"Method for supplying power to light sources, corresponding power supply unit and light source"

In the name of: OSRAM Gesellschaft mit Beschrankter Haftung, a German company, Hellabrunner Strasse 1 - 81543 Munich (Germany) and OSRAM S.P.A. - SOCIETA' RIUNITE OSRAM EDISON CLERICI, an Italian company, via Savona, 105 - 20144 Milan (Italy)

Named inventors: Tobias FROST, Paolo DE ANNA and Filippo BRANCHETTI

Filed on: February 16, 2010

-k -k -k -k

DESCRIPTION Technical field

The present description relates to methods for regulating light sources .

The present description makes particular reference to the possible application for controlling light sources comprising a plurality of light generators which emit respective chromatically different light radiations that can be mixed with one another to produce a mixed light radiation. The generators in question may be, for example, electrical generators, for example LED light generators, the light emission intensity of which is a function of the (average) supply power, for example a function of the average value of the electrical current flowing through them. These are therefore light generators which are susceptible to selective activation, i.e. in on/off mode, with respective supply powers over respective activation intervals included in an activation period. Description of the prior art

In LED lamps, it is common practice to mix radiations of different color (i.e. "chromatically different" radiations) produced by individual LED generators so as to obtain, as a result of the mixing, a light radiation with chromatic characteristics corresponding to specific coordinates Cx, Cy in the CIE 1931 diagram or to a specific color temperature value or CCT (Correlated Color Temperature) .

In practice, this corresponds to mixing a certain set of colors originating from different generators so as to obtain the desired color. The result of the mixing action (i.e. the coordinates Cx, Cy or the CCT of the radiation resulting from the mixing) may be regulated by modifying the contribution of each mixed color, i.e. by varying the intensity of the radiation emitted by each of the generators; this result may be obtained, for example, by varying the duty cycle of the power supply (i.e. the activation interval) of the generator, that is to say by operating in accordance with a pulse width modulation (PWM) diagram, in ways which are wholly analogous to those commonly used to carry out the function known as "dimming". In a light source (or "lamp") having a number of light generators (or "channels") subjected to mixing, a PWM modulation function can therefore be used:

i) both to vary the comprehensive intensity of the radiation emitted by the lamp as a whole - by intervening on all the channels, i.e. on all the light generators active in the lamp - keeping the chromatic characteristics substantially unchanged (except for possible shift phenomena, which will be explained in more detail below) ,

ii) and to reciprocally vary the contribution of the individual "channel" to the radiation obtained by the mixing action, so as to vary the chromatic characteristics of said radiation - keeping the comprehensive light intensity substantially unchanged. In this respect, it can be seen that the individual "channels" (i.e. the individual generators which provide respective radiations intended to be mixed with one another) may be activated at different power levels, insofar as there can be channels which make a particularly considerable contribution to the mixed radiation and are activated at a power level (duty cycle) which is higher than the power level at which other channels which make a less considerable contribution to the mixing action are activated. The text which follows refers principally to the second function outlined above, i.e. that which makes it possible to vary the chromatic characteristics of the light radiation resulting from the mixing, keeping the comprehensive light intensity substantially unchanged.

In the text which follows, it is therefore assumed for simplicity that what is involved is a plurality of light generators or "channels" CHI, CH2, CH3, ... supplied with respective supply powers PI, P2, P3, ... over respective activation intervals Tl, T2, T3, ... included in an activation period T, which is assumed to have a constant duration value. According to known criteria, the value of the period T is chosen such as to utilize the effect known as "persistence of the image on the retina" such that, even if the individual light generators are activated periodically (i.e. once for each period T) , the radiation emitted thereby and the mixed radiation generated by the lamp L as a whole are perceived as a continuous light by a human observer, without "flickering" phenomena. Typical values of the interval T can be chosen in the range between 1 and 10 ms, corresponding to repetition frequencies of the pulses emitted by the individual generators CHI, CH2, CH3 of between 100 and 1000 Hz.

