WO1994013045A1

WO1994013045A1 - Method and apparatus for operating a plurality of laser diodes

Info

Publication number: WO1994013045A1
Application number: PCT/GB1993/002450
Authority: WO
Inventors: Anthony Raven; Robin Michael Lee
Original assignee: Diomed Limited
Priority date: 1992-11-30
Filing date: 1993-11-29
Publication date: 1994-06-09
Also published as: AU5570694A; GB9225050D0

Abstract

The laser diodes (2) of a light source (1), which may be coupled to an optical fibre (4) are operated by a microprocessor (10) and driver (11) such that the number of diodes (2) activated depends on the desired power of the output beam (6), as input via input means (12). Preferably, the number of diodes (2) activated is chosen such that each activated diode (2) operates at its optimal demand (about 75 %). The diodes (2) may be selected on the basis of which have been used the least and/or to produce an output beam (6) of desired configuration, e.g. shape or intensity profile. The demands on the diodes (2) may take account of losses in the system, and a check may be made to ensure that the desired output power will be produced by the selected diodes (2) at the selected demands. Damaged diodes (2) are either not used, or only used if the combination of all the undamaged diodes (2) is insufficient.

Description

Method and Apparatus for Operating a Plurality of

Laser Diodes.

The present invention relates to a method of operating a plurality of laser diodes to produce a combined output of a desired power, and also relates to a light source implementing the method. WO-A-92/02844 discloses a system in which the output beams from a plurality of laser diodes are coupled into an optical fibre to produce a light source capable of operating at high powers.

The applicants have realised that problems can arise in operating such a source efficiently at powers less than the maximum, and especially in the lower power ranges. It is therefore an object of the present invention to provide a method of operating such a source, and sources comprising multiple laser diodes in general, in an efficient manner, particularly at lower power ranges.

Thus, viewed from one aspect, the present invention provides a method of operating a plurality of laser diodes to produce a desired combined output power therefrom, wherein selected ones of said laser diodes are activated by electronic control means, the number of said selected diodes depending on the desired combined output power.

This method contrasts to the known method in which all of the laser diodes are always activated no matter what the desired output power may be, and in which the output power of each laser diode is set relatively low when the desired output power is low and relatively high when the desired output power is high. Instead of this, the method of the present invention allows some of the laser diodes to remain unactivated so that the desired output power is produced in appropriate circumstances by running less than the full number of laser diodes available, at higher average output powers than would be necessary if all of the laser diodes were being activated. This method actually increases the complexity of the control means, as the laser diodes do not all receive the same amount of current, and means are needed to allocate which laser diodes are to be used and at what powers they are to be operated. These inconveniences are, however, more than compensated for by the increase in efficiency which this method can provide.

The increased efficiency may be explained with reference to Figure 1, which shows a graph of input current versus output laser power for a typical laser diode. As can be seen, the output laser power is generally proportional to input current, but, importantly, the laser diode needs to be supplied with a threshold current (typically -0.5A or -25% of maximum current drawable) before a laser output is produced. This means that for each laser diode activated there is effectively a current loss of about 25% of the laser diode's maximum current. The present invention reduces this loss by allowing fewer laser diodes to be activated at higher average powers, where appropriate. As an example of the greater efficiency available, if a laser source comprises sixteen laser diodes, and if all sixteen diodes are used to produce an output power of only about 10% of the maximum possible output power, then each laser diode would need to be supplied with 32.5% of its maximum current, i.e. a 25% threshold current and a further 7.5% current to produce the required 10% of maximum output power (one tenth of the 75% of maximum current which produces laser output) . The same output power can, however, in accordance with the present invention, be produced, for example, by operating only two laser diodes, as opposed to all sixteen, at 80% of their maximum output power. Each diode would then need to be supplied with 85% of its maximum current, but the combined current required by the two diodes would only be 10.6% of the combined current required when using all sixteen of the laser diodes. Thus, by using only two laser diodes at higher powers, the number of threshold current losses are reduced.

A further advantage of the present invention is that less heat is produced because of the lower overall supply current. This enables smaller heat sinks to be used to cool the laser diodes, and so enables smaller, more compact sources to be produced. Also, there is less thermal stress, which increases the reliability and service life of the source, and the laser diodes can be controlled more accurately due to the lower heat levels and the higher operating powers. Further, in the medical and surgical fields, the circumstances under which a source such as in WO-A-92/02844 is used at low powers, are usually those in which the source is operated for a relatively long period of time. In these situations, therefore, the problems of low power operation discussed above are exacerbated and the present invention is especially effective.

