METHOD FOR DRIVING NON-LINEAR LOAD ELEMENTS

Abstract

A method for driving a non-linear load element. On account of the non-linear interrelationship between the voltage and the current at the load element and the related non-linear dependence of the power loss on the quantities “voltage” and “current”, an adjustment of the switching speed only on the basis of the power loss in the switching element cannot be carried out with non-linear load elements without being confronted with undesirable switching losses and related electromagnetic noise fields. Therefore, the load current currently flowing in the load element is picked up in addition to the currently determined power loss in the switching element, and the switching speed of the switching element is controlled in dependence on the determined power loss and on the current picked up. The switching speed can be optimally adjusted when driving the non-linear load elements by means of PWM.

Description

The invention relates to a method for driving a non-linear load element according to the preamble of Claim 1.

Many electric load elements, particularly electric load elements in motor vehicles, such as lamps, heating spirals etc., are driven by means of pulse-width modulation (PWM), wherein the power delivered to the load element can be regulated or controlled, wherein it is possible to minimize the losses in the drive electronics by the switching operations.

However, when driving load elements in motor vehicles by means of pulse-width modulation, electromagnetic fields that may interfere with the radio reception in the vehicle are emitted via the battery supply lines and the load supply lines. Therefore, appropriate limiting values have been laid down in various standards, such as IEC, ISO, CISPR, said limiting values reducing the interference with the radio receiver in the corresponding spectra to a tolerable degree. These emitted fields may be reduced by filtering at the inputs and outputs of the control device, for example. The new methods actively influence the switching edges, as it is described in the unexamined application WO2005/057788, for example.

In the new series of vehicles, the lamps are being increasingly replaced by light emitting diodes (LED). However, the non-linearity of their current-voltage characteristic results in a sudden break-off of the current and so in increased interfering emissions.

Conventional methods attenuate the high-frequency alternating currents in the supply lines by means of filters in the input lines and output lines. However, the disadvantage of the filters consists in the facts that they are very expensive and require a lot of space, whereby they raise the price of the electronic components, and that they cannot be miniaturized (integration in silicon).

Another possibility of reducing electromagnetic radiation is the reduction of the switching speed in the switching element, whereby the high-frequency current portions can be reduced to the necessary degree, but here the undesirable switching losses heating up the electronic components are increasing with decreasing switching speed.

In the new methods, for example according to WO2005/057788, the switching speed of the switching element is varying in dependence on the instantaneous power loss. FIG. 1 shows the normalized power loss in a switching element when driving an ohmic load element (linear load element) as well as a stepped convergence of the course of the rate of change of the output voltage according to WO2005/057788.

However, such methods, as disclosed in said application, fail when driving load elements that show a non-linear behaviour within a switching process (load elements with a non-linear voltage-current interrelationship, such as LEDs).

On account of the non-linear interrelationship between the voltage and the current in the load element, the power loss in the switching element is not linearly dependent on the output voltage or the load current, respectively. Therefore, when driving the non-linear load elements, an adjustment of the switching speed of the switching element that is only related to the power loss or output voltage at the switching element or to a quantity depending thereon is not applicable without being confronted with increased switching losses or a bad electromagnetic radiation, respectively. This difference between linear and non-linear load elements is illustrated in FIG. 5, for example. With an ideal linear load element (e.g. linear resistor), the current changes proportionally to the voltage at the load element (see dashed line L1 in FIG. 6). Thus, with a linear load element, there is a quadric-polynomial interrelationship between the power loss P_V at the switching element and the output voltage U_a: P_V˜(U_a)², i.e. the power loss P_V is linearly proportional to the square of the output voltage U_a, as shown in FIG. 1 (see continuous polynomial curve).

On the other hand, with a non-linear load element, the current flowing in this non-linear load element does not change proportionally to the output voltage (see continuous line L2 in FIG. 6). Therefore, there is no linear interrelationship between the power loss and the current or the voltage, respectively. This is illustrated in FIGS. 2 and 8, wherein FIG. 2 shows a non-linear interrelationship in the L-range of the output voltage and FIG. 8 a non-linear interrelationship both in the L-range and in the H-range of the output voltage. As a result of the non-linear interrelationship between the voltage and the current, an adjustment of the switching speed only on the basis of the power loss or the output voltage is not practicable with the non-linear load elements, which means that the radiation of electromagnetic noise fields cannot be reduced effectively by means of the method according to WO2005/057788.