A situation of this type is shown schematically in figure 1 in the accompanying drawings, which refers for simplicity to the case of one light source comprising only two "channels" CHI and CH2, i.e. only two generators, for example two LEDs .

In the situation shown in figure 1, the first "channel" CHI is activated over an interval Tl, whereas the second generator or second "channel" is activated in an interval of duration T2 (T2 < Tl) which partially overlaps the interval Tl. In figure 1, AP and PP respectively indicate the average power level and the peak power level consumed by the system containing the two channels CHI and CH2. In the case under consideration, in which T2 is much shorter than Tl, the average power AP is slightly less than the power PI consumed by the first channel CHI, whereas the peak power PP (supplied over the interval of duration T2) corresponds to the sum PI + P2 of the power consumptions of the first channel CHI and of the second channel CH2 put together. Therefore, the electrical power supply unit of such a light source has to be designed such that it is able to supply a power corresponding to the peak power PP. This is the case even if in reality, for most of the period T, the power supplied is that which corresponds to the power consumed by the first channel CHI alone.

Figure 2 shows a situation which is likely to be established in the exemplary case of three channels CHI, CH2, CH3 activated respectively over an interval Tl, over an interval T2 and over an interval T3, all three of which have a shorter duration than the period T fixed by the PWM modulation diagram. In this case, the activation takes place at respective power levels PI, P2, P3 which are not necessarily the same, i.e. with each channel CHI, CH2, CH3 being able to consume a respective percentage (%PPS) of the peak supply power.

By way of example, it is conceivable that the power PI consumed by the channel CHI corresponds to a percentage equal to 37.5% (%PPS=37.5%) of the overall peak power; this is also applicable to the channel CH2, whereas the channel CH3 consumes only 3.25% of the peak power (%PPS=3.25%) .

With respect to the right-hand side of figure 2, assuming that all three intervals Tl, T2 and T3 (it is assumed here that Tl > T3 > T2, but this is just an example) start at the same time as the period T, it can be seen that the overall consumption of the three channels CHI, CH2 and CH3 reaches the peak value P1+P2+P3 only for a (sub) period equal in length to T2, and then at first drops to the value P1+P3 for a subsequent interval of duration T3-T2, and then again to the value PI for a subsequent interval of duration T1-T3. For the rest of the period T, of duration T-Tl, the consumption is zero.

In order to sustain the operating modes referred to in figure 2, the power supply unit of the lamp has to be dimensioned to supply a peak power P1+P2+P3. However, this peak power is supplied only for a small fraction (the interval T2) of the period T, whereas for a good part of this period the power consumption is less (even much less) if not even zero.

As a corollary to that mentioned above, it can also be noted that the power supply units for light generators, such as LED light generators, mostly use commutated (switched-mode) power supply devices, which provide maximum efficiency when they operate at the maximum proposed power. This condition is far from being achieved either in the situation shown in figure 1 or in the situation shown in figure 2. It should also be stated that the commutation (on-off) of high-intensity currents inevitably causes electromagnetic interference (EMI).

Object and summary of the invention

The object of the invention is to provide a solution for driving light sources of the type specified above which is able to overcome the disadvantages outlined above.

According to the present invention, this object is achieved by means of a method having the characteristics specifically claimed in the claims which follow.

The invention also relates to a corresponding power supply unit as well as to a corresponding light source.

The claims form an integral part of the technical teaching given here in relation to the invention.

Brief description of the accompanying drawings

The invention will now be described, purely by way of non- restrictive example, with reference to the accompanying drawings, in which:

figures 1 and 2 have already been described above, figures 3 to 5 show the operating principle of various embodiments,

figure 6 is a diagram showing various embodiments, and figure 7 is a flow chart relating to various embodiments . Detailed description of embodiments

The following description sets out various specific details aimed at giving an in-depth understanding of the embodiments. The embodiments may be realized without one or more of the specific details, or with other methods, components, materials, etc. In other cases, known structures, materials or operations are not shown or described in detail so as not to obscure various aspects of the embodiments.