As few laser diodes as possible could be used to produce the desired output power, which would necessitate running the selected laser diodes at or close to 100% demand, that is at 100% of their maximum output power. Typically, however, a laser diode will have an optimum output power, or range of output powers, corresponding to an optimum demand or demand range, at or over which it operates with the greatest efficiency, and/or, for example, has its greatest life expectancy and/or controllability. This need not necessarily be at 100% demand, and preferably, therefore, the number of laser diodes which are selected to be activated is calculated so that as many as possible of those selected may operate at or in an optimal power or power range. This power range need not necessarily be the same for each laser diode, but may typically be between about 70-90% demand, and more preferably between about 75-85% demand. There are a number of ways in which the number of laser diodes to be used, and the power at which each selected diode is operated, could be determined. For example, each diode could be set to operate at a specific demand, e.g. 75%, and the output powers of successively selected diodes could be added until the next addition would make the combined output power greater than that desired. The demand on that next laser diode could then be reduced below the set percentage so that exactly the desired combined output power is provided. This however could mean that the last selected diode might be run at an inefficient output power. As an alternative, the optimum number of laser diodes for a particular desired output power could be looked up in a map having predetermined diode numbers for each output power, the number of diodes being set in dependence on the above and, optionally, further considerations as discussed below.

In the presently preferred mode of operation, laser diodes are selected in sequence and their yields, i.e. output powers at maximum demand, added together until a set percentage of the sum of the yields is equal to or exceeds the desired output power. Each selected laser diode is then run at a power level which is calculated on the basis of the ratio between the maximum power available from the selected diodes and the desired output power, so that their outputs combine to produce the desired output power. This provides a quick and simple method of calculating the number of laser diodes to use, and the set percentage may be such that in practice the laser diodes end up being run at about an optimum output power or within an optimum power range. The laser diodes could be selected in a set order so that when eight laser diodes are needed the first eight in the order are always used, and when ten laser diodes are needed the first ten in the order are always used, and so on. This, however, would mean than the lower order laser diodes would be used much more often than the higher order ones, and so the lower order diodes would tend to fail more quickly than the others, thereby reducing the service life of the source as a whole. Preferably, therefore, the diodes are selected in such a way that in the long term each is on average used to substantially the same extent, or as near to the same extent as possible given any other considerations which may affect the selection, such as those mentioned below. This more equal division of use amongst all of the laser diodes allows each to be used less and increases the service-life of the source.

One method of equalising the usage is to select the diodes following a predetermined rota. However, other considerations, such as those mentioned below, may take precedence over usage, and may effect the order of the laser diode selection. In such a case, a simple rota system may not be applicable, and in a preferred embodiment, the selection of diodes favours those which have been used less, subject to any other considerations which may also effect selection as discussed below. In one embodiment, the number of times each laser diode is used is recorded, and those which have been used less are in general selected in preference to those which have been used more, but this selection may also be subject to other considerations.

Counting the number of times each laser diode is used, or using a rota selection system, may only approximately equalise the amount of use .of each diode, as the duration for which a laser diode is used may be different on each occasion. It can be expected that it will not be too inaccurate an approximation to assume that each use is of equal duration, since the differences may even out with time. In a preferred method, however, instead of, or as well as, counting the number of times each laser diode is used, the length of time each laser diode has been used for is recorded and used to select which of the diodes are to be activated. It would also be possible to record the laser diodes' operating powers during each use, and to use this value as a weighting factor on the number of uses or length of use counts. This may not be an important consideration, however, where, for example, the number of laser diodes chosen is selected to ensure that they operate approximately within a set demand range.

The term "usage" should be taken to cover the number of times a laser diode is activated, and/or the amount of time a laser diode has been activated, with or without any consideration being given to the power levels at which this use has been made.

In a preferred embodiment, the outputs from the laser diodes are combined and focused into the upstream end of an optical fibre in use. In this case, a selection consideration which may take precedence over the amount of usage of each diode is the shape or cross- sectional intensity profile of the combined output beam emerging from the fibre, as differently positioned laser diodes within an array affect the shape and/or profile of the combined output beam, and in many circumstances, such as in laser surgery where a beam is focussed onto tissue, this is of practical importance. As an example, it may be desirable to provide a consistent cone of illumination output from the fibre, or an output beam having as small as possible a cone angle.

