METHOD FOR OPERATING A DISCHARGE LAMP AND DISCHARGE LAMP

Abstract

The invention describes a method for operating a discharge lamp by adapting a current signal. A distribution function is defined for gathering several time span values that define several different time spans. The several time span values are determined depending on the distribution function by one or more random numbers. The current signal is commutated at every instant of time according to an expiry of each of the several time spans. A light apparatus is provided with a control unit to perform any method.

Claims

1. A method for operating a discharge lamp by adapting a current signal by performing the following steps: (a) defining and/or providing a distribution function for gathering several time span values that define several different time spans, (b) determining the several time span values depending on the distribution function by one or more random numbers, and (c) commutating the current signal at every instant of time according to an expiry of each of the several time spans.

2. The method according to claim 1, wherein the several time spans are arranged directly adjacent.

3. The method according to claim 1, wherein several distribution functions are used and for each time span value one of those distribution functions is selected according to a further random number.

4. The method according to claim 1, wherein step a), step b) and/or step c) are performed repeatedly.

5. The method according to claim 1, wherein the distribution function is defined depending on one or more discharge lamp parameters, in particular a lamp voltage, a power level of the discharge lamp, a position and orientation of the discharge lamp, a current flow through the discharge lamp, an abrasion degree of a pair of electrode tips and/or the voltage ratio of time spans of opposite polarity.

6. The method according to claim 5, wherein the distribution function is defined dynamically based on a measured quantity or several measured quantities that describe the discharge lamp parameter and/or the parameter of the pair of electrodes of the discharge lamp.

7. The method according to claim 5, wherein the defining of the distribution function is based on a dynamical behavior of an average lamp voltage, a dynamic behavior of the lamp voltage during each time span and/or a dynamical behavior of the current flow through the discharge lamp.

8. The method according to claim 5, wherein several distribution functions are given and the distribution function for determining the time span values is selected based on a threshold value concerning the discharge lamp voltage.

9. The method according to claim 5, wherein at least two different probability functions are given and the distribution function for determining the time span values is a superposition of the least two different probability functions, wherein the superposition is depending on the lamp voltage.

10. The method according to claim 5, wherein the distribution function is defined by a characteristic diagram of the discharge lamp voltage, in particular the distribution function is depending on a threshold value of the discharge lamp voltage.

11. The method according to claim 1, wherein for different types of discharge lamps and/or for different groups of discharge lamps a separate distribution function is defined.

12. The method according to claim 1, wherein the distribution function contains a boundary condition.

13. The method according to claim 12, wherein a probability value with regard to a corresponding predetermined time span value is defined by the boundary condition or several probability values with regard to corresponding predetermined several time span values are defined by the boundary condition, wherein in particular the probabilities of the several time span values are defined by a ratio to each other.

14. The method according to claim 12, wherein a maximum and a minimum value is defined by the boundary condition.

15. The method according to claim 1, wherein the distribution function is superposed with a predetermined function, or one of several predetermined functions, wherein at least one of the several predetermined functions is selected by the at least one random number in order to determine the time span values.

16. The method according to claim 1, wherein distribution function is defined based on a lifetime of the discharge lamp.

17. The method according to claim 1, wherein the current signal is a wave-shaped signal, a square-wave signal or a mixture of wave-shaped and square-waved signal.

18. The method according to claim 1, wherein the distribution function is defined as a uniform distribution, a normal distribution, an exponential distribution, a power law distribution and/or an overlaid normal distribution.

19. A lighting apparatus comprising a discharge lamp with an arc tube with a pair of electrodes, a ballast unit for providing a current signal for the discharge lamp and a control unit that is configured to define and/or provide a distribution function for gathering several time span values that define several different time spans, determine the several time span values depending on the distribution function by one or more random numbers, and commutate the current signal at every instant of time according to an expiry of each of the several time spans.