It is therefore the object of the present invention to set forth a method of the above-mentioned type, by means of which the radiated electromagnetic fields with the non-linear load elements can be reduced effectively by means of a load-dependent, active influencing of the switching edges of the switching device and the valid standards can be met.

This object is achieved by the characterizing features of Claim 1. The advantageous further developments of the invention are subject matter of the subclaims.

On account of the non-linear interrelationship between the voltage and the current at the load element and the related non-linear dependence of the power loss on the quantities “voltage” and “current”, an adjustment of the switching speed only on the basis of the power loss or the voltage, respectively, cannot be carried out with non-linear load elements without being confronted with undesirable switching losses and related electromagnetic noise fields.

Therefore, according to the invention, the load current currently flowing in the load element is picked up in addition to the currently determined power loss at the switching element, and the switching speed of the non-linear load elements is controlled in dependence on the determined power loss and on the measured load current.

By including the current flowing in the load element as a further measured quantity in addition to the quantity “power loss”, the switching speed of the non-linear load elements can be optimally adjusted, wherein the switching speed of the non-linear switching element is controlled as follows:

• • the switching speed is set high in the range where the current-voltage characteristic of the load element dI_a/dU_a is approximately zero;
  • the switching speed is set high in the range where the current-voltage characteristic of the load element dI_a/dU_a is not approximately zero and the power loss is high; and
  • the switching speed is set low in the range where the current-voltage characteristic of the load element dI_a/dU_a is not approximately zero and the power loss is low.

The present invention is based on an active influencing of the switching speed of the switching element in such a manner that the switching element is operated at a high switching speed in that range where both the output voltage and the current in the load element have exceeded the respective lower threshold. Accordingly, the switching element is operated at a reduced switching speed when the output voltage and the current in the load element are below the respective threshold.

For optimally adjusting the switching speed, two thresholds each are used for the power loss at the switching element and for the load current. For the power loss, the output voltage is picked up as a measured quantity, and two voltage thresholds are used as a reference quantity for the output voltage.

In the following, the invention will be explained in greater detail on the basis of the exemplary embodiments and with the aid of the figures in which

FIG. 1: shows the relative power loss in dependence on the relative output voltage with a linear load element, shown in a right-handed coordinate system, wherein the right-hand abscissa axis is the relative output voltage U_a/U_KL30 and the upward ordinate axis is the relative power loss P_V/P_{V max};

FIG. 2: shows the relative power loss in dependence on the relative output voltage with a non-linear load element, shown in a right-handed coordinate system, wherein the right-hand abscissa axis is the relative output voltage U_a/U_KL30 and the upward ordinate axis is the relative power loss P_V/P_{V max};

FIG. 3: shows the variations in time/the switching behaviour of the relative power loss, of the relative load current and of the relative output voltage with a non-linear load element according to FIG. 2, shown in three right-handed coordinate systems, wherein the first coordinate system shows the variation in time of the relative load current I_a/I_{a max}, the second coordinate system shows the variation in time of the relative output voltage U_a/U_bat, and the third coordinate system shows the variation in time of the relative power loss P_V/P_V-max;

FIG. 4: shows an equivalent circuit diagram of a circuit arrangement for carrying out the method according to FIG. 3;

FIG. 5: shows the course of the normalized current-voltage characteristic of a linear and of a non-linear load element, wherein the maximum current with this non-linear load element has an upper limit, shown in a right-handed coordinate system, wherein the right-hand abscissa axis is the relative output voltage U_a/U_KL30 and the upward ordinate axis is the relative load current I_a/I_{a max};

FIG. 6: shows the variations in time/the switching behaviour of the relative power loss, of the relative load current and of the relative output voltage with a non-linear load element, wherein the maximum current with this load element is limited according to FIG. 5, shown in three right-handed coordinate systems, wherein the first coordinate system shows the variation in time of the relative load current I_a/I_{a max}, the second coordinate system shows the variation in time of the relative output voltage U_a/U_bat, and the third coordinate system shows the variation in time of the relative power loss P_V/P_V-max;

