high frequency clocked switch is closed when the lamp voltage or output voltage is below a predetermined threshold.

The invention also relates to a method for operating a gas discharge lamp, comprising a full bridge circuit, to which a DC voltage (Uo) is applied and which comprises four controllable switches (S1-S4), wherein a gas discharge lamp (EL) in a back branch, the one Junction between the first switch (Sl) and the second switch (S2) with a node between the third switch (S3) and the fourth switch (S4) connects, and wherein alternately one of the two bridge diagonals is activated, each of a high-frequency clocked

Switch and a low-frequency clocked switch, wherein a high-frequency clocked switch a bridge diagonal is always closed when a measurement signal meets a reclosing condition, the low-frequency switch clocked the same

Brückendiagonale after a predetermined period (T2) after opening the high-frequency clocked switch is opened (Figure 2b), if until then the reconnection condition for the high-frequency clocked switch has not yet met, to lower the bridge branch current faster, the closing of the low-frequency and / or the high frequency clocked switch temporally (conscious delayed) occurs after the reconnection condition of the high-frequency clocked switch occurs.

The invention further relates to a method for operating a gas discharge lamp, comprising a full bridge circuit, to which a DC voltage (Uo) is applied and which comprises four controllable switches (S1 - S4), wherein a gas discharge lamp (EL) in a bridge branch, which has a Node between the first switch (Sl) and the second switch (S2) with a node between the third switch (S3) and the fourth switch (S4) connects, and wherein alternately one of the two bridge diagonals, each consisting of a high-frequency clocked switch and a low-frequency clocked switch, is activated, wherein the control circuit (1) closes a high-frequency clocked switch a bridge diagonal whenever a measurement signal meets a reclosing condition, whereby the low-frequency switch of the same bridge diagonal after a predetermined period (T2) after opening the high-frequency clocked 2b) when the reclosing condition for the high-frequency clocked switch has not yet been fulfilled in order to allow the bridge branch current to decrease more rapidly, the time of the additional opening and / or the time of the subsequent closure of the low-frequency and / or or high frequency clocked switch is externally adjustable and / or adaptive, ie is set as a function of the event.

Finally, the invention also relates to a control unit, in particular an integrated circuit such as an ASIC or microcontroller designed to carry out a method according to any one of the preceding claims.

The present invention will be described below with reference to preferred embodiments with reference to the accompanying drawings.

1 shows a circuit diagram of a circuit arrangement according to the invention according to a preferred embodiment of the present invention,

FIG. 2a shows a first diagram illustrating time-dependent voltage and current characteristics in the circuit arrangement shown in FIG. 1, FIG.

FIG. 2b shows a second diagram which shows the time-dependent current profile and switching states in the circuit shown in FIG

Circuit arrangement according to a further development,

FIG. 3 shows an electronic ballast in which the circuit arrangement shown in FIG. 1 is used, FIG.

Fig. 4 shows a circuit arrangement according to the known prior art, and

Fig. 5 shows a further circuit arrangement according to the known prior art.

6 is a schematic circuit diagram for explaining the present invention;

FIG. 7 shows waveforms which may occur in this circuit according to FIG. 6, FIG.

Fig. 8 shows waveforms using the present invention, and

9 shows a circuit diagram of a circuit arrangement according to the invention in accordance with a further preferred exemplary embodiment of the present invention.

The circuit arrangement shown in Fig. 1 comprises controllable switches Sl - S4, which are connected to form a full bridge. To the full bridge, a DC voltage U _{0 is} applied, which comes from a suitable DC voltage source of the corresponding electronic ballast, in which the circuit arrangement is used. To the switches Sl - S4 freewheeling diodes are connected in parallel, wherein for the sake of simplicity in Fig. 1, only the switch S1 parallel-connected freewheeling diode Dl is shown. As a switch Sl - S4 field effect transistors are preferably used, which already contain the freewheeling diodes. In the bridge branch of the full bridge shown in FIG. 1, a gas discharge lamp EL to be controlled, in particular a high-pressure gas discharge lamp, is arranged. The shown in Fig. 1 Circuitry is particularly suitable for the operation of metal halide high-pressure gas discharge lamps, which require very high ignition voltages. As already mentioned, high-pressure gas discharge lamps differ from low-pressure gas discharge lamps in particular in that they require higher ignition voltages and a higher pressure occurs in their smaller lamp body. Furthermore, high-pressure

Gas discharge lamps have a higher luminance, but the color temperature of the respective high-pressure gas discharge lamp changes with the supplied power.