The reference to "an embodiment" in this description means that a particular configuration, structure or feature described in relation to the embodiment is included in at least one embodiment. Hence, phrases such as "in an embodiment", which may occur at various points in this description, do not necessarily refer to the same embodiment. Moreover, particular shapes, structures or features can be combined in a suitable way in one or more embodiments. The references used here are purely for convenience and do not therefore define the scope of protection or range of the embodiments .

Figure 3 schematically shows, with reference to the simple case of only two channels CHI and CH2, for better understanding, the general principle on which various embodiments are based. Figures 4 and 5 refer to cases which involve three channels CHI, CH2 and CH3. The examples under consideration, like those already considered in relation to figures 1 and 2, therefore refer to the presence of a plurality of light generators, i.e. "channels" CHI, CH2, CH3, which are assumed to emit respective chromatically different light radiations that can be mixed with one another to produce a mixed light radiation by selectively activating said generators with respective supply powers PI, P2, P3, over respective activation intervals Tl, T2, T3, ... included in an activation period T. The block diagram shown in figure 6 shows a plurality of LED light generators 10, 20, 30, ... which can emit respective radiations with chromatically different characteristics and correspond exactly to the channels CHI, CH2, CH3, .... The LED generators 10, 20, 30, ... are supplied with power by a power supply unit 1000, which can be connected to the electric mains N.

According to criteria known per se, the unit 1000 is provided with two regulation functions 1002 and 1004 which respectively supervise :

- the regulation of the comprehensive intensity of the light radiation emitted by the source L and resulting from the mixing of the individual radiations produced by the generators 10, 20, 30, ... (which may be present in any number in the plurality) ,

- the control of the mixing function obtained by varying the intensity of the radiation emitted by the individual generator 10, 20, 30, ... so as to selectively vary the contribution made thereby to the radiation resulting from the mixing, hence varying the chromatic characteristics (Cx, Cy, CCT) of said radiation.

The recourse to light generators such as LEDs, even if preferred, is by no means to be considered imperative, insofar as the solution described here lends itself to general use, in connection with any light generator having a light emission intensity which is a function of the (average) supply power, for example a function of the average value of the electrical current flowing through it. Accordingly, even if the rest of the present description makes reference to a method for regulating the radiation intensity of the individual generators effected with PWM modality, it is to be understood that the invention also encompasses other regulation options, such as, for example, regulation by means of alternating current, pulse density modulation and statistical dimming methods.

In turn, whereas the rest of the description, provided purely by way of example, makes reference at least implicitly to control effected on the supply current of the channels CHI, CH2, CH3, ... (i.e. of the LED generators 10, 20, 30, ...) , various embodiments may provide for a regulation method conducted, in addition or as an alternative to a current control function, by means of a function for controlling the supply voltage of the generators 10, 20, 30,

The description which follows refers principally to the method for regulating the light intensity of the channels CHI, CH2, CH3, which is directed at regulating the chromatic characteristics of the mixed radiation emitted by the lamp L (control function indicated by 1004 in figure 6) .

Figure 3 refers to a source L comprising two generators, i.e. two "channels" CHI and CH2, which are assumed to be intended for activation over two respective intervals equal to Tl and to T2 with respective power levels PI, P2, which, for simplicity, are assumed to be substantially identical.

The solution shown schematically in figure 3 is based on the recognition that, when it is necessary to supply power to the two channels CHI and CH2 over two respective intervals Tl and T2 (with T1+T2<T - where T is the "general" power supply period fixed in accordance with the criteria already indicated above) , it is advantageous to avoid simultaneous activation of the two channels CHI and CH2.

