BALLANTYNE, Alexander (6 Craiglockhart Park, Edinburgh EH14 1HE, GB)
1. A control system for a laser comprising: a drive circuit for providing a periodic drive current to a laser element; an input for receiving a real time feedback signal representative of the power output of that laser element; sampling means for sampling the level of the real time feedback signal at predetermined discrete times and producing a feedback control signal therefrom; and a drive current controller for determining the periodic drive current of the drive circuit according to a desired laser firing pattern and the feedback control signal.
2. The control system of claim 1 in which the drive circuit is adapted to provide pulsed drive current to the laser element, and in which the sampling means is adapted to sample the level of the real time feedback signal at one or more predetermined times relative to the commencement of a pulse.
3. The control system of claim 2 in which the one or more predetermined times are arranged to be after a predetermined settling time after the rising edge of a drive current pulse.
4. The control system of claim 2 in which the feedback control signal generated from a given drive current pulse is used to determine the drive current for a subsequent pulse.
5. The control system of claim 2 in which the feedback control signal generated from a given drive current pulse is used to determine an adjusted drive current for the remaining portions of the same pulse.
6. The control system of claim 1 further including calibration means for storing calibration data for the laser element, the drive current controller determining the periodic drive current also according to the calibration data.
7. The control system of claim 1 further including a plurality of channels, each channel corresponding to a different laser element and a respective laser output detection element, the drive circuit adapted to provide separate periodic drive currents to each different laser element and the sampling means adapted to sample the level of real time feedback signals from each of the respective detection elements.
8. The control system of claim 7 in which the sampling means is adapted to sample the level of real time feedback signals for different detection elements at different times.
9. The control system of claim 7 in which the sampling means is adapted to sample the level of real time feedback signals for different detection elements at the same time, and to transmit the sample values to the drive current controller in at least partially serial manner.
10. A laser array comprising: an array of laser elements; an array of laser output detection elements, each detection element corresponding to a laser element; and a control system according to claim 7 having a channel corresponding to each laser element and detection element pair.
11. A method for controlling a laser comprising: providing a periodic drive current to a laser element; receiving a real time feedback signal representative of the power output of that laser element; sampling the level of the real time feedback signal at predetermined discrete times and producing a feedback control signal therefrom; and determining the periodic drive current of the drive circuit according to a desired laser firing pattern- and the feedback control signal.
CLOSED LOOP CONTROL OFLASEROUTPUT
The present invention relates to controlling the optical output of solid state lasers, and in particular to the use of feedback in determining appropriate laser drive conditions to achieve a desired laser output.
The use of arrays of semiconductor lasers is becoming increasingly popular in a large number of applications, including thermal printing, computer-to-plate printing, computed radiography, to mention but a few.
In particular, many of these applications require strict control of the optical output of the individual lasers in an array since this determines the 'exposure' of the substrate being subjected to the laser output. The substrate is typically some thermally or optically sensitive medium from which an image can be derived based on the exposure, in printing applications and in computer-to-plate printing, unwanted variations in this exposure clearly affect the quality of the printed image. High quality exposure by the lasers in an array depends on delivering the required amount of power to the required position of the substrate. The laser control system is critical because it determines both the amount of power and the time that it is delivered.
A number of factors influence the output of a laser in addition to the current supplied to it, such as operating temperature, crosstalk from adjacent lasers in an array, ageing of the device, etc. A laser control system must be able to adapt the drive current to take into account these other factors and so maintain the required output power. Adaptation of the drive current can be achieved in a number of ways that can be categorised as closed or open loop systems.
Closed loop systems (sometimes referred to as APC or Automatic Power Control) measure laser output power and use feedback to match the output to a required level. Such a system is described, for example, in WO 02/45007 in which individual laser elements in an array are each optically coupled to a respective photodetector. hi this way, the optical output of each laser can be monitored and
adjustments made to the drive current to compensate for variations in efficiency of each laser element.
Open loop systems do not measure the laser output but control the drive current based on presumed operating conditions and on calibration data. Output power is related to drive current so this method can be effective provided that factors such as temperature and laser history are taken into account when the required current is calculated. Accurate control under all operating conditions is difficult to achieve by this method.
Closed loop schemes are therefore attractive for a number of reasons. Direct control of the critical parameter, laser power, is obtained resulting in consistent output despite temperature, ageing and other effects. Closed loop systems typically offer automatic laser behaviour control after initial calibration.