Electronic pre-scanners for high-pressure

Gas discharge lamps should therefore on the one hand provide high ignition voltages and on the other hand allow a constant maintenance of the supplied power.

With the bridge branch of the full bridge shown in Fig. 1, a series resonant circuit is coupled to a

Inductance Ll and a capacitance Cl, wherein the capacitance Cl acts on a tapping point of the inductance Ll and is connected via a further controllable switch S5 in parallel with the switch S4. In addition, a smoothing or filtering circuit is provided which has a further inductance L2 and a further capacitance C2, these components being connected as shown in FIG. To the full bridge, a resistor Rl is also connected, which serves as a current measuring or shunt resistor.

The aforementioned series resonant circuit with the inductance Ll and the capacitance Cl serves in combination with the further capacitance C2, in particular for igniting the gas discharge lamp EL. For this purpose, the series resonant circuit is excited to resonance, ie, a frequency corresponding to the resonant frequency of the lamp or a Vilefaches thereof (harmonic) supplied. The excitation of the resonant circuit is effected by alternately switching the switches S3 and S4. This will be explained in more detail below.

To ignite the gas discharge lamp EL, two switches connected in series, for example the switches S1 and S2, are opened by means of a suitable control circuit and the switch S5, which is in series with the capacitance C1, is closed. The other two switches, for example, the switches S3 and S4, the full bridge are alternately opened and closed, this being done at a relatively high frequency (about 150 kHz). The switching frequency is slowly lowered in the direction of the resonance frequency of the series resonant circuit formed by the inductance Ll and the capacitance Cl. The ignition voltage of the gas discharge lamp EL is usually reached before reaching the resonance frequency. In this case, the switching frequency for the switches S3 and S4 is maintained at this frequency until the lamp EL is lit. The voltage dropping at the right half of L1 is, due to the autotransformer L1 implemented autotransformer principle, for example, in the ratio 1: 2 up to the left half, which is coupled to the gas discharge lamp EL, transformed, wherein the voltage occurring at the left half of the inductance Ll forms the actual ignition voltage for the gas discharge lamp EL, which is applied to the lamp via the capacitor C2 becomes. In order to detect the ignition of the gas discharge lamp EL, the voltage dropping at the tapping point of the inductance L 1 is measured, which is proportional to the Zündbzw. Lamp voltage Uj: _L is there after the. Igniting the lamp EL this dampening acts on the series resonant circuit. After ignition of the gas discharge lamp EL, the switch S5 is opened for the subsequent normal operation.

In addition, it should be noted that the switch S5 is not absolutely necessary for the functionality of the circuit arrangement according to the invention. Rather, the switch S5 could remain closed even after the ignition of the gas discharge lamp EL or basically be replaced by a corresponding bridging. With the help of the switch S5, which is opened after the ignition of the gas discharge lamp EL, however, a cleaner operation of the gas discharge lamp EL is possible. Furthermore, it should be noted that the ignition coil Ll is designed in particular such that it operates in the following explained in more detail normal operation in the saturation and thus does not affect the rest of the circuit. This can be achieved, for example, in that a coil with an iron core is used as the ignition coil L1, which is operated in saturation in normal operation, so that the coil Ll after ignition of the gas discharge lamp EL in normal operation forms only a negligible inductance. In normal operation, therefore, only the inductance L2, which is also provided in the bridge branch, is effective in limiting the current.

Subsequently, the initiated after the ignition of the gas discharge lamp EL normal operation will be explained in more detail, wherein during normal operation, the circuit arrangement according to the invention or full bridge is operated in a so-called. Discontinuous mode. In principle, the full bridge shown in Fig. 1 with the controllable switches Sl S4 operated in a conventional manner during normal operation, ie the two bridge diagonals with the switches Sl and S4 or S2 and S3 are alternately activated and deactivated and thus the corresponding switches of the two bridge diagonals alternately or complementary to each other switched on and off, wherein also activates the bridge diagonal with the switches Sl and S4, the switch high-frequency alternately switched on and off, while according to activation of the bridge diagonal with the switches S2 and S3 the controllable switch S2 high frequency alternately turned on and off. That is, the full bridge is reversed with a relatively low frequency, which may be in particular in the range 80-400 Hz, while the switch Sl or S2 of each activated bridge diagonal also high frequency, for example, with a frequency of about 90 kHz, alternately on and is turned off. This high-frequency switching on and off of the switch Sl or S2 is carried out with the aid of a high-frequency pulse width modulated control signal of a corresponding control circuit, which is screened by means of existing from the components L2 and C2 filter or smoothing circuit, so that at the gas discharge lamp EL only the linear average of the current flowing over the bridge branch branch current i _L2 is applied. With the help of the pulse width modulated control signal, the power supplied to the full bridge can be kept constant, which - as has been mentioned - especially for the Operation of high-pressure gas discharge lamps is important. The low-frequency component of the gas discharge lamp El supplied current is generated by switching or reversing the polarity of the two bridge diagonals, ie by switching from Sl and S4 to S2 and S3. In this case, the lamp EL is applied to the supply voltage U ₀ or to ground via the right bridge branch with the switches S3 and S4, so that essentially only the low-frequency component is applied to the terminals of the lamp EL.