Description

[0051] In this context the figures show in:

[0052] FIG. 1 a schematic illustration of a discharge lamp;

[0053] FIG. 2 a schematic operation scheme for a discharge lamp;

[0054] FIG. 3 examples of determining or calculating time span values from random numbers;

[0055] FIG. 4 predetermined probability density functions PD1 and PD2 corresponding to the inverse cumulative distribution functions DF1 and DF2 shown in FIG. 3 for calculating the time span values;

[0056] FIG. 5 an output current signal with pseudo-random sequence of several different time spans between commutations based on first distribution function according to FIG. 3;

[0057] FIG. 6 example of a superposition of two different probability density functions depending on the lamp voltage;

[0058] FIG. 7 example of a simple pattern or a building block for constructing a randomized current signal;

[0059] FIG. 8 example of an output current signal with a random sequence of patterns according to FIG. 7, wherein the duration of each pattern is based on the first distribution function according to FIG. 3.

[0060] In FIG. 1, a lighting apparatus 200 with a discharge lamp 100, a control unit 115 and ballast unit 125 as operation unit is shown. The discharge lamp 100 comprises an arc tube 110. Within the arc tube 110, a pair of electrode tips 105 is indicated. Between these two electrode tips 105, an arc discharge may appear. The discharge lamp 100 is able to emit light if a current A flows between the electrode tips 105. Within the arc tube 110, a noble gas, such as helium, argon, crypton, etc., or a metallic gas, such as mercury or natrium, may be present. If the discharge lamp 100 is operated with an alternating current AC at a single frequency, the discharge lamp 100 may suffer from uneven wear and tear. An important aspect of this invention is to avoid such drawbacks. This can be achieved by operating the discharge lamp 100 with a current signal w that is rather random instead of deterministic.

[0061] A random current signal w does not mean that the current signal w can take any possible value for the current flow A. For example, it is possible that the current signal of the lamp current is constructed as a continuous stream of time spans dt of direct current separated by the change of current direction or commutation. The amplitude of the direct current can be fixed according to the voltage of the lamp such that a prescribed nominal lamp power is maintained. The time span values dtv can be chosen randomly according to a prescribed distribution function DF. The distribution function DF may be limited or influenced by boundary conditions. Such boundary conditions may be a minimum or a maximum value for the time span values dtv. Furthermore, the distribution function DF may follow a pre-given distribution. Such a pre-given distribution may be a uniform distribution, a normal distribution or any other function. These functions may also consider physical lamp parameters 120 like the discharge lamp voltage U as well as environmental parameters 120 of the discharge lamp 100. This means that the distribution function DF may consider statistical parameters and/or physical parameters.

[0062] Statistical parameters may be considered by different pre-given distributions, whereas physical parameters may be considered by the lamp voltage U as well as other environmental parameters 120 of the discharge lamp 100. All these parameters may be considered by the distribution function DF. This means that a degree of randomness may be influenced. Since the time span values dtv are calculated by at least one random number ri, the current signal w may be determined randomly. The concrete definition of the distribution function DF may limit the degree of randomness, if necessary. A control unit 115 may gather and/or detect discharge lamp parameters and/or other environmental parameters of the discharge lamp 100 or lighting apparatus 200.

[0063] FIG. 2 shows an exemplary overview of several components of the lighting apparatus 200. The lighting apparatus 200 can comprise the operation unit 125 and a DC/DC converter 10. The current flow A can be detected by a current detector 11 and a voltage detector 12. The lamp operation unit 125 comprises a polarity switch 13. The control unit 115 can switch the polarity by the polarity switch 13. The operation unit 125 can be part of the lighting apparatus 200. The DC/DC converter 10 is used to control the current flow A according to a set value determined by the control unit 115. The set value can be determined based on measurements of the output voltage. Additionally, the control unit 115 can gather values for discharge lamp parameters 120 and/or environmental parameters 120. This means that the control unit 115 is able to measure and/or gather parameters concerning the current flow A and concerning environmental parameters 120 for the discharge lamp 100. An ignition device 14 can be used to create a starting voltage for the discharge lamp 100 at the start of the lamp operation.