However there are significant disadvantages with existing closed loop implementations using continuous feedback of the measured power. The accurate measurement of laser output inevitably incurs some delay and limits the speed of the system. For high throughput printing systems, this delay can be a significant penalty. Ideally, the control loop determines the output at all times but to maintain stability the rate of change of the output must be significantly slower than the speed of the slowest element in the measurement and feedback system. There are also significant implementation difficulties in individually addressed laser arrays associated with noise and crosstalk between laser elements due to the effect of high rate of change of laser drive current, the small value of the signal representing the output power and the close proximity of many lasers and their drive circuits. These issues exist with both analogue and digital implementations.
It is an object Of the present invention to provide an improved laser control system that has particular, though not exclusive, application in control of solid state laser arrays.
According to one aspect, the present invention provides a control system for a laser comprising: a drive circuit for providing a periodic drive current to a laser element; an input for receiving a real time feedback signal representative of the power output of that laser element; sampling means for sampling the level of the real time feedback signal at predetermined discrete times and producing a feedback control signal therefrom; and a drive current controller for determining the periodic drive current of the drive circuit according to a desired laser firing pattern and the feedback control signal.
According to another aspect, the present invention provides a method for controlling a laser comprising: providing a periodic drive current to a laser element; receiving a real time feedback signal representative of the power output of that laser element; sampling the level of the real time feedback signal at predetermined discrete times and producing a feedback control signal therefrom; and determining the periodic drive current of the drive circuit according to a desired laser firing pattern and the feedback control signal.
Embodiments of the present invention will now be described by way of example and with reference to the accompanying drawings in which: Figure 1 shows a schematic diagram of a laser control circuit;
Figure 2 shows a more detailed schematic diagram of the laser control circuit of figure 1; and
Figures 3a to 3e illustrate exemplary pulse waveforms representing laser output.
With reference to figure 1, there is shown a laser control system 10 using sampling feedback according to one preferred embodiment of the present invention. The control system 10 comprises a custom driver circuit 11 controlled by digital logic
circuit 12. In a preferred arrangement, the driver circuit 11 is implemented using an application-specific integrated circuit (ASIC) and the digital logic circuit 12 is implemented using a field programmable gate array (FPGA).
The control system 10 is connected to an optical sub-assembly 1, comprising laser element 2 and corresponding photodiode detector element 3. Figure 1 illustrates one channel of the laser control system. It will be understood that in most applications, the optical sub-assembly 1 will include a plurality of lasers 2, preferably formed in one or more monolithic arrays, and a plurality of photodiode elements 3, also preferably formed in one or more monolithic arrays. Preferably, the number of photodiode elements 3 is exactly correspondent to the number of laser elements 2, on a one-to-one basis, but it will be understood that other configurations are possible. For example, several proximate laser elements 2 could be monitored by the same photodetector element 3. This may be appropriate for monolithic laser arrays where there is only a very small divergence in laser output characteristic between very closely adjacent lasers in the same array. Preferably, the laser elements 2 and photodiode detector elements 3 are mounted on or formed within a common substrate. In preferred arrangements, the photodiode detector elements 3 may be configured to receive optical output from the rear facet of an edge emitting laser 2.
Where there are multiple laser elements 2 and detector elements 3, it will be understood that the control system 10 is scaled to have multiple channels, each channel controlling a separate laser element 2. The control system 10 can be extended to control arrays of any size.
A control interface 20 is used to receive and load external data. This external data may include calibration data for initial set up of the control system, for example data relating to the operating characteristics of the laser elements 2, such as feedback and/or output gain. Calibration data is stored in a calibration register 22 and is usually necessary because of manufacturing tolerances in the laser elements and/or detection elements. The external data received through control interface 20 also includes a dot pattern and/or dot intensity values representing the laser drive
pattern to be implemented. In a printing application, this dot pattern would define the image to be printed or transferred to a printing plate. The dot pattern is loaded into a dot pattern register 21. In the case of a single channel, the dot pattern may be simple 'on' or 'off data, but may also include a desired intensity value if variable intensity or spot size laser output is required. For a multi-channel system, the dot pattern may relate to one or two dimensional arrays of dots at a time. The required dot patterns are periodically loaded into the register 21 ready for driving the laser elements 2. The dot intensity, i.e. the power level of a drive pulse delivered to the laser element, is calculated based on: (i) the required intensity from register 21, (ii) on calibration data from register 22 and (iii) on measured intensity of output of the laser element from a feedback loop to be discussed below. The power level is determined by the digital control system 23 which produces a digital control signal on line 24. The power level may be determined by magnitude and/or duration of a drive pulse delivered to the laser element.