According to the aforementioned low-frequency discontinuous mode, the controllable switch S1 or S2 of the respectively activated bridge diagonal is always closed when the branch current i _L2 flowing across the inductance L2 has reached its minimum. By "minimum" is meant the lower reversal point of the current \ _L2 , whereby this minimum can well be in the slightly negative current value range.

To consider the current flow is to be assumed below that initially the bridge diagonal is activated with the switches S2 and S3, while the bridge diagonal is disabled with the switches Sl and S4. That is, the switches S2 and S3 are closed while the switches Sl and S4 are opened. At the time of closing of the switches S2 and S3 begins to flow through the inductance L2, a current i _L 2, which increases in accordance with an exponential function, in the region of interest here, a quasi-linear increase of the current ^A 2 can be seen, so that subsequently the For simplicity, a linear increase or decrease of the current ii2 is spoken. By opening the switch 52, this current i _{L2 is} interrupted, wherein - as already mentioned - the switch S2 in particular high-frequency and independent of the switching state of the switch

53 is alternately opened and closed. The opening of the switch S2 has the consequence that the current i _L2 while initially on the freewheeling diode Dl of the open switch Sl in the same direction continues to flow, but continuously decreases and may even eventually reach a negative value.

This is especially the case until the electrons have been eliminated from the barrier layer of the freewheeling diode Dl. The achievement of this lower reversal point of the current ii ₂ is monitored and the switch S2 closed after detecting this lower reversal point, so that the current rises again. That is, that high-frequency switching of the switch S2 is always carried out when the lower reversal point of the current i _L 2 has been reached. The opening of the switch S2 can be chosen arbitrarily in principle, wherein the time of opening of the switch is particularly crucial for the power supply of the gas discharge lamp EL, so that by appropriate adjustment of the opening time, the power supplied to the lamp can be controlled or kept constant. As a switching criterion for this example, the time or the maximum value of the branch current i _L2 can be used. By the measure that the respective high frequency alternately switched on and off switches Sl and S2 respectively in the lower reversal point of the current i _L 2, ie in the vicinity of the current value zero, is turned on again, the respective field effect transistor Sl or S2 is conserved, ie protected from destruction, and it can Field effect transistors are used as switches Sl and S2, which have relatively long Ausräumzeiten for the corresponding freewheeling diode. This will be explained in more detail below.

Before the switch S2 is closed, a voltage is applied to it, which in the present case is approximately 400 volts. When switch S2 is closed, this voltage collapses, i. It drops very rapidly from 400 volts to 0 volts. The special feature of a field effect transistor, however, is that the current already begins to flow upon activation of the corresponding field effect transistor, before the corresponding voltage has fallen to 0 volts. In this short period of time between the increase of the current flowing through the field effect transistor and the reaching of the voltage 0 volts, a product supplied to the respective field effect transistor is formed by the product of the current and the voltage, which can destroy the field effect transistor. Therefore, it is advantageous to switch the field effect transistor at a lowest possible current flow, in particular in the vicinity of the current value zero.

Furthermore, it should be noted that the current flowing through the inductor L2 current i _L 2 flows through the freewheeling diode of Dl when the switch Sl is open and the switch S2 is still open. If the switch S2 is closed and the switch S1 opened, it takes a certain period of time until the electrons could be eliminated from the barrier layer of the freewheeling diode Dl. During this time, the field effect transistor Sl is practically in a conductive state. This means that the field effect transistor S2 during a relatively short period of time until clearing of the barrier layer of the freewheeling diode Dl, the

Field effect transistor Sl is assigned to the full operating voltage U ₀ , which is approximately 400 volts, is applied, which may also lead to the above-described overloading and possibly even destruction of the field effect transistor S2. Due to the previously proposed approach, namely the turning on the switch S2, whenever the current i flowing through the inductor L2 is _L2 reaches its minimum, the effect described above with reference to the Ausräumzeit of the switch or field effect transistor Sl is almost irrelevant, so that for the switches Sl - S4 also field effect transistors can be used which have relatively long Ausräumzeiten for the associated freewheeling diodes. Although there are already switching elements with very short Ausräumzeiten, such. As the so-called. IGBT (Insulated Gate Bipolar Transistor), but these devices are very expensive. With the help of the present invention can thus be dispensed with the use of such expensive components.