The laser diodes may therefore be selected to produce a desired output beam configuration, e.g. shape and/or cross-sectional intensity profile, and this may advantageously be done by dividing the laser diodes into a number of groups, such that the use of a laser diode from any one group has the same general affect on the combined output beam as the others in that group. The number of laser diodes selected from each group may then be selected to produce a desired configuration of output beam, and ^" which of the laser diodes in each group are selected to make up that number may be determined in accordance with the usage criteria described above.

The laser diodes may be arranged into an inner group which are nearer the centre axis of the fibre, and an outer group which are further from the axis. Then, in order to make the output beam cone angle as small as possible, laser diodes from the inner group may be selected in preference to laser diodes from the outer group, whereas, in order to provide a more divergent output beam of approximately constant intensity across its whole cross-section, a substantially equal number of laser diodes may be selected from each group. For example, a laser diode may be selected alternately from the inner group and then the outer group. Alternatively, it may in certain circumstances be desirable to select a greater number of diodes from the outer group, so that an annular image can be produced at the fibre output.

The demand on each laser diode may also be modified somewhat to produce a combined output beam of a desired configuration, e.g. shape and intensity profile. For example, the laser diodes in one group may be run at higher outputs than those in another group.

When calculating the laser diode power levels necessary for producing a desired output power, losses dependent on the efficiencies of the source optics may need to be taken into consideration. The losses can normally be divided into those related to each laser beam's path, which are the same for each.specific type of source, and those which are specific to each source and relate to variations in each individual source's assembly. One way of compensating for these losses is to measure, during manufacture, the maximum output power of each laser diode at the source's output, for example at a fibre connection port of the source. These amounts, or efficiencies based on them, may then be stored in a control means, and used to adjust the demand on each diode accordingly.

Where an optical fibre is used, it may be preferable to specify the desired output power of the fibre tip, rather than at, for example, the fibre connection port of the source. The fibre may therefore be calibrated by firing a pulse of laser light of a set power (measured at the fibre port) into the fibre, and measuring the power produced at the fibre tip so as to obtain the fibre's efficiency. This value may then be used to uprate the desired output power inputted by the operator.

For safety, a check may be made to verify that a correct number of laser diodes at correct output levels have been selected. This may be done by multiplying the calculated demand by the yield for each selected laser diode to produce each laser diode's theoretical operating output power, then adding these output powers, and comparing the addition to the desired total output power to ensure that it falls within set tolerance values. Checks may be periodically made to determine if any of the laser diodes are damaged or have failed, in which case the laser diode may need to be eliminated from availability. When a laser diode is damaged, but still operable, it may be reserved for use only at high or maximum powers when the use of the other diodes is not enough. This then reduces the usage on the damaged diodes and increases the service life of the source. As in WO-A-92/02844, it is possible.to combine beams from a pair of laser diodes into a single beam using, for example, a polarising beam combiner and then to combine a number of these beams into an optical fibre by a lens. In situations such as this, each diode pair may be treated as a single diode, in which case both are operated in the same manner and may receive the same supply current, or they may each be operated separately in accordance with the above considerations, so that, for example, at 50% of maximum power, only one diode from each pair is operated.

The invention also extends to a light source which uses a number of laser diodes efficiently to produce a desired combined output power, and from a further aspect, provides a light source comprising a plurality of laser diodes, means for inputting a desired source output power requirement, and means for determining the number of diodes to be activated in dependence on the desired output power. Preferably, the source includes means for recording the amount of usage of each diode, such as the number of times each diode has been used, the total length of time each diode has been used, and the power at which each diode has been used, during each use, the determining means selecting the diodes to be activated on the basis of these records.