By way of example, with reference to the part on the far left of figure 3, it can be seen that, when supplying power to the channel CH2 over an interval T2 which starts at the same time as the activation interval Tl of the channel CHI, it is necessary to provide over the initial interval of duration T2 an overall electrical power P1+P2 equal to the power consumed by the two channels CHI, CH2 put together. In this case, it is possible to defer the activation of the channel CH2 to a time when the interval Tl has elapsed, as indicated schematically by the arrow which can be seen in figure 3. In this case, while obtaining the same results in terms of the power supply to the two channels CHI and CH2, there is the considerable advantage that the power supply unit 1000 may be dimensioned without having to take into account the peak power given by the sum of the power PI consumed by the channel CHI and by the power P2 consumed by the channel CH2.

In the case (given for simplicity) where P1=P2, the set peak power of the source 1000 no longer has to be equal to P1+P2, but in practice has to be equal to P1=P2, which in essence means half.

The solution shown in figure 3 is therefore based on the identification of at least one pair of generators, CHI and CH2, for which the sum of the respective activation intervals Tl, T2 is less than the activation period T (i.e. with T1+T2 < T) , in order to temporally offset the activation of the generators CHI, CH2, avoiding simultaneous activation.

In practice, the principle shown in figure 3 corresponds to an "alignment" function for driving the various channels CHI, CH2 (which can be implemented, according to criteria known per se, for example, in the programming of the unit 1000) such as to "minimize" the number of channels active at the same time: with reference to the simple example shown in figure 3, instead of there being two channels active at the same time, there is at most only one active channel, first the channel CHI and then the channel CH2.

By minimizing the number of channels which are active at the same time, it is no longer necessary for the unit 1000 to be dimensioned so as to be able to take into account the requirements for supplying power to all the channels of the source L (as would occur in the case of an operation corresponding to the solutions shown in figures 1 and 2) . By virtue of the abovementioned optimization function, it is possible to dimension the power supply unit 1000 in such a manner that it is able to supply power - at most - to all the channels for which simultaneous activation is anticipated.

This makes it possible to achieve a number of advantages:

firstly, the power supply unit 1000 can be designed on the basis of a lower peak power, with advantages both in terms of dimensions and in terms of cost, and

secondly, the average power supplied by the power supply unit 1000 will be next to the peak power, with significant advantages in terms of efficiency. The diagram shown in figure 3 is substantially based on the idea of utilizing the "empty" space left free, during period T, by the interval Tl for supplying power to the first channel CHI in order to supply power to the channel CH2. Such a condition, i.e. the availability of an "empty" interval, is not always present, particularly in the case of sources L comprising more than two channels. By way of example, figure 4 refers to a source with three channels CHI, CH2 and CH3 intended to be supplied with power over respective intervals equal to Tl, T2 and T3 in conditions in which the interval Tl takes up virtually all of the period T (CHI supplied with power with 100% duty cycle) .

However, in these conditions, too, various embodiments may allow for an optimization function, temporally shifting (i.e. bringing forward or delaying) the occurrence of the intervals for supplying power to some of the generators.

By way of example, the diagram shown in figure 4 refers to a situation where the interval T2 for supplying power to the channel CH2 and the interval T3 for supplying power to the channel CH3 satisfy the condition T2+T3 < T.

In this case, the activation criterion shown on the left-hand side, indicated by a), of figure 4, for which the two channels CH2 and CH3 are supplied with power over an interval of duration T3, can be replaced by the activation criterion shown on the right-hand side, indicated by b) , of said figure 4, which provides for "bringing forward" the interval T2 for supplying power to the channel CH2 such that it temporally precedes the interval T3 for supplying power to the channel CH3.

The top part of figure 4 compares, respectively for the solution shown in part a) and for the solution shown in part b) , the effect on the power consumption of the channels CHI, CH2 and CH3 and therefore on the dimensioning of the power supply unit 1000. With regard to the solution shown in part b) of figure 4, it can be seen that the peak power PP is virtually returned to the value of the average power AP, again achieving the advantages outlined above.