For the procedure described above, it is necessary to know the instantaneous value of the current i _L2 and the time at which it reaches its reversal point.

The instantaneous value of the current For example, be determined by measuring the voltage drop across the resistor Rl voltage. The lower reversal point of the current i _L2 is preferably determined by a voltage tapped off transformer-wise at the coil L2. For this purpose, a (not shown in Fig. 1) winding or coil are transformer-coupled to the coil L2, which leads to a differentiation of the current flowing through the coil L2 i _L2 stream and thus permits a statement about the point of reversal of the current i _L2.

The lower reversal point can also be detected indirectly by other feedback signals, for example the midpoint voltage at the connection point of the switches S1 and S2. It is only important that the time of the lower minimum of the bridge branch current is detectable. By contrast, quantative statements on this current are not required for determining the switch-on time of the high-frequency clocked switch.

The normal operation of the circuit arrangement shown in Fig. 1 will be explained below with reference to the diagram shown in Fig. 2, wherein in Fig. 2 time-dependent the course of the voltage applied to the node between the switches Sl and S2 voltage Ui, the lamp voltage U _EL and over the coil L2 flowing current i _{L2 is} shown. In particular, in Fig. 2 the case is shown that during a first period Tl of the circuit arrangement shown in Fig. 1, the bridge diagonal with the switches S2 and S3 is activated, whereas during a subsequent period T2, the bridge diagonal with the switches Sl and S4 is activated , That is, during the period Tl, the switch S3 is permanently closed, and the switches Sl and S4 are permanently open. Furthermore, the switch S2 is switched on and off in a high frequency alternately during this time period T1. From Fig. 2 it is particularly apparent that the switch S2 is always closed when the over the current flowing through the coil L2 L2 _{L2 has reached} its lower reversal point, ie its minimum value, so that the pulse-like course of the voltage U ₁ results. The slope of the edges of the current i _L2 is determined by the inductance of the coil L2. By changing the peak value of the current i _L2> ie the timing of the opening of the switch S2, the current average value of the current i _{L2 can be} changed and thus the power supplied to the lamp EL and its color temperature can be regulated or kept constant. The high-frequency course of the current i _L 2 is smoothed by the components _L 2 and C 2, so that the smoothed curve of the voltage U _EL applied to the gas discharge lamp EL is shown in FIG. 2.

After expiration of the time period Tl, the switches S2 and S3 are permanently opened, and the switch S4 is switched on permanently. Analogous to the switch S2 during the

Period Tl now the switch Sl high-frequency alternately on and off, so that the in Fig.

2 shows the course of the voltages ui and U _EL as well as of the current i _L 2. As already mentioned, a control circuit is used to switch repeatedly between the operating phases during the periods T 1 and T 2, these being

Umpolfrequenz especially in the range 80 - 400 Hz may be, while the high-frequency clock frequency of the switch S2 (during the period Tl) and the switch Sl

(During the time period T ₂ ) may be in the range around 90 kHz.

Due to the low-frequency switching or polarity reversal between the bridge diagonals Sl - S4 and S2 - S3 inevitably creates a hum, which due to its low Frequency in itself is relatively quiet and not disturbing. Due to the steep edges at the switching time between the periods Tl and T2, however, harmonics occur, which have a disturbing effect. For this reason, the control circuit, which controls the switches S 1 _-S 4, advantageously designed such that it reduces the current peaks of the current i _L 2 before and after switching between the operating phases Tl and T ₂ . This can be done, for example, by a special software or by a special adaptation of the hardware of the control circuit 5, which reduces the last current peaks during the period Tl and the first current peaks during the period T ₂ , in this way the edges when switching between the operating phases Tl and T ₂ flatten. In this case, the dashed lines shown in Fig. 2 results in the current i ^ _Σ and the lamp voltage u _EL . From this dashed representation it can be seen that before and after the switching time, the current peaks are slightly reduced compared to the original curve and thus a slightly softer transition of the lamp voltage U _{EL is} achieved.