The digital control signal on line 24 is received by a digital to analogue converter (DAC) 30 which produces an analogue current control signal on line 31 to driver 32. Driver 32 delivers the required drive current to laser element 2 via output 33.
Any suitable compensation algorithm can be implemented within the digital logic circuit 12 using established digital signal processing techniques for discontinuous sampled systems and statistical sampling systems.
The driver circuit 11 and digital logic circuit 12 include timing elements 38, 28 respectively, implementing a suitable timing synchronisation as will be understood by one skilled in the art.
The feedback loop comprises photodiode element 3, an amplifier 35 and an analogue to digital converter (ADC) 36. The digital feedback signal from ADC 36 is fed to the digital control system 23 on line 37.
Thus, in a general aspect, it will be recognised that the digital control system 23, DAC 30, and driver 32 effectively provide a means for providing a periodic drive current to one or more laser elements 2.
A particular feature of the present invention is that the feedback circuit is adapted to provide fast and efficient feedback on the output of laser elements 2. Conventional closed loop feedback circuits rely on a continuous representation or indication of the laser output being fed back to a control circuit to vary the drive current being delivered to the laser. An exemplary waveform 100 showing a laser output pulse as might be expected using a prior art closed loop laser control circuit (e.g. as described in WO 2005/041370, US 2003/138010 or EP 0468416) is illustrated in figure 3 a. This is characterised by a relatively slow turn-on slope (rising edge) 101, a relatively slow turn-off slope (falling edge) 102, and a period of instability 103 following the rising edge. The rate of turn-on and turn-off are generally restricted to maintain the stability of the control loop. The period of instability following the 'rising edge can occur for a number of reasons including: (i) marginal loop stability resulting from minimising the rise-time such that phase shifts in the feedback loop become significant; (ii) optical or electrical cross-talk from adjacent lasers turning on (or off) at the same time; and (iii) parasitic effects in the control device or the sensor feedback due to the rapid rate-of-change of current, especially if many lasers turn on (or off) together. In practice, a variable combination of these effects occurs depending on operating conditions. However, the target output power is achieved during a period 104 after the transient instability 103 has decayed.
By contrast, as shown in figure 3b, open loop control of a laser output 110 is characterised by a relatively rapid turn-on slope 111, a relatively rapid turn-off slope 112, a limited overshoot period 113 and a relatively long stable output period 114 at or near to target power. An improvement in pulse form can be obtained by eliminating instability caused by external sensor feedback, and by reducing sensitivity of a controller to the causes of the instability following the rising edge. Current control circuitry is less sensitive to noise than automatic power control circuitry. However, the waveform 110 shows that the stable output period 114 is
likely to exceed (or fall short of) the target power due to the indirect relationship between controlled parameter (laser current) and the required output (light amplitude).
The present invention recognises that a good representative discrete sample of output from the laser is all that is required to provide good feedback. By taking only discrete samples of the laser output, at predetermined times, it is possible to realise a number of advantages not envisaged in the prior art.
With reference to figure 2, a preferred embodiment of a feedback circuit is now described. The real time output 4 from photodiode detector element 3 is passed to amplifier 35. The output of the amplifier 35 is sampled by a sample and hold circuit 40 which captures the level of the real time feedback signal from the detector element 3 at predetermined discrete times. Those sampled values are converted to the digital domain by ADC 36. The sample and hold circuit 40 is preferably adapted to capture the output of the amplifier 35 at a predetermined time after commencement of a drive pulse, e.g. following a rising edge thereof. The predetermined time is preferably established to be shortly after any set-up delays and settling times in the driver 32, laser element 2, photodiode element 3, and amplifier 35. In other words, the predetermined time at which a sample of laser output is taken is when the laser output is stable, or expected to be stable.
Figure 3 c illustrates a suitable period or window 124 for sampling the level of the real time feedback signal 120, shortly after any instability period 123. It will be noted that, in the absence of continuous closed loop feedback, the turn-on and turn- off times 121, 122 can be similar to those of the open loop system. In exemplary embodiments of the invention, it is desirable to have a laser pulse width of ≤ 1 microsecond with a rise time 121 and fall time 122 of < 100 ns each. The sample window 124 is preferably of a duration of the order of 100 ns and is located after a settling time 125 at any suitable point in the pulse waveform 120 after instability period 123.
Although sampling only a single discrete sample value is preferred for each drive pulse to the laser element 2, it will be understood that multiple discrete values for each drive pulse could be used and filtered appropriately. Alternatively, subsequent discrete values attributable to the same drive pulse may be used to verify that the laser output has stabilised. Then, one or more of the discrete values known to be representative of a stable laser output may be used by the digital control system 23.