FIG. 7: shows the variations in time/the switching behaviour of the relative power loss, of the relative load current and of the relative output voltage with a non-linear load element according to FIG. 8 with the current limitation according to FIG. 5, shown in three right-handed coordinate systems, wherein the first coordinate system shows the variation in time of the relative load current I_a/I_{a max}, the second coordinate system shows the variation in time of the relative output voltage U_a/U_bat, and the third coordinate system shows the variation in time of the relative power loss P_V/P_V-max;

FIG. 8: shows the relative power loss in dependence on the relative output voltage with a non-linear load element with the current limitation according to FIG. 5, shown in a right-handed coordinate system, wherein the right-hand abscissa axis is the relative output voltage U_a/U_bat and the upward ordinate axis is the relative power loss P_V/P_{V max};

FIG. 9: shows an equivalent circuit diagram of a circuit arrangement for carrying out the method according to FIG. 7;

FIG. 10 a: shows the switching behaviours in an inventive method for adjusting the switching speed of the switching element when driving a non-linear load element, with a course of the relative power loss relative to the relative output voltage according to FIG. 8 according to Table 1, shown in four right-handed coordinate systems, wherein the first coordinate system shows the variation in time of the relative gate voltage U_g/U_{g max}, the second coordinate system shows the variation in time of the relative output voltage U_a/U_bat, the third coordinate system shows the variation in time of the relative load current I_a/I_{a max}, and the fourth coordinate system shows the variation in time of the relative power loss P_V/P_V-max;

FIG. 10 b: shows the variation in time of the relative load current I_a/I_{a max} with the associated pulse duration for the switching behaviour according to Table 1 and FIG. 10a;

FIG. 10 c: shows the switching behaviours in an inventive method for adjusting the switching speed of the switching element when driving a non-linear load element, with a course of the relative power loss relative to the relative output voltage according to FIG. 8 according to Table 2, shown in four right-handed coordinate systems, wherein the first coordinate system shows the variation in time of the relative gate voltage U_g/U_{g max}, the second coordinate system shows the variation in time of the relative output voltage U_a/U_bat, the third coordinate system shows the variation in time of the relative load current I_a/I_{a max}, and the fourth coordinate system shows the variation in time of the relative power loss P_V/P_V-max;

FIG. 1 shows the change of the relative power loss P_V/P_{V max} in dependence on the relative output voltage U_a/U_KL30 with a linear load element, wherein P_V is the instantaneous power loss, P_{V max} is the maximum power loss, U_a is the output voltage, and U_KL30 is the supply voltage. According to that, the relative power loss P_V/P_V-max shows a quadric polynomial of the output voltage U_a/U_KL30, which is a result of the linear interrelationship between the output voltage U_a and the current I_a in a linear load element.

FIGS. 2 and 8, on the other hand, show a non-linear interrelationship between the power loss P_V and the output voltage U_a in a non-linear load element. This non-linear interrelationship between the power loss P_V and the output voltage U_a is a result of the non-linear interrelationship between the output voltage U_a and the current I_a in the load element, which is the case with a non-linear load element. Consequently, such a non-linear load element, such as an LED (light emitting diode), cannot be operated by means of a method as described in WO2005/057788 without causing increased interfering radiation or increased switching losses.

As shown in FIG. 2, the curve of the relative power loss P_V/P_V-max has a range (range B1 in the figure) where the power loss P_V is zero in spite of the increasing output voltage U_a. Outside this range B1, in range B2, the relative power loss P_V/P_V-max shows an approximately linear interrelationship with the relative output voltage U_a/U_KL30. The transition point between these two ranges B1 and B2 is the lower voltage threshold U_au of the output voltage U_a that is still to be determined.

Since the voltage threshold U_au of the LED is not exactly known or is varying in dependence on the operational conditions, the current currently flowing in the load element (load current I_a) is introduced as a new physical measured quantity as against the known method according to WO2005/057788. Also, a current threshold I_lim1 is introduced for the currently flowing load current I_a.

So, the switching element is operated at a high switching speed in that range where both the output voltage U_a and the load current I_a have exceeded the lower thresholds U_au and I_lim1, respectively. Accordingly, the switching speed is reduced when the output voltage U_a and the load current I_a are below the thresholds U_au and I_lim1, respectively. The variations in time are illustrated in FIG. 3. The range of high power loss is dependent on the voltage threshold U_au of the LED.