If there are more than two channels, too, and also if there is no "empty" space present during the period T, it is possible to minimize the number of channels activated simultaneously, for example, as shown in part b) of figure 4, by bringing forward the activation interval T2 of the channel CH2 with respect to the activation interval T3 of the channel CH3 or (considering the same function in a different way) by delaying the activation of the channel CH3 until the channel CH2 is deactivated .

It is possible to provide for dead times between the deactivation of one channel and the activation of the other, thereby ensuring that the activation of two channels does not overlap (which could produce an undesirable peak consumption) .

The example shown in figure 4 therefore corresponds to the case in which the plurality of light generators of the lamp L identify a first CHI, a second CH2 and a third generator or channel (respectively CHI, CH2 and CH3) in conditions for which the sum of the respective activation intervals - T2 and T3 - of the second generator CH2 and of the third generator CH3 is less than the activation period, and in which:

- the first generator CHI and the second generator CH2 are simultaneously activated over the respective activation interval T2 of the second generator CH2, and

the first generator CHI and the third generator CH3 are simultaneously activated over the respective activation interval T3 of the third generator CH3.

Here, too, the optimization method is based on the identification of at least one pair of channels or generators (CH2, CH3 in the case shown in figure 4) for which the sum of the respective activation intervals T2 and T3 is less than the activation period T. This is done in order to temporally offset the activation of the generators CH2 and CH3 in said pair, avoiding simultaneous activation.

Restrictions such as that shown above (T2+T3 ≤ T) are not usually particularly pressing insofar as, in Color Management Systems (CMS) , it is found that said condition is satisfied almost always for at least two colors in the system.

By way of example, in a CMS system with three or more colors or channels, there is usually a "white" color which provides the majority of the light flux (take, for immediate reference, the channel CHI shown in figure 4, activated with a duty cycle substantially equal to 100%, i.e. for an interval Tl virtually identical to the period T) , whereas the other channels CH2 and CH3 may have the function of "harmonizing" or maintaining a certain color dot (Cx, Cy in the CIE 1981 diagram) or the corresponding CCT value.

For example, the other colors mentioned above may realize a function of adding blue and red to a fundamentally white system (where the red gives a "warmer" CCT, while the blue gives a "colder" CCT) , so both of these components (red and blue) are never activated over intervals T2 and T3 longer than 50% of T.

As seen in this perspective, a strategy of the type referred to in figure 4 ignores, as it were, the power requirements of the "white" channel CHI and applies the optimization strategy to the other channels CH2, CH3 so as to reduce the requirements in terms of peak power. The "alignment" function of the channels CH2, CH3 can thus be carried out independently of the channel CHI .

The example shown in figure 4 can be considered to be an illustrative embodiment in which one of the channels (for example the "white" channel which, in the text above, is assumed to correspond to the channel CHI) is extensively dominant, being activated with a much higher duty cycle than the other channels (for example with a duty cycle close to if not equal to 100%, so the interval Tl is roughly equal to the period P) , and the source or lamp (luminaire) L has to be subjected to dimming. For this purpose, it is possible to effect a driving strategy for the channels, in which the channel with the highest duty cycle (for example CHI) is driven at 100% (or at the maximum value according to the optimization algorithms of the driving) , additionally ensuring that the current of said channel can be reduced so as to make the light flux of said channel correspond to the desired level of dimming (this can take place according to known criteria, for example by means of a communication function with the power supply unit 1000 or an intervention on the control loop of said unit) .