[0064] A lamp operation unit 125 (ballast unit) may comprise a random number generator 17. The random number generator 17 may generate a set or stream of random numbers ri in a predetermined range. The predetermined range can be between the values 0 and 1. Several random numbers ri can be generated with a function following a uniform distribution. Usually, the random number generator 17 is based on a uniform probability. This means that the random numbers ri follow a uniform distribution. An adaption of the random number generator 17 is not necessary because physical and/or statistical influences may be considered by a distribution shaping unit 18. In the distribution shaping unit 18, these random numbers ri can be used to calculate values for the time spans dt. The distribution shaping unit 18 and/or the control unit 115 may calculate several time span values dtv from a distribution function DF or a corresponding distribution table.

[0065] In FIG. 3, two different distribution functions are exemplarily shown. The time span values dtv may be passed to a timer unit 19. The timer unit 19 may provide the several time span values dtv for the time spans dt. According to the time spans dt of the time unit 19, the control unit 115 can regulate or operate the polarity switch 13. Since the time span values dtv are influenced by one or more random numbers ri determined by the random number generator 17, the resulting time span values dtv may be randomly generated. This may lead to a randomized current signal w. The randomized current signal w expresses in the form of different time spans dt or durations between a change in polarity or between the commutations of the current flow A. The control unit 115 can switch the polarity or commutate the current flow A after every expiry of a corresponding time span dt. This means that the control unit 115 may commutate the current signal w at every instant of time according to the expiry of every time span dt.

[0066] The resulting output current signal w of the operational unit 125 may be a continuous stream of directly aligned periods of pieces of the current signal w, wherein each piece is represented by its time span dt with the duration of the time span value dtv. Thereby, these periods or pieces may have a different duration according to the several time span values dtv that may be calculated by the control unit 115 and/or the distribution shaping unit 18. According to the underlying distribution functions DF, the degree of randomness may be influenced and/or controlled. In an extreme situation, the distribution function DF may eliminate this random influence. Although this may be possible, such an operation is not intended.

[0067] In FIG. 3, two different distribution functions DF are shown. On the left side, a first distribution function DF1 and on the right side, a second distribution function DF2 is presented.

[0068] The first and second distribution functions DF1, DF2 may assign a time span value dtv to a random number ri. The random numbers ri may be within an interval that is defined by a minimum and a maximum value. The minimum value is indicated by “min”, the maximum value is indicated by “max”. By means of the distribution functions DF1 and DF2, values for the time spans dt or the duration for the time spans dt may be calculated by the random numbers ri. Instead of the distribution functions DF1 or DF2, a table or a lookup table may be used.

[0069] The first distribution function DF1 is an increasing function with increasing random numbers ri. All time span values dtv are within an interval defined by a first and second value T1 and T2 for the time span values dtv. The first value is called T1 and the second value is called T2.

[0070] The second distribution function DF2 shows another behavior. The second distribution function DF2 looks like a function similar to an inverse sign function, also known as arcus sinus. This means that the second distribution function DF2 may lead to other time span values dtv. This is due to the fact that the second distribution function DF2 is completely different to the first distribution function DF1. For example, it is possible that depending on the lamp voltage U, the control unit 115 chooses the first distribution function DF1 or the second distribution function DF2. For example, the second distribution function DF2 may be applied or used if the lamp voltage U is above a pre-given value, a threshold value, for the lamp voltage U. This example can be extended to any other parameter or combination of parameters. Such parameters may be temperature, pressure, inclination, and so on.

[0071] The distribution shaping unit 18 can be controlled by the control unit 115. This distribution shaping unit 18 may contain more than one or two distribution functions or distribution tables. The control unit 115 can determine which distribution function DF or table is used or applied based on quantities, such as the average lamp voltage U, the lamp voltage U depending on a direction of the current A or the dynamic behavior of the lamp voltage U during preceding time spans dt. This means that a selection of one distribution function DF can be depending on said discharge lamp parameters and/or environmental parameters 120 of the discharge lamp 100.