Embodiment of the present invention will now be described, by way of example only, with reference to the accompanying drawings, in which:

Figure 1 shows a graph of light output power versus input current for a typical laser diode;

Figure 2 is a schematic view of apparatus in accordance with a first embodiment of the present invention;

Figure 3 shows a schematic plan view of a 4 x 4 laser diode array according to the first embodiment of the invention;

Figure 4 is a flow diagram for a laser diode allocation algorithm in accordance with the present invention; Figure 5 is a schematic diagram showing the input of an off-axis laser diode beam into an optical fibre, and the beam shape on emergence at the output end of the f ibre ; and

Figure 6 shows schematically the light intensity distribution across a combined output beam, and inner and outer laser diode components thereof. As shown in Figure 2, a laser light source unit 1 comprises a plurality of laser diodes 2, the output beams 3 of which are coupled into an optical fibre 4 by a focusing lens 5 to produce a fibre output beam 6 of relatively high power. The fibre 4 is detachably connected to the source at a fibre port 7, which may be able to accommodate a number of different fibre sizes and types.

Each beam 3 is shaped by an anamorphic beam shaping means 8, such as a prism pair, to ensure efficient coupling of the beam into the fibre 4, and is deflected towards the lens 5 by a prism 9. The prisms 9 are arranged in a stepped configuration so that the beams 3 from a row of laser diodes 2 impinge in a line one above the other onto the lens 5. A 4 x 4 array of laser diodes 2 is used, as shown in Figure 3, the laser diodes being labelled LD0-LD15 for convenience. Details of apparatus for use in such a system may be found in WO-A-92/02844.

A microprocessor CPU 10 controls the current to the laser diodes 2, via driver means 11, in accordance with the desired output power entered by an operator through input means 12, such as a keyboard, and also in accordance with instructions and data held in an associated memory means 13. This data may include a look-up table showing, for example, for each laser diode, whether or not that laser diode has been selected for use, its yield, i.e. its output power at 100% demand, and whether or not is it available for use, e.g. whether or not it is damaged. A display 14 is provided so that the CPU 10 may display error signals and other messages, such as, for example, to confirm the desired output power input by the operator. Once a desired output power Pop has been entered by an operator, the CPU 10 executes the routine shown in the flow diagram of Figure 4. This routine selects the laser diodes 2 which are to be activated to produce the desired output power.

Generally, the routine runs as follows: The loop S4-S14 is run through a number of times until enough laser diodes have been selected such that their combined output power is 125% that of the desired output power. During each loop, the least used laser diode is selected for use from the as yet unselected laser diodes. The routine then goes on to calculate the demand of each laser diode, i.e. the percentage of maximum power at which the diode needs to operate. This is determined by dividing the desired output power by the total output power available from the selected diodes. A check is then made in steps S18-S26 to ensure that the operation of the selected diodes at their determined demand will indeed produce a total output power which is equal to the desired output power within a given tolerance.

Error signals are issued at steps S17 and S25 should the power demand be unable to be achieved with the selected diodes operating at full demand, or should the selected diodes be determined to produce an output power outside the acceptable tolerance. Otherwise, the selection is considered successful, and the apparatus is enabled to allow the operator to activate the source.

Considering the routine of Figure 4 more specifically, the routine is entered at step SI, and, at step S2, the total output power available Pav is set to 0, as are all of the laser diode selection flags LDsel(i) , in step S3. These flags may be found in a look-up table in the memory means 13 , and indicate whether or not a particular laser diode i has yet been selected (LDsel(i)=l or 0, respectively) . This zeroing corresponds to the situation of no laser diode yet selected and no output power available. At step S4 , a minimum usage variable Umin is set to an initial high value, and the laser diode variable i is set to 0, corresponding to laser diode LDO of Fig. 3.