However, this method for regulating the light flux may result in a color shift in the radiation emitted by the channel CHI and therefore in the mixed radiation emitted by the lamp L as a whole. In this case, it is possible to intervene with a color management algorithm or a Color Management System (CMS) algorithm and an LED color algorithm, in order to ensure that at least one of the other channels (CH2 and/or CH3 in the example under consideration here) can be accordingly regulated (by varying the interval T2 and/or the interval T3, always referring to the example under consideration here) , bearing in mind the possibility to correct the color shift described above . In this case, the intensity of the radiation emitted by the first generator (CHI in the example under consideration) is varied, and said variation establishes a color shift in said radiation. Then, the respective activation interval (T2 and/or T3) of at least one of the other radiation generators CH2 and CH3 included in the source L is selectively varied so as to compensate for the color shift.

It will be appreciated that the requirements in terms of power consumed by the various channels are not necessarily the same (as already considered to a certain extent also with reference to figure 4 ) .

These requirements may be known when designing the light source L, where it is known how many generators 10, 20, 30, (therefore channels CHI, CH2, CH3, ...) will be used and where the relative requirements in terms of current and voltage consumption are known. In sophisticated light systems, it is also possible to measure some or all of these parameters in real-time, or to provide at least an estimation so that the requirements in terms of supply power are known. It is also conceivable for said parameters to be neither design parameters nor calculated/estimated parameters, but instead regulation parameters supplied to the light source L by an external system (for example by means of a digital communication channel, etc . ) .

Figure 5 shows, according to the same modes of representation, a situation which is entirely similar to that shown in figure 4, again referring to three channels CHI, CH2 and CH3 intended to be activated, during period T, over respective activation intervals Tl, T2 and T3 with respective power levels indicated by PI, P2 and P3. This is a situation where the power level P2 of the generator/channel CH2 is at least approximately equal to the sum P1+P3 of the power levels PI and P3 of the two other generators/channels CHI and CH3, i.e. P2 is roughly equal to P1+P3. Once again, the part on the left-hand side of figure 5, indicated by a) , refers to a non-optimized activation strategy, in which all three activation intervals Tl, T2 and T3 commence at the same time as the period T. In this case, the set peak power of the unit 1000 has to be the sum of the three powers PI, P2 and P3, insofar as all three channels CHI, CH2 and CH3 are simultaneously active over an interval of length T2 at the start of the period T.

On the other hand, the part on the right-hand side, indicated by b) , shows the result of the optimization operation, which initially provides for the activation of the channel CH2 over an interval T2 and then the simultaneous activation of channels CHI and CH3 only after the channel CH2 has been deactivated. In the exemplary embodiments shown in figure 5, there are therefore at least three generators/channels CHI, CH2 and CH3 present in a situation in which the respective activation power P2 of one of the generators (i.e. CH2) is at least approximately equal to the sum of the respective activation powers PI, P3 of the other two generators, i.e. CHI and CH3. In this case, the first generator CH2 is then activated over the respective activation interval T2, keeping said other two generators CHI and CH3 deactivated. Then, these other two generators CHI, CH3 are activated simultaneously once the first generator CH2 has been deactivated .

This example therefore also provides for the identification of at least one pair of generators (CH2 and CHI, or CH2 and CH3) for which the sum of the respective activation intervals (T2 and Tl or T2 and T3) is less than the activation period T, and for the temporal offsetting of the activation of the generators (CH2 and CHI or CH2 and CH3) in said at least one pair, avoiding simultaneous activation (CH2 is not activated together with CHI or together with CH3) . Once again, the optimization results in conditions in which the average power supplied is very close to, if not even coincident with, the peak power. Owing to the optimization process, the power supply unit 1000 can be dimensioned so as to supply a peak power virtually equal to P2=P1+P3, much less (in the case of the present example roughly equal to half) than the peak power required in non-optimized conditions.

It can be seen from the part on the right-hand side of figure 5 that the accomplished optimization and the consequent "packing" of the activation functions could also have the effect that, for a certain residual part of the period, there will be no power consumption, insofar as all the channels are deactivated. This condition may be advantageous insofar as the supply source 100 may be placed in standby mode with a reduced power supply, virtually zero.