[0072] For example, the first distribution function DF1 may be applied if a growth of the electrode tips is desired. A selection of the first distribution function DF1 may be depending on a lamp voltage U that exceeds a pre-given first threshold value U1 for the voltage. The second distribution function DF2 may be used in order to facilitate a shrinking of the electrode tips 105. The second distribution function DF2 is preferably selected if the control voltage U lies below the first threshold value U1. This means that the first distribution function DF1 is known to allow a growth of the electrode tips 105 and the second distribution function DF2 is known to facilitate a shrinking of the electrode tips 105.

[0073] In FIG. 4, the probability density functions PD1 and PD2 corresponding to the distribution functions DF1 and DF2 are shown. The first distribution function DF1 is the inverse function of a cumulative of the first initial probability function PD1. The second distribution function DF2 is the inverse function of a cumulative of the second initial probability function PD2.

[0074] The probability density function PD1 may rely on a modified power law distribution. It may be expressed by the equations

t−D1=0; for t<T1,

t−D1=K/[t{circumflex over ( )}n]; for T1≤t≤T2,

t−D1=0; for t>T2.

[0075] t−D1 may be considered as a probability density as a function of possible values t for the time span values dtv. K is a normalization constant such that the integral value of t−D1 from t=0 to t=infinity results in unity. The value n is a constant characterizing the probability distribution, particularly the ratio of the probability density values at the boundaries T1 and T2.

[0076] The probability density function PD2 may rely on a normal or Gaussian distribution. It may be expressed by the equation

t−D2=1/(S*sqrt(π)*2)*exp(−½*((t−T3)/S){circumflex over ( )}2

[0077] t−D2 may be considered as a probability density as a function of possible valued t for the time span values dtv. T3 is the mean or expectation value of the distribution. S is the width or standard deviation of the distribution.

[0078] By means of the probability distribution functions PD1, PD2 probability values like p1 or p2 for certain time span values can be modified. This means that predetermined boundary conditions may be realized by appropriate probability functions PD1 or PD2. This would lead to new modified first and second distribution functions DF1 and DF2. The functions DF1 and PD1 are preferably not independent, they relate to each other. For example according to the first probability function PD1 the probability for time span values dtv is zero for time span values larger than T2 and smaller than T1. Therefore, all time span values dtv lie between T1 and T2. The probability is zero concerning time span values dtv outside this interval.

[0079] The second probability distribution PD2 function PD2 shows a local maximum around the value T3. This adequately influenced the shape of the second distribution function DF2. This means that the distribution functions DF1 and DF2 are preferably a result of the probability distribution functions PD1 and PD2. In particular, the distribution function DF1 is an inverse function of an integral over the time of the probability distribution function PD1. The same is true for DF2 and PD2.

[0080] A status of the electrode tips 105 may be estimated or determined by the control voltage U that is used to operate the lighting apparatus 200 or discharge lamp 100. A control voltage U above the first threshold value U1 indicates that the first distribution function DF1 shall be applied, whereas a control voltage below the first threshold value U1 indicates that the second distribution function DF2 should be applied for the operation of the discharge lamp voltage 100.

[0081] FIG. 5 shows a current signal w according to different time spans dt and time span values dtv. The current signal w in FIG. 5 is based on the first distribution function DF1 and accordingly the probability density function PD1. The first probability distribution PD1 influences the distribution of the time span values dtv.

[0082] In FIG. 5 different values for the time spans are expressed by dt−1, dt and dt+1. dt−1 is the preceding time span of dt, whereas dt+1 is the following time span of dt. In case of FIG. 5 the duration of the time spans dt are between T1 and T2 according to the time span values dtv resulting from the first distribution function DF1. The first distribution function DF1 may be adapted or influenced by the first probability distribution function PD1. The same can be valid concerning DF2 and PD2. This first probability distribution PD1 influences the generation of time span values dtv in the distribution shaping unit 18. For example, the control unit 115 may apply the first probability distribution PD1 in order to influence the determining of the time span values dtv.