In the control just described, after opening the high-frequency switched switch, the current continues to flow across the free-wheeling diode and decreases relatively slowly when the second switch of the bridge diagonal that is currently activated remains closed. This leads to a smaller current peak value and accordingly also to a smaller power loss. However, it may happen that at a time when the electrons have been eliminated from the barrier layers of the freewheeling diodes and thus the lower reversal point of the current i _L2 has been reached, this has not fallen off sufficiently and thus the switches are still exposed to a high load when closing. In order to eliminate these stresses, the switches can be controlled in accordance with the diagram in FIG. 2b in a further development.

This diagram shows the current profile i _L2 and the state of the second and the third switch 2, 3 during the period Tl. The two other switches are open in this period Tl. During a first phase [tau] l both switches are closed and the current i _L2 increases continuously. As with the control just described, during a second phase [tau] 2 ₂ , the beginning of which may be determined by reaching a maximum value of i _L2 or by a predetermined duration of [tau] 1, the second switch S2 is opened and i _L2 decreases slowly. In addition, however, the third switch S3 will now also be opened at a predetermined point in time after the opening of the second switch S2 in a third phase X3. The current now flows through the two freewheeling diodes of the first and fourth switches and now decreases more than during the second phase [tau] 2. This ensures that i _L 2 actually reaches a negative value before the barrier layers of the freewheeling diodes are cleared out. When i _L 2 reaches the lower reversal point, both switches are closed again and the controller is again in the state of the first phase [tau] l. The opening of the third switch S3 - ie the third phase [tau] 3 - is omitted, however, if the current i _{L2 has} previously fallen to zero, since in this case no high loads when opening switch occur. Instead, the first phase [tau] l, is continued immediately, and the second switch S2 is opened again. The low-frequency switching between the two bridge diagonals is analogous to the previous embodiment, wherein also here advantageously the current peaks of the current i _L2 before and after the switching between the operating phases Tl and T ₂ can be reduced.

It is a known feature of high pressure gas discharge lamps that they have a relatively poorly controllable and unstable behavior until complete heating. The complete heating occurs approximately after 1 - 2 minutes. In the

Warm-up phase, the voltage across the lamp may be lower than in normal operation. If the ballast were operated in the warm-up phase as in the normal operation described above, the reduced lamp voltage would result in a current i _L2 flowing through the inductance L2 with a correspondingly small slope _di.sub.L2 / dt, so that the reversal point of i.sub.1, if applicable, would be sufficient _L 2 can not be reliably detected by means of the aforementioned transformer tap. It is therefore advantageous, during the warm-up phase, ie after ignition and before normal normal operation, to clock high-frequency the switches S3 and S4 analogously to the switches S1 and S2, with low-frequency switching between the bridge diagonals S1-S2 and S2 S3, that is, it is switched low frequency between two states, wherein in the first state, the switches S1 and S4 clocked high-frequency and the switches S2 and S3 are open, while in the second state, the switches S2 and S3 clocked high-frequency and the switches Sl and S4 are open. By this measure it is achieved that over the freewheeling diodes of the switches S4 and Sl, a current flows through the coil L2, whereby in the transformer coupled to this coil L2 and not shown in Fig. 1 winding, which detects the reversal point of the current i _L 2 is provided, a higher voltage is generated, so that a reliable detection or monitoring of the current i _L 2 is possible. In particular, the switching time can be monitored exactly. The change from the warm-up phase to normal operation takes place after reaching the operating temperature of the lamp, for example, after exceeding a threshold (about 45V) by the lamp voltage, preferably still waiting for a certain period of time until the actual switching.

3 shows the use of the circuit arrangement shown in FIG. 1 according to the present invention in an electronic ballast for operating gas discharge lamps, in particular high-pressure lamps.

Gas discharge lamps.

On the input side, the electronic ballast on a radio-interference filter with a balancing transformer L4, L5 and capacitors C3 and C4, which are connected to a current-carrying conductor L, a neutral conductor and a ground conductor of a supply voltage network. The radio noise filter is connected to a 5 rectifier comprising diodes D5 - D8. This rectifier is followed by a circuit which acts as a boost converter and resistors R2 - R6, capacitors C5 and C6, a diode D9, a transformer L6, L7, a field effect transistor Sβ and one of a Supply voltage VCC supplied integrated control circuit 4, which in particular the field effect transistor S6, which serves as a switch, using a pulse width modulated signal depending on the tapped off at the resistor R3 voltage. In this way, it is achieved that the times in which the transistor S6 is conductive during a power half-wave are controlled so that the envelope of the absorbed current is substantially sinusoidal. This output voltage is rectified by means of the diode D9 and sieved with the aid of the capacitor C6, so that the already explained with reference to FIG. 1 DC supply voltage U _{0 is provided} for the operation of the gas discharge lamp EL circuit provided. On the output side, the electronic ballast shown in Fig. 3 comprises the circuit arrangement shown in Fig. 1, wherein the corresponding components are provided with identical reference numerals, so that a repetition of the description of these components can be dispensed with. In addition, however, it should be noted that FIG. 3 also shows the previously mentioned winding L3, which is transformer-coupled to the inductance L2 in the bridge branch of the full-bridge and for detecting the reversal point of the current i _L 2 (see FIG ) serves.