At step S5, a check is made to see if laser diode i has already been selected. If so, then that laser diode does not need to be further checked, and^' the routine jumps to step S9 where the laser diode variable i is incremented by 1 so that the next laser diode, e.g. LD1, may be tested. If laser diode i has not yet been selected, a comparison is made, at step S6, between the usage U(i) of laser diode i, and the minimum usage Umin. If U(i)<Umin, then Umin is set to U(i) in step 7, and a variable j is set to i at step S8, before i is incremented to indicate the next laser diode to be tested in step S9. As Umin takes the usage value U(i) of the least used of the selected laser diodes so far checked in the present S5-S10 loop, a negative determination at step S6, i.e. a determination that U(i)>Umin, indicates that the presently considered laser diode i is not the least used of the as yet unselected laser diodes. Accordingly, that laser diode is not eligible for selection on this loop, and the routine jumps to step S9 to increment i to allow the next laser diode to be tested. From S9, the routine goes to step S10, where it is determined whether or not i is less than the number of diodes available (in this case 16) . If i is less than 16, the routine returns to step S5 so that the remaining diodes in the array may be tested. When all of the laser diodes have been tested, the laser diode indicated by the j variable is selected at step Sll by setting the LDsel(j) flag to 1 (this diode is the least used of the unselected diodes) . The yield YLD(j) available from this laser diode j is then read out from the look-up table of memory means 13, and added to the total output power Pav available from the selected laser diodes at step S12. The procedure returns to step S4, through steps S13 and S14, to allow the selection of further laser diodes j , until enough laser diodes are selected such that the total output power available Pav is 125% that of the power desired Pop. In this case, the routine passes from step S13 to step S15, in which the demand Dm at which each individual laser diode needs to operate is calculated. The demand Dm for each laser diode is set at a percentage of its maximum power output, and depends on the ratio of the desired output power to the power available from the selected laser diodes (100 x Pop/Pav) . The routine also passes to this step S15, through step S14, if the total power available Pav has not yet reached 125% of the power Pop, but all of the laser diodes possible have been selected. In this case, the selected laser diodes will either be able to provide the desired output power by running at or near maximum demand, or the desired output power Pop cannot be achieved even when all of the laser diodes are combined at maximum demand. In this latter case, the comparison at S16 will show that each laser diode will need to be run at a demand Dm greater than 100% to provide the output power, and an error signal will be displayed at step S17 to show that the desired output power is unachievable.

If, at step S16, the demand Dm of each diode is deemed to be less than 100%, the routine passes to steps S18 and S19 where both the power available variable Pav and the laser diode variable i are reset to 0, in readiness for the verification steps S20-S26.

At step S20, a check is made to see whether or not the laser diode i has been selected. If it has, the routine proceeds to step S21, where the total power available Pav is increased by that laser diode ' s proposed output power, i.e. the yield YLD(i) of laser diode i multiplied by the demand Dm at which the laser diodes are to be run. Then, at step S22, the variable i is incremented so that the next laser diode may be checked. If it is determined at step S20 that the laser diode i is not a selected diode, the routine jumps straight to step S22, as there will be no output from that diode.

At step S23, the routine returns to step S20 until all of the laser diodes have been checked, at which point the determination at step S23 is positive and the total power available Pav, calculated by adding the power available from each selected laser diode, is compared with the desired output power Pop to check that the total available power Pav is equal to the desired output power within a preset tolerance. Where this is not true, an error has occurred in the selection, and the routine branches to step S25 to inform the operator of the error by displaying a signal on the display means 14. When the determination is positive, the routine proceeds to step S26, where the apparatus is enables to allow the laser diodes to be activated by the operator. Each laser diode 2 may be calibrated to produce a front face yield of, for example, 2 watts, but losses in the beam combining optics may reduce this power to about 1.6 watts at the fibre port 7. The amount of loss is made up of an amount related to the laser beam path, which is the same for each light source unit 1 of any one specification, and an amount which is specific to each individual light source unit 1 and relates to variations in each unit's assembly.

During manufacture, therefore, the power at fibre port 7 for each diode 2 is measured, and the optical efficiencies of the source calculated from this, and stored in the memory means 13. When the user requests a particular power, the routine of Figure 4 may be modified to allow for these optical efficiencies when calculating the demand required to produce the desired output power.

For some applications, it is preferable to specify the desired power at the fibre tip 4a, rather than at the fibre port 7. The light source unit 1 may therefore be calibrated for a particular fibre 4. This may be done by inserting the fibre tip 4a into a calibration port of calibration means within or external to the unit 1, and firing, for example, a 10 watt (at fibre port 7) pulse into the fibre. The unit 1 then measures the power at the calibration port, and uses this to calculate the percentage efficiency of the fibre. The fibre efficiency can then be used to uprate the desired output power inputted by the operator, so that the power at the fibre tip is the power desired by the user. In some circumstances, it may be important to produce a specific shape or profile of output beam, and if laser diodes 2 are selected for use without consideration of their physical positions, this shape may change, because the position of a diode in the array determines the shape of that laser diode' s beam when it emerges from the fibre tip 4a. For example, as can be seen from Figure 5, for a beam 3 which is incident on the end of an optical fibre 4 at a position off-axis from the fibre's central axis, the resultant beam emerges from the output end of the fibre as a "doughnut mode", i.e. a ring-shaped beam. Figure 6 shows schematically the intensity distributions A and B across a fibre output, as produced, respectively, by a centrally located laser diode (LD0- LD3) and an off-axis diode (LD4-LD15) . The routine of Figure 4 may therefore be modified to choose the laser diodes 2 with regard to their positions as well as with regard to their usage, or the labelling of the laser diodes may be modified accordingly.