[0083] For example, the first probability distribution PD1 may be relevant for certain circumstances. For example, control unit 115 chooses the first probability distribution PD1 according to a lamp voltage value that is above or below a threshold value. It is also possible that the control unit 115 selects a second probability distribution PD2. In this case, the distribution of the calculated time span values dtv may follow a normal distribution that is represented by the second probability distribution PD2. Although the random number generator 17 provides uniform distributed random numbers ri, the distribution of the time span values dtv may be influenced by the first or second probability distribution PD1 or PD2.

[0084] It is also possible that the control unit 115 selects one of several probability distributions by means of an additional separate random number ri. In case of FIG. 5, one of the two shown probability distributions may be chosen by a second random number. In this case, each probability distribution has a probability of 50% to be selected by the control unit 115. In this case, a further random factor influences the calculation or determination of the time span values dtv.

[0085] The resulting distribution of calculated time span values dtv may be a smooth blend of all probability distributions according to assigned probabilities for the selection of the probability distributions. This means that another probability distribution can be pre-given or defined in order to select one of the several probability distributions. In particular, it is possible that the predetermined probability distribution for selecting the probability distributions is not uniform. The sum of such probability distribution for the selection is preferably 1 since a selection of a probability distribution may be necessary in order to operate the discharge lamp 100. An example of such selection of probability distributions is the mentioned selection between the first and second probability distribution PD1 and PD2 shown in FIG. 4. Depending on the lamp voltage U, the first probability distribution or the second probability distribution PD2 may be chosen or selected for determining the time span values dtv. It is possible to perform the process of selection for every time span dt.

[0086] In FIG. 6, an example of a third probability distribution PD3 is shown that is a result of a superposition of the first and second probability distribution PD1 and PD2. The superposition may be a linear superposition. In the example of FIG. 6, a blending factor bf is calculated in order to compute the third probability distribution PD3. The calculation of the blending factor bf preferably depends on the lamp voltage U. In this case, the first threshold value U1 and a second threshold value U2 are relevant for the computation of the blending factor bf. The blending factor bf depends on two variable probability factors p1 and p2. The first probability factor p1 represents the probability for the first probability distribution PD1. The second probability factor p2 represents the probability for the second probability distribution PD2. The first and second probability factors p1 and p2 depend on the lamp voltage U.

[0087] If the lamp voltage U is below the first threshold value U1, the first probability factor p1 is 1 and the second probability factor p2 is 0. If the lamp voltage U is larger than a second threshold value U2 concerning the lamp voltage, the second probability factor p2 is 1 and the first probability factor p1 is 0. Between the first and second threshold values U1, U2 concerning the lamp voltage U the blending factor bf is generated by a linear superposition of the first probability factor p1 and the second probability factor p2. At the third threshold value U3, the first and second probability factors are equal. The first probability factor p1 is defined according to the following equation 1.

p1=1−[U−1]/[U2−U1];  equation 1

[0088] The second probability factor p2 can be expressed by the following equation 2.

p2=[U−U1]/[U2−U1];  equation 2

[0089] It is also possible to use a non-linear superposition of the first and second probability distribution. In FIG. 6, the superimposed third probability distribution PD3 is shown above the value for the third threshold value U3 for the voltage U. It can be seen that PD3 is a mixture of PD1 and PD2. According to the example of FIG. 6, the degree of superposition is depending on the lamp voltage U of the discharge lamp 100. The ratio of the first probability distribution PD1 and second probability distribution PD2 for the superposition that leads to the third distribution PD3 may be calculated by the equations 1 and 2. The third distribution PD3 may represent the superposition of the first probability distribution PD1 and second probability distribution PD2.

[0090] A further variation or embodiment of this invention in shown by FIGS. 7 and 8. In this case, a segment pattern or a building block DFB may be used for creating a randomized current signal w. In FIG. 8, the output current A is shown, wherein the output current A is created by a building block DFB according to FIG. 7. The building block DFB of FIG. 7 consists of two segments or parts with equal length and opposite polarity. The building block DFB may follow a pre-given regimentation or rule. In case of FIG. 7, the rule would be to change the polarity in the middle of a time span dt. The rule or regimentation may further be depending on parameters, such as the lamp voltage U, and other lamp parameters 120 of the discharge lamp 100.