Furthermore, in Fig. 3, the central control circuit 1 is shown, which is fed by a supply voltage VDD and on the one hand with the aid of the coil L3 the reversal point of the current i _L2 and detected by means of the voltage picked up at the resistor Rl the instantaneous height of the current i _L 2 , Furthermore, this control circuit 1 monitors, in particular as a user-specific Integrated Application Specific Integrated Circuit (ASIC) circuit can be designed, the voltage applied at the tap point of the coil Ll of the series resonant circuit, with the aid of which the ignition of the gas discharge lamp EL can be detected. The outputs of the control circuit 1 are coupled to bridge drivers 2 and 3, which each serve to drive the field effect transistors Sl and S2 or S3 and S4. The also serving as a switch field effect transistor S5, which is connected in series with the resonant circuit capacitor Cl, is driven directly by the control circuit 1.

FIG. 6 shows a further embodiment of the present invention. As is known, the invention requires that in a full-bridge circuit with four switches (now called A, B, C and D in FIG. 6), one of the bridge diagonals A, B or C, D can be activated alternately by one of the switches A , B or C, D high-frequency and the other of the same bridge diagonal is clocked low frequency.

While in the exemplary embodiment of FIGS. 1 and 5 it has been explained by way of example that switches S1 and S3 are clocked at high frequency, reference will now be made to FIG. 6 to explain that the switches A, C are each high-frequency and switches B, D are each clocked low frequency.

In particular for the switches (here A, C), which are clocked at high frequency, field-effect transistors (FETs) with a fast body diode, ie a body diode with a short evacuation time, have recently been used. Out Cost reasons are on the low-frequency side (in the example of Fig. 6, the switches B, D) normal field effect transistors used, which thus have no diode with fast Ausräumzeit. Meanwhile, a field effect transistor with a body diode with a short evacuation time can also be provided on the low-frequency side.

The following procedure can occur:

In a phase 1, both the switch A and the switch B is turned on, provided that in this example the bridge diagonal A, B is activated. Thus, as shown in FIG. 6, a current flows from the supply voltage V _bus via the switch A through the load circuit to the lamp and then via the switched-on switch B to ground. The current through the bridge branch with the lamp increases continuously until a switch-off condition for the high-frequency clocked switch A is reached. After reaching the off condition, the switch A is turned off, while the low-frequency clocked switch B is still turned on. This results in a current flow in a phase 2, in which therefore the coil L continues to drive the current through the lamp and through the closed switch B. This current decreases continuously until it reaches its point of reversal, and then normally, upon reaching the reversal point of the bridge current, the switch A is turned on again and the phase 1 process begins again.

If, on the other hand, the current in phase 2 is not lowered to its minimum after a predetermined time, according to the invention, in addition to the open switch A, too the switch B is opened, so that the current waveform according to phase 3 is formed. In this phase, the current now flows through the body diodes of the switches C and D and thus drops accelerated to its minimum value. This phase 3 will now take until the body diodes of the switches C, D are eliminated. However, the problem may arise that the body diodes of the high-frequency side (switches A, C) have shorter evacuation times than the body diodes of the switches B, D, ie the switch of the low-frequency clocked branch.

In this case, therefore, the midpoint voltage U _x at the high-frequency clocked branch will commutate faster than the midpoint voltage U _γ in the case of the low-frequency clocked branch. This can cause the problem that on the high-frequency clocked switching side, the reclosing condition for each activated high-frequency clocked switch is reached and the switch is turned on, while the body diode is not completely eliminated on the low-frequency clocked bridge branch. Thus, in the prior art, there is a so-called hard switching of the transistor on the low-frequency clocked side, since according to the prior art the reconnection of the additionally opened switch on the low-frequency clocked side comes simultaneously to the reconnection of the switch of the same activated bridge diagonal, the high-frequency is clocked. Under a hard switching is to be understood that switching on the low-frequency clocked switch of the activated bridge diagonal is not power.