The power supply unit 1000 may be managed, for example, by an element, such as a microcontroller, which is able to verify some reference data, i.e.:

the fact that the %PPS values of all the generators/channels, if added together, never exceed 100%, while verifying that the sum value is as close as possible to 100%, and

the fact that the activation/deactivation relationship (duty cycle on/off) of the various individual channels is maintained, for example so as to ensure that the desired CCT is maintained.

The management element mentioned above, aware of the requirements of the various channels, may be configured (for example in programming effected according to criteria known per se) such that it calculates the %PPS value for the various channels in order to identify, among the plurality of light generators CHI, CH2, CH3, ... of the source L, at least two subsets of generators to be activated at a given time, each of said at least two subsets comprising one generator having a respective supply power or a number of generators having a sum of the respective supply powers. The management element mentioned above may then intervene so as to activate, at said given time, the generator or the generators in the subset for which the respective supply power or the sum of the respective supply powers is closer to the maximum supply power which can be supplied by the unit 1000.

Said maximum suppliable power may be defined, for example during the design of the unit 1000, as a function of the peak value of the supply power which can be supplied to the various generators/channels CHI, CH2, CH3 during the activation period T. The reference, when introducing the optimization strategy explained in figure 5, to at least two subsets of generators to be activated - at a given time - takes into account the fact that various channels activated simultaneously can be switched off or deactivated at different times so as to maintain the respective on/off relationships, and therefore, at a certain time, the channels to be considered for the optimization calculations are only those for which the required on time has not yet been reached.

The exemplary strategy on the right-hand side of figure 5 corresponds to the following steps shown in the flow chart of figure 7. After a start step START, the respective activation power values are determined (for example in terms of %PPS) in a step 100 - for all the channels to be activated or still to be activated (in the present case all three channels CHI, CH2 and CH3) .

Next, in a step 102, the subsets of one or more channels are found in a condition where the sum of the relative %PPS does not exceed 100%. These subsets may be only one subset or a plurality of subsets. In the latter case, that subset for which the sum of the %PPS is closest to 100% is selected in a step 104.

In a step 106, the only subset or the preselected subset (P2 in the example shown in figure 5) is used to activate the respective channel or channels (the channel CH2 in the present case) .

The step 108 represents the verification of the fact that the channel or the activated subset of channels may be switched off once it has accomplished its activation function. If the outcome of step 108 is negative, the activation is continued .

If, on the other hand, the outcome of step 108 is positive, indicating the fact that the channel or the activated subset of channels may be switched off, the sequence described above is repeated for the remaining channels to be activated. The step 110 aims precisely at verifying that said channels are present (positive outcome of step 110) until all the channels to be activated have been activated (negative outcome of step 110), and the commutation cycle is then repeated. In step 112, which is reached if the outcome of step 110 is negative, the need to repeat the commutation cycle is verified (positive outcome of step 112, returning to before step 110), with evolution toward the stopping of the system (STOP) if the outcome is negative.

The reference in this description (and, if appropriate, in the claims which follow) to "simultaneously" activated channels naturally does not exclude the fact that, in various embodiments of the "simultaneous" activation, it is possible to provide for small separation intervals between one activation and the other in order to avoid voltage and/or current fluctuations, which may possibly generate undesirable electromagnetic interference phenomena.

Various embodiments make a positive contribution to the operational efficiency of the power supply unit 1000, insofar as some of the losses (for example the resistive losses and the rectification losses) are lower in the power supply source 1000 at lower currents.

The alignment function described here can be effected by the unit 1000 (or in general by the light engine or lamp L as a whole) , since these elements can communicate with one another to hypothesize an operating situation by avoiding overload situations and also carrying out an action for predicting the power supply requirement so as to adapt to temporally variable load requirements (for example in the subsequent duty cycles) .

Without departing from the principle of the invention, the details of construction and the embodiments may vary, even significantly, from those illustrated purely by way of non- restrictive example without thereby departing from the scope of the invention as defined in the appended claims.