[0091] It may be possible that several building blocks DFB can be given and the current signal w may be determined randomly by a random combination of different building blocks DFB. In this case, a further random input to the current signal w is possible. The application of several different building blocks DFB may be seen as a special form of determining the several time span values dtv depending on the distribution function DF. The distribution function DF may be represented by the corresponding building blocks DFB. By the one or more random numbers ri, a certain building block DFB may be chosen or selected from several building blocks DFB. It is possible to assign different building blocks DFB to the different environmental or physical lamp parameters 120. The control unit 115 may select one or more building blocks out of the several building blocks depending on boundary conditions and/or by the at least one random number ri. Preferably, the building blocks DFB comprise a polarity switch so that several time span values dtv may be determined depending on the building blocks DFB as distribution function DF. If the control unit 115 executes the current flow A according to the current signal w derived from the building blocks DFB, a commutation of the current signal w or a current flow A appears.

[0092] A total duration of all building blocks DFB or each building block DFB may be determined randomly according to a prescribed distribution function DF. Further variation may be introduced by a random choice of the patterns or building blocks DFB from a pool of pre-given building blocks DFB.

[0093] Overall, with these methods, a randomness of the current flow A or the current signal w may be created and may help to smooth an interference with image sensors. An acoustical noise related to the current flow A, for example in inductors, may spread over a large spectral width and thereby less perceptible than noise emitted at singular frequencies. The use of the distribution function DF or related functions like the building blocks DFB allow to specify the duration of the time spans dt. This allows a level of control superior to using a simple frequency modulation or temporal multiplexing. At the same time, the random nature of the commutation pattern helps to avoid the appearance of frequency peaks in the current spectrum. This may help to reduce visible interference in recordings made using imaging sensors as any remaining artifacts will be randomly patterned in contrast to embodiments of prior art where artifacts are typically patterned according to heterodyne beat. Also, any acoustical noise related to the lamp current flow A, for instance caused by the current flow A through inductors, will be spread across a large frequency range and is thus considerably less audible. This may help to operate the discharge lamp 100 more efficiently and to increase the quality of the light emission of the discharge lamp 100 as well as its lifetime.

[0094] In general, this invention offers a method and a lighting apparatus 200 with a discharge lamp 100 that uses current signals w which are created or assembled by random influence. The random influence is ensured by a combination of the distribution function DF together with the random number(s). By means of the random number(s) ri values relevant for the current signal w are determined. Since this calculation depends on the random number(s) ri, a certain degree of randomness can be maintained. Nevertheless, the distribution function DF may contain boundary conditions so that a good balance between a deterministic control and randomness can be achieved. This may help to a better discharge lamp operation with respect to quality and/or lifetime.

LIST OF REFERENCE SIGNS

[0095] 200 lighting apparatus [0096] 100 discharge lamp [0097] 110 arc tube [0098] 105 electrode tip [0099] 10 DC/DC converter [0100] 11 current detector [0101] 12 voltage detector [0102] 13 polarity switch [0103] 14 ignition device [0104] 115 control unit [0105] 17 random number generator [0106] 18 distribution shaping unit [0107] 19 timer unit [0108] 125 operation unit [0109] 120 environmental parameters [0110] 100 discharge lamp [0111] dt time span [0112] dtv time span values [0113] T1, T2, T3 first, second, third time value [0114] DF1, DF2 first, second distribution function [0115] DF distribution function [0116] ri random number [0117] A current, current flow [0118] w current signal [0119] PD1, PD2 first probability distribution, second [0120] probability distribution [0121] dt time span, time interval [0122] min, max minimum and maximum value [0123] DFB building block, pattern [0124] U voltage [0125] bf blending factor [0126] p1, p2 first and second probability factor/value [0127] U1, U2, U3 first, second and third threshold value for the voltage