The invention now addresses this problem and presents a technology that prevents this hard switching can be.

According to the invention, in particular the restarting of the additionally opened low-frequency clocked switch of a bridge diagonal is now delayed beyond the occurrence of the reclosing condition for the high-frequency clocked switch.

Between the triggering of the switching operation by a control unit detecting the reclosing condition and the actual physical switching operation, i. the current carrying the switch are circuit inherent delays, for example, caused by driver circuits and may typically be of the order of 0, lμS and lμS.

According to the invention, the control unit selectively waits for a predetermined, in particular programmed, time duration after the reclosing condition has occurred, before it triggers the switch-on process by outputting a switching signal, wherein the actual closing of the switch then occurs, as explained above, with a further delay. The switching signal thus appears at an output of the control unit only after the expiration of the delay time.

This predetermined period of time is "deposited" in the control unit itself and thus not the product of external delay effects that can not be definitely calculated due to different components.

Preferably, of course, the triggering of the reconnection of both the high-frequency clocked Switch as well as the low-frequency switch (if it has been opened to lower the current faster as explained above) an activated bridge diagonal delayed by the control unit beyond the occurrence of the reconnection condition for the high-frequency switch clock.

Thus, if the reclosing condition is to reach the minimum of the bridge branch current (which may be accomplished by monitoring the bridge branch current or a variable dependent thereon), then both the high frequency clocked switch and the additionally opened low frequency clocked switch of an activated bridge diagonal will be switched back to a period of time, the time after reaching the reversal point (minimum value) of the bridge branch current.

Other reclosing conditions, such as the increase in the center voltage at the junction of two switches of a half-bridge over a predetermined threshold also.

This is shown schematically in FIG. 8, where it can be seen that the switch-on condition has already been reached, the switches A, B being switched on synchronously (in the case of the activated bridge diagonals A, B) by the control unit but with a delay of, for example, 0, 5 μs is triggered. This delay in triggering the switching operation ensures that the center voltages U _x and U _γ (see FIG. 6) are at the same potential when the two switches A, B or C, D are actually switched on. This is no longer true the hard switching of the low-frequency clocked switch.

In order to achieve this powerless switching, a delay value can be set, for example, between 300 ns and 2.5 μs, but preferably below 1 μs.

The timing of closing the low frequency clocked switch may be after the high frequency clocked switch is closed when the lamp voltage, output voltage or other voltage monitored in the output circuit such as the midpoint voltage on the low frequency clocked switch side (switches B and B at low frequency) D this is the midpoint voltage Uy) is below a predetermined threshold. Since the center point voltage U _x for detecting the reclosing condition is also monitored, by monitoring the midpoint voltage U _γ can be easily determined.

In addition to the time at which the low-frequency clocked switch is switched on again, the time at which the low-frequency clocked switch opens (ie the transition from phase 2 to phase 3 in FIG. 6) can also be set according to the invention.

Both the setting of the delay time of the triggering of the restarting and the setting of the timing off the low-frequency clocked switch, for example, the control unit from the outside, for example, be predetermined depending on the type of transistors used. In particular, the delay of the reconnection of the low-frequency clocked switch after entering the Restart condition for the high frequency clocked switch may be programmable.

As is clear from a comparison of Figures 1 and 9, the switch S5 is omitted in Figure 9 and bridged, so to speak, permanently. Furthermore, the capacitor Cl of the series resonant circuit consisting of the autotransformer Ll and just this capacitor Cl is connected to one end to ground.

For this purpose, an additional capacitor CN is provided, on the one hand with the connection point between the inductance Ll and the

Capacitance Cl of the series resonant circuit is connected.

On the other hand, the additional capacitor CN at the connection point between the switches Sl and S2 and the

Connection point between the inductance L2 and the

Capacitance C2 of the smoothing or filtering circuit is provided

(which incidentally forms its own series resonant circuit).

Alternatively, as shown by broken lines in FIG. 9, the additional capacitance CN ^{Λ may} also be connected between a connection point between the inductance Ll and the capacitance Cl of the series resonant circuit and a connection point of the third and fourth switches S3 and S4. This additional capacity CN ^λ , which may thus be provided as an alternative or in addition to the additional capacity CN already explained, is thus connected in parallel with the right branch of the autotransformer L 1 in FIG. 9. The already explained additional capacity CN, however, is connected in parallel to the gas discharge lamp EL and the left branch of the autotransformer Ll, in which the ignition voltage is transformed.