FIG. 2 shows the basic course of the relative power loss P_V/P_V-max in the switching element. As in the method that is already known, the switching speed is adjusted to the instantaneous value of the power loss P_V, wherein the dashed curves S1 and S2 represent a stepped convergence of the switching speed to the ideal curve with different numbers of steps, whereby the emitted spectrum can be reduced in the high-frequency ranges also for non-linear load elements without significantly increasing the switching losses.

FIG. 4 shows a possible realization of this switching process according to FIG. 3 with a circuit arrangement. As the non-linear load element, an LED with a series resistor 110 is shown. In this circuit arrangement, when switching, the gate of the MOSFET switching element 200 is supplied with currents of varying amperage in dependence on the power loss P_V, namely the output voltage U_a, and on the load current I_a, and so the switching speed, within the switching process, is specifically adjusted to the instantaneous values of the measured quantities power loss P_V/output voltage U_a and load current I_a.

The switching speed is set high in the range of high power loss P_V and in the range of small change dI_a/dU_a of the load current I_a compared to the change of the output voltage U_a. The switching speed is set low in the range of low power loss P_V and of great change dI_a/dU_a of the load current I_a compared to the change of the output voltage U_a.

By means of the controllable power sources 400 and corresponding control logic 300, the transitions between the high and the low switching speeds can be preset very precisely and, if necessary, a very fine adjustment of the switching speed be realized. The functioning of the control logic 300 is illustrated in Tables 1 and 2.

In order to extend the service life of the LEDs and/or to control the brightness/the colour spectrum independently of the supply voltage, it is also usual to operate LEDs in series with power sources 120 for current limitation. The continuous line L2 in FIG. 5 shows an exemplary course of the normalized current-voltage characteristic of such a load element. When driving such a load element, current limitation results in the following variations in time as in FIG. 6. The shown variations in time illustrate that the spectrum of the current includes considerable high-frequency portions. These portions can be reduced by reducing the switching speed shortly before reaching the threshold current of the current limitation.

Since the current threshold of the current limitation is not exactly known or is varying in dependence on the operational conditions, this current threshold has to be determined for the next falling/rising edge during operation, unless the course of the current has already been rounded off in the load element, wherein the lower current threshold I_lim1 is a quantity that is dependent on the maximum current of the switching element and/or on the instantaneous load.

In addition to the lower current threshold I_lim1 already described, an upper current threshold I_lim2 is defined so that the switching element is operated at a high switching speed only when the current in the load element has fallen below this upper current threshold I_lim2. Accordingly, the switching speed is reduced when the current I_a in the load element has exceeded the upper current threshold I_lim2. The difference ΔI between the current threshold and the maximum current I_max in the load element is defined absolutely or with reference to the maximum current I_max. The variations in time are illustrated in FIG. 7.

The range of high power loss is dependent on the current threshold I_lim2 of the current limitation as well as on the voltage threshold U_ao of the LED. FIG. 8 shows the basic course of the relative power loss in the switching element.

The threshold may be defined in advance or during operation. For determining the upper current threshold I_lim2 during operation, the maximum value I_max of the current is first determined, for example by means of a so-called peak detector or by single or multiple sampling of the current, said sampling being synchronous with the PWM signal. From this maximum value I_max, the current threshold I_lim2 is then determined, as described above, and preferably continuously compared with the instantaneous load current I_a.

However, those periods of time during switching in which the current is almost constant may negatively influence the pulse-pause ratio so that the pulse-pause ratio of driving deviates from that of the load current I_a. As a countermeasure, a further sample-and-hold element may be used for measuring, synchronously with the upper current threshold I_lim2 of the load current I_a, the associated voltage value at the load element. In dependence thereon, the upper voltage threshold U_ao may be defined above this voltage value, above which the switching speed is increased in order to improve the correspondence between the set pulse-duty factor and the actual pulse-duty factor by reducing the dead times. In addition to that, the maximum value of the voltage of the load element of the same pulse or the voltage value occurring simultaneously with the maximum current, respectively, is determined and compared with the lower voltage threshold U_ao. In dependence on the difference between the two values, the associated upper voltage threshold U_ao and the current threshold I_lim2 are accepted as valid values or rejected in order to prevent that wrong thresholds are adopted when the current limitation is not reached on account of a low supply voltage.