For example, the unit 1 may be operated in a normal mode, in which the cone angle is kept approximately the same each time the unit is used, and, to do this, the laser diodes may be divided into an inner group consisting of laser diodes LD0-LD3 and an outer group consisting of laser diodes LD4-15. The laser diodes may then be allocated in pairs, so that for each inner diode activated, an outer diode is also activated, each pair then effectively producing an output corresponding to the curve C of Figure 6. Subject to this constraint, the diodes may be allocated so as to spread their usage, as in the routine of Figure 4.

The unit 1 may also be operated in, for example, a research mode, in which the cone angle is kept as small as possible. To do this, the diodes may be selected starting at the centre of the array, and then working out from there. The diodes D0-LD15 of Figure 3 are labelled so that the routine of Figure 4 produces such a result . This obviously has the effect that the diodes at the centre will wear out sooner than those further out .

The above embodiments are only examples of how the present invention may be put into effect, and variations of the above are also within the scope of the present invention. For example, diagnostic checks may be made by the source to see if any of the laser diodes have failed or are damaged, and flags may be set for each laser diode to indicate if they are unavailable for selection because they have failed, or that they should only be used when high power outputs are required and the outputs of the other laser diodes are insufficient .

Claims

1. A method of operating a plurality of laser diodes to produce a desired combined output power therefrom, wherein selected ones of said laser diodes are activated by electronic means to produce the desired output power, the number of diodes selected being dependent on the desired combined output power.

2. The method of claim 1, wherein the number of laser diodes selected to be activated is chosen so that a substantial number of the selected diodes operate in optimal power ranges.

3. The method of claim 2, wherein said optimal power ranges are between about 70% and about 90% demand.

4. The method of any preceding claim, wherein the laser diodes are selected in sequence until a set percentage of the sum of the yields of the selected laser diodes is equal to or greater than the desired output power, and each selected laser diode is run at a power level set in dependence on the ratio of the sum of the yields to the desired output power in order to produce the desired output power.

5. The method of any preceding claim, wherein the laser diodes are selected so as to tend to equalise the usage of each laser diode.

6. The method of any preceding claim, wherein the selection of said laser diodes favours those of least usage.

7. The method of any preceding claim, wherein the laser diodes are selected to produce a desired output beam configuration.

8. A method according to any preceding claim, wherein the laser diodes are divided into two or more groups, such that the use of a laser diode from any one group has the same general effect on the combined output as the others in that group.

9. A method according to claim 8, wherein the number of laser diodes selected from each group is selected to produce a desired output beam configuration.

10. A method according to any preceding claim, wherein the outputs from the laser diodes are coupled into an optical fibre, and wherein the laser diodes are arranged into an inner group nearer the centre axis of the fibre and an outer group further from the axis.

11. A method according to claim 10, wherein the laser diodes in the inner group are selected in preference to the laser diodes in the outer group.

12. A method according to claim 10, wherein a substantially equal number of laser diodes are selected from each group.

13. A method according to any preceding claim, wherein the demands on the selected laser diodes are chosen so as to produce a combined output beam of a desired configuration.

14. A method according to any preceding claim, wherein the demand on the selected laser diodes are adjusted to take account of losses in beam delivery apparatus associated with the laser diodes.

15. A method according to any preceding claim, wherein a check is made to ensure that a correct number of laser diodes operating at correct demands have been selected.

16. A method according to claim 15, wherein the demand determined for each selected laser diode is multiplied by its yield to produce its output power, the output powers are added together, and their sum is compared to the desired output power to check that they match within a set tolerance limit.

17. A method according to any preceding claim, wherein damaged or failed laser diodes, which are still operable, are non-selectable unless use of the normally selectable laser diodes is insufficient.

18. A light source comprising a plurality of laser diodes, means for inputting a desired source output power requirement, and means for determining the number of diodes to be activated to produce the desired output power in dependence on the desired output power.

19. A light source according to claim 18, wherein the source includes means for recording the usage of each laser diode, and wherein the determining means selects the laser diodes to be activated on the basis of their usage.