Claims

claims

1. Circuit arrangement for operating a

A gas discharge lamp, comprising a full-bridge circuit to which a DC voltage (U ₀ ) is applied and which comprises four controllable switches (S1-S4), wherein a gas discharge lamp (EL) in a bridge branch, which is a node between the first switch (Sl) and the second switch (S2) to a node between the dr itten S c ha older (S 3) and the fourth switch (S4) connects, is arranged, and with a control circuit (1) which alternately activates one of the two bridge diagonals, respectively consisting of a high-frequency clocked switch and a low-frequency clocked switch, wherein the control circuit (1) triggers the closure of a high-frequency clocked switch a bridge diagonal whenever a measurement signal meets a reclosing condition, characterized in that the control circuit (1) closing the low-frequency and / or the high-frequency clocked switch according to a predetermined delay of the control circuit gszeit after the entrance of the

Restart condition of the high frequency clocked switch triggers.

2. Circuit arrangement according to claim 1, wherein the delay time is at most 3uS, preferably 30OnS to 2.5uS.

3. A circuit arrangement for operating a gas discharge lamp, with a full bridge circuit to which a DC voltage (U ₀ ) is applied and the four controllable switch (S1-S4), wherein a gas discharge lamp (EL) in a bridge branch, which is a node between the first Switch (Sl) and the second switch (S2) to a node between the third door (S 3) and the fourth switch (S4) connects, and to a control circuit (I) _/ alternately activates one of the two bridge diagonals, each consisting of a high-frequency clocked switch and a low-frequency clocked switch, wherein the control circuit (1) always closes a high-frequency clocked switch of a bridge diagonal, when a measuring signal fulfills a reclosing condition, the control circuit ( 1) also the low-frequency clocked switch the same bridge diagonal after a predetermined period (T ₂ ) after the Publ 2 (b), when the reclosing condition for the high-frequency clocked switch has not yet been fulfilled in order to allow the bridge branch current to decrease faster, characterized in that the control unit determines the time of the additional opening and / or the time the subsequent closure of the low frequency and / or high frequency Clock set switch according to an external default setting and / or the time by the control circuit (1) itself adaptive, ie adjusts depending on a defined event.

4. Circuit arrangement according to one of the preceding claims, wherein the time of triggering the closure of the low-frequency clocked switch is delayed only after the reclosing condition of the high-frequency clocked switch when the control unit supplied lamp voltage or the output voltage is below a predetermined threshold.

5. A method for operating a gas discharge lamp, with a full bridge circuit to which a DC voltage (U ₀ ) is applied and the four controllable switch (S1-S4), wherein a gas discharge lamp (EL) in a bridge branch, which is a node between the first Switch (Sl) and the second

Switch (S2) with a node between the third switch (S3) and the fourth switch (S4) connects, and wherein alternately one of the two bridge diagonal is activated, each consisting of a high-frequency clocked switch and a low-frequency clocked switch, wherein a high frequency clocked switch of a bridge diagonal is always closed when a measurement signal fulfills a reclosing condition, characterized in that the triggering of the closing of the low frequency and / or the high-frequency clocked switch takes place only after a defined delay time after the occurrence of the reconnection condition of the high-frequency clocked switch.

6. A method for operating a gas discharge lamp, comprising a full bridge circuit, to which a DC voltage (U ₀ ) is applied and the four controllable switch (S1-S4), wherein a gas discharge lamp (EL) in a bridge branch, which is a node between the first Switch (Sl) and the second switch (S2) to a node between the third switch (S3) and the fourth switch (S4) connects, and wherein alternately one of the two bridge diagonals, each consisting of a high-frequency clocked switch and a Low-frequency clocked switch, is activated, wherein the control circuit (1) always closes a high-frequency clocked switch of a bridge diagonal when a measurement signal meets a reclosing condition, whereby the low-frequency clocked switch the same bridge diagonal after a predetermined period (T ₂ ) after opening the high-frequency clocked switch is opened (Figure 2b), we nn until then the reconnection condition for the high-frequency clocked switch is not yet met, to let the bridge branch current fall faster, characterized in that the time of additional opening and / or the time of the subsequent closure of the low-frequency and / or high-frequency clocked switch an external default is set and / or adaptive, ie is set event-dependent.

7. The method of claim 6, wherein the timing of triggering the closing of the low-frequency clocked switch is delayed after the reclosing condition of the high-frequency switch only when the control unit supplied lamp voltage or the output voltage is below a predetermined threshold.

8. Control unit, in particular integrated circuit, in particular ASIC, which is designed to carry out a method according to one of claims 5 to 7.