Accordingly, when the lower current threshold I_lim1 is reached, the associated voltage value at the load element may be measured and the lower voltage threshold U_au be defined just below it, below which the switching speed is increased as well.

At the beginning, the thresholds U_au, U_ao, I_lim1, I_lim2 are fixed at the respective expected value or its minimum or maximum value, respectively.

FIG. 9 shows an equivalent circuit diagram of a circuit arrangement for carrying out the method for adjusting the switching behaviour according to FIG. 7.

FIGS. 10 a, 10b, 10c show the detailed switching behaviour in an inventive method for adjusting the switching speed in a non-linear load element, with a course of the relative power loss relative to the relative output voltage according to FIG. 8.

FIG. 10 a shows the switching behaviour as shown in Table 1.

TABLE 1
	Gate voltage:	Output	Switching	Power loss
Range	Ug	current: I	speed	in this range
A	Ug < Ugu	I < Ilim1	High	Low
B, C	Ugu < Ug < Ugo	I < Ilim1	Low	Low
D	Ugu < Ug < Ugo	Ilim1 < I < Ilim2	High	High
E, F	Ugu < Ug < Ugo	I > Ilim2	Low	Low
G	Ug > Ugo	I > Ilim2	High	Low

Table 1 and FIG. 10a show a first case, wherein the switching speed is set low in ranges B and F. Here, particularly with the transitions between B and C and between E and F, the thresholds U_au and U_ao are not used yet.

FIG. 10 b shows the course of the current with the associated pulse duration for the switching behaviour according to Table 1 and FIG. 10a, and also the idealized course of the current for driving without pulse shaping and the pulse duration thereof.

It is obvious that the correspondence between the two pulse durations, and thus the duty cycle faithfulness, depends on the symmetry of ranges A, B, C with ranges G, F, E. The difference between the pulse durations results from the difference between the duration of ranges E, F, G and the duration of ranges A, B, C:

ΔT_on=T_on-effektiv−T_on-ideal=t_E+t_F+t_G−t_A−t_B−t_C

Since t_F, the duration of range F, depends on the supply voltage and thus does not have to be constant for a short period of time during operation, this may cause interfering variations of brightness.

For this reason, the switching speed is to be increased in ranges B and F, as it has been done for ranges A and G for the same reason. The result of this is a further improved switching behaviour (see Table 2 and FIG. 10c). According to that, the reduction of the times t_B and t_F also reduces their influence on the error of the pulse duration ΔT_on. Here, the thresholds U_gu and U_go are not used any more. They are replaced by the thresholds U_au and U_ao, respectively.

TABLE 2
	Output		Switching	Power loss in
Range	voltage: Ua	Output current: I	speed	this range
A	Ua < Uau	I < Ilim1	High	Low
B	Ua < Uau	I < Ilim1	High	Low
C	Uau < Ua < Uao	I < Ilim1	Low	Low
D	Uau < Ua < Uao	Ilim1 < I < Ilim2	High	High
E	Uau < Ua < Uao	I > Ilim2	Low	Low
F	Ua > Uao	I > Ilim2	High	Low
G	Ua > Uao	I > Ilim2	High	Low

The switching element is operated at a high switching speed only when the current in the load element has exceeded a lower current threshold I_lim1, as it is the case in range D.

Accordingly, the switching speed is reduced when the current in the load element is below the lower current threshold I_lim1, as at the transition from range D to range C.

The upper current threshold I_lim2 cannot be preset for changing loads or for loads where a change in temperature, for example, results in a variation of the current limitation of the load element. This threshold has to be determined during operation and can then be assumed as being temporarily constant.

For determining this current threshold I_lim2 during operation, the maximum current I_max of a previous pulse is determined, and so the threshold I_im2 has to be just below this maximum value:

I _lim2 =I _max −ΔI

ΔI may be a fixed value or a percentage value δ. If it is a percentage value δ, the upper current threshold I_lim2 is determined as follows:

I _lim2=(1−δ)×I_max.