Taking only discrete sample values of the feedback signal offers some advantages. Firstly, it is ideally suited to laser drive currents that comprise a series of 'on-off type pulses. The feedback signal is generally relevant only during an on-pulse, and it is easy to ignore or screen sample values corresponding to an off pulse. The feedback signal can be captured / sampled at any appropriate time during the pulse to optimise noise immunity and processing time. Thus, faster and more accurately controlled pulse sequences become possible for a given technology.
Secondly, for an array of lasers, the digital control system 23 may be capable of handling feedback signals from many laser elements in a serial fashion. This substantially reduces the circuitry required for large laser arrays. The number of connections between the driver circuit " 11 and the digital logic circuit 12 is minimised to facilitate the control of a large number of channels.
A demultiplexer 41 retrieves the individual channel feedback samples. These are logically combined with any required calibration adjustments in process 42 to provide an adjusted power measurement. The adjusted measurement is processed b) ? a filter algorithm in loop filter 43 to be used as input to a discrete time feedback control system to adjust a subsequent laser drive pulse. Process 44 combines the calibrated feedback data with a predetermined power set point 45 according to the laser intensity required. This generates a drive output 46 which, is switched on or off according to dot pattern data 21. TMs can be implemented for each channel in serial or serial/parallel fashion, depending upon processing resource available. The drive current data for each channel, adjusted according to a feedback signal from a previous drive pulse, may then be transmitted in serial fashion, using multiplexer
47 to the drive circuit 11. For typical pulse widths of 1 microsecond, and sampling windows of 100 ns or less, it may be possible to multiplex a number of feedback channels on one serial link 37.
The drive circuit 11 initiates laser activation by turning on the driver 32.
New dot patterns are loaded into the register 21 at suitable intervals to keep pace with laser activation, e.g. printing.
In preferred arrangements, the feedback signal for each drive pulse is used to recalculate the drive current required for the immediately subsequent drive pulse, although it will be understood that the feedback signal may be used for one or more later subsequent pulses. Depending upon the speed of the feedback circuits and drive pulse durations, it may be possible to use a sampled feedback signal from an early portion of a pulse to provide feedback to adjust the same drive pulse, e.g. by adjusting its amplitude mid pulse or its total duration, e.g. delaying the off time of a pulse according to early feedback data.
An example of the pulse waveform 130 for such an arrangement is given in figure 3d. After a feedback sample is taken in sampling window 134, a small correction may be made to the laser drive current in order to bring the output to the target amplitude for period 135. It will be understood that the correction could be made in a downward direction, or upward direction as shown by dashed line 136, dependent upon the displacement error from target amplitude during the sampling window. In a still more sophisticated arrangement, the correction made to the laser output during period 135 may be deliberately overcompensated in order to take into account over- or under-power established during the sampling window 134.
In this way, the feedback circuit can more accurately control the total energy of the pulse (e.g. area under the curve 130). Another way of achieving this, as discussed above, is to extend or foreshorten the overall duration of the pulse according to whether the pulse height during the sampling window is below or above target.
In most applications, the accuracy afforded by subsequent pulse feedback is adequate, as illustrated in figure 3e. A first pulse 141 may be slightly off target amplitude (e.g. similar to open loop control, figure 3b). However, subsequent pulses 142 will be accurately controlled. Providing that there is a reasonably high pulse repetition rate, drift from target amplitude is held at insignificant or very low levels. In exemplary embodiments, subsequent pulses 142 are expected to achieve a flat pulse height within 2% of " target compared with corresponding figures of within 20% for an open loop system and within 1% for a continuous closed loop system. At the same time, the invention offers an instability period 133 of similar order to open loop systems (e.g. less than 100 ns for a 1 microsecond pulse width), and substantially less than continuous closed loop systems which may be expected to be of the order of 200 ns.
It will be recognised that the present invention has particular benefit in printing and other imaging applications where large arrays of lasers, usually linear arrays, are used for scanning across an imaging substrate and in which lasers in the array are typically fired simultaneously when an image dot is required for each laser location. By processing feedback data serially for many lasers, the overall circuit real estate on-chip can be substantially reduced allowing smaller device sizes and therefore higher laser spot density.
The preferred implementations of figures 1 and 2 use digital processing circuits in logic circuit 12, but it will be understood that principles of the invention can be implemented in the analogue domain. The implementation of figures 1 and 2 partitions particular function blocks into programmable and application specific devices but other arrangements are also envisaged.
Other embodiments are intentionally within the scope of the accompanying claims.