If the load element is an LED with a series resistor without current limitation as in FIG. 4, the value I_max changes almost linearly with the supply voltage U_bat. However, since the supply voltage U_bat can be measured continuously, it is more useful not to select the current threshold I_lim2, but to select a threshold U_a2 (not U_au or U_ao) with reference to the output voltage U_a that depends on the supply voltage. U_a2 is selected as follows:

U _a2 =U _KL30 −R _DS-on ×I−ΔU,

with

• • U_bat: supply voltage,
  • R_DS-on×I: voltage drop at switching element,
  • ΔU: fixed value, or percentage value relative to supply voltage.

On account of the small voltage drop at the switching element, also U_a2 may be indicated as a percentage value:

U _a2=(1−δ)×U_KL30

The lower voltage threshold U_au is determined as follows. The current threshold I_lim1 is already known. During the cycle of a pulse, the output voltage U_a and the current I are sampled simultaneously. At the point in time t₁, the current I reaches the lower current threshold: I(t₁)=I_lim1. At the same point in time, the output voltage U_a reaches the value U_a(t₁). The voltage threshold U_au is fixed just below this value U_a (t₁): U_au=U_a (t₁) ΔU₁. When fixing the magnitude of ΔU₁, it has to be taken into consideration that the current I is still or already almost zero at the lower voltage threshold U_au.

Similarly, the upper voltage threshold U_ao is determined as follows. The upper current threshold I_lim2 is already known. During the cycle of a pulse, the output voltage U_a and the current I are sampled simultaneously. At the point in time t₂, the current reaches the upper current threshold: I(t₂)=I_lim2. At the same point in time, the output voltage U_a reaches the value U_a (t₂). The voltage threshold U_ao is fixed just above this value U_a (t₂): U_ao=U_a(t₂)+ΔU₂. When fixing the magnitude of ΔU₂, it has to be taken into consideration that the current I is still or already almost the maximum value I_max at the upper voltage threshold U_ao.

List of reference numerals
110                 Light emitting diode (LED) in series with series resistor
120                 Light emitting diode in series with power source
200                 MOSFET switching element
300                 Control logic
400                 Controllable power source
510, 520, 530       Schmitt trigger, comparator
600                 Comparator
700                 Inverter
810, 820            Switch
K1                  Relative power loss PV/PV-max as a function of the relative
                    output voltage Ua/UKL30 with a linear load element
K2                  Relative power loss PV/PV-max as a function of the relative
                    output voltage Ua/UKL30 with a non-linear load element
K3                  Relative power loss PV/PV-max as a function of the relative
                    output voltage Ua/UKL30 with a non-linear load element, wherein
                    the maximum load current has an upper limit
S1, S2              Stepped convergence of the switching speed to the ideal curve K1,
                    K2 and K3, respectively, with different numbers of steps
B1                  Range where the power loss PV is zero in spite of the increasing
                    output voltage Ua
B2                  Range where the relative power loss PV/PV-max shows an
                    approximately linear interrelationship with the relative output voltage
                    Ua/UKL30
B3                  Range where the load current Ia is limited by the power source the
                    light emitting diode 120 is connected in series with
T1                  Separating line between ranges B1 and B2
T2                  Separating line between ranges B2 and B3
L1                  Normalized current-voltage characteristic of a linear load element
L2                  Normalized current-voltage characteristic of a light emitting diode
                    120 that is operated in series with a power source
A, B, C, D, E, F, G Individual range in the switching process

Claims

1.-14. (canceled)

15. Method for driving a load element, wherein an output voltage (Ua) at the load element is controlled at a variable switching speed,wherein during a switching process, a load current (Ia) currently flowing in the load element is picked up several times, and an instantaneous switching speed is controlled in dependence on the load current (Ia),wherein during the switching process, both an instantaneous power loss and the load current (Ia) currently flowing in the load element are picked up several times and the instantaneous switching speed is controlled in dependence on the power loss and the load current (Ia).

16. Method according to claim 15, wherein during the switching process, the load current (Ia) currently flowing in the load element is picked up continuously.

17. Method according to claim 15, wherein during the switching process, both the instantaneous power loss and the load current (Ia) currently flowing in the load element are picked up continuously.

18. Method according to claim 15, wherein the method is a method for driving a load element for the pulse-width modulated driving of a load element with a non-linear interrelationship between the voltage and the current by means of an electronic switching element in the load circuit.

19. Method according to claim 18, wherein during the switching process, an instantaneous output voltage at the load element and/or an instantaneous voltage at a gate of the switching element as well as the load current (Ia) currently flowing in the load element are picked up several times and the instantaneous switching speed is controlled in dependence on the instantaneous output voltage (Ua) at the load element and/or the instantaneous voltage (Ug) at the gate of the switching element and on the load current (Ia).

20. Method according to claim 15, wherein the switching speed is set high in the range where a ratio (dI/dU) of the load current (Ia) to the output voltage (Ua) is approximately zero.

21. Method according to claim 15, wherein the switching speed is set high in a range where a ratio (dI/dU) of the load current (Ia) to the output voltage (Ua) is not approximately zero and a power loss is high.

22. Method according to claim 15, wherein the switching speed is set low in a range where a ratio (dI/dU) of the load current (Ia) to the output voltage (Ua) is not approximately zero, but a power loss is low.

23. Method according to claim 15, wherein at least two current thresholds (Ilim1 and Ilim2) are provided for the load current (Ia).

24. Method according to claim 23, wherein at least two voltage thresholds (Ua, and Uao) are provided for the output voltage (Ua).

25. Method according to claim 24, wherein the voltage thresholds (Uau and Uao) for the output voltage (Ua) are determined in dependence on the thresholds (Ilim1 and Ilim2) of the load current (Ia).

26. Method according to claim 24, wherein a maximum value (Imax) of the load current (Ia) is picked up during a pulse and one or more upper thresholds (Uao, Ilim2) are determined in dependence thereon, and a currently flowing load current (Ia) is compared with the thresholds (Uao, Ilim2) and the switching speed is controlled in dependence thereon.

27. Method according to claim 15, wherein the driving is carried out in such a manner that, in a range of high power loss, the output voltage (Ua) is adjusted to a maximally preset switching speed and the switching speed at a beginning and at an end of the switching process is lower than a maximum switching speed.

28. Method according to claim 24, wherein a voltage value of the load element occurring simultaneously with the lower threshold (Ilim1) of the load current (Ia) is determined and a lower voltage threshold (Uau) is defined in dependence thereon and the switching speed is controlled in dependence thereon so that the switching speed below this voltage threshold (Uau) is higher than a minimally preset switching speed.

29. Method according to claim 26, wherein a voltage value of the load element associated with the maximum current (Imax) is determined and an upper voltage threshold (Uao) is defined in dependence thereon and the switching speed is controlled in dependence thereon so that the switching speed above the upper voltage threshold (Uao) is higher than a minimally preset switching speed.

30. Method according to claim 26, wherein the maximum value of the voltage of the load element, or a voltage value that occurs simultaneously with the maximum current (Imax), is determined and compared with the voltage threshold (Uao) that occurs in a same pulse and depends on the maximum current (Imax) and, in dependence on a difference therebetween, the voltage threshold (Uao) and/or the current threshold (Ilim2) that depends on the maximum current (Imax) is/are adopted or rejected.

31. Device for driving a load element, wherein control logic is provided that controls an output voltage (Ua) at the load element at a variable switching speed,wherein, during a switching process, a load current (Ia) currently flowing in the load element is picked up several times and an instantaneous switching speed is controlled in dependence on the load current (Ia),wherein means are provided that, during the switching process, pick up both an instantaneous power loss (PO and the load current (Ia) currently flowing in the load element and control the instantaneous switching speed in dependence on a power loss (Pv) and the load current (Ia).

32. Device of claim 31, wherein the device is configured for pulse-width modulated driving of a load element with a non-linear interrelationship between a voltage and a current by an electronic switching element in the load circuit.

33. Device of claim 31, wherein during a switching process, the load current (Ia) currently flowing in the load element is picked up continuously.

34. Device of claim 31, wherein during the switching process, said means pick up both the instantaneous power loss (PV) and the load current (Ia) currently flowing in the load element several times.

35. Device of claim 31, wherein during the switching process, said means pick up both the instantaneous power loss (PV) and the load current (Ia) currently flowing in the load element continuously.