GB2562138B

GB2562138B - Laser power controller

Info

Publication number: GB2562138B
Application number: GB1800531.4A
Authority: GB
Inventors: Coue Dominique; Redman-White William
Original assignee: Hilight Semiconductor Ltd
Current assignee: Hilight Semiconductor Ltd
Priority date: 2018-01-12
Filing date: 2018-01-12
Publication date: 2019-06-12
Anticipated expiration: 2038-01-12
Also published as: GB201800531D0; GB2562138A; CN110034487B; JP7305354B2; CN110034487A; JP2019145785A

Description

Laser Power Controller

Description of Invention

In a fibre optical communications system, it is important to be able to control the output power of the transmitting laser diode for a number of reasons. Firstly, the average and peak power of the laser must not exceed certain limits in order to avoid damage. Secondly, the different power levels corresponding to binary (or other radix) data values must be set so that the modulation index (alternatively defined as extinction ratio) is within the overall system specifications to ensure reliable reception at the end of the link. One difficulty to be addressed in any control system is that the characteristics of the laser can change significantly with temperature and also over time with ageing, and diverging from an ideal linear response, so that a conventional factory set up of the “high” and “low” drive current levels is not adequate.

Numerous techniques exist in prior art that describe methods intended to estimate the instantaneous values of the minimum and maximum transmitted optical output and compensate for the changes in device characteristics. Most are limited in their effectiveness due to the restricted bandwidth of the monitor diode and its associated circuitry. Others require specific patterns to be present in the data stream, or to be deliberately inserted into the data stream in some defined way.

Monitoring the transmitted output power is even more challenging in an optical communications link that transmits the data in a series of discrete bursts, as a simple average value of the optical output may vary greatly over time, and the instantaneous levels are not stable enough for most methods described in prior art to reach adequate estimates of minimum and maximum levels. The temperature related effects are likely to be even more severe, as the transmitting laser diode may be in an off state for a long period before being activated for a data burst, and hence may have cooled to ambient temperature before heating up during a data burst.

Hence it is desirable to be able to sense the minimum and maximum optical outputs corresponding to logic “1” and logic “0” during data bursts on a near continuous basis. It is further desirable to be able to make such measurements using a transmit power monitoring function with only moderate bandwidth, and by means that do not disturb the transmitted data payload nor compromise the received signal to noise performance. Such an approach has been proposed in prior art in application GB1611938.0. However, the aforementioned method has the benefit of only one data point per data burst, and thus has some susceptibility to noise inasmuch as it may have an unacceptable effect on the calculation of the required laser current values. It is an object of the invention to achieve an improved accurate and robust control of a laser output in a burst-mode optical communications system by using further measurements of the optical levels during each data burst.

According to a first aspect there is provided a system for transmitting a sequence of at least two data bursts in a fibre optical communications system, the system comprising: selection circuitry configured to select one of a data input value, a logical high value or a logical low value such that the selection circuitry is configured to select the data input value during a data transmission period during a defined burst period and one of the logical high value and the logical low value during an extension time period during the defined burst period and immediately following the data transmission period, such that for the sequence of at least two data bursts, at least one data burst is a logical low value burst and at least one data burst is a logical high value burst; drive circuitry configured to apply a current to a laser diode, the current corresponding to the one of a data input value, a logical high value or a logical low value selected by the selection circuitry during the defined burst period or a zero value otherwise, the current being such that the laser diode is configured to provide an optical output; an optical sensor module configured to provide a sensor module output corresponding to the optical output of the laser diode; wherein the sensor module output is configured to provide electrical outputs proportional to the optical outputs of the laser diode corresponding to the logical high value and the logical low value in the sequence of bursts; and further configured to provide an output corresponding to an average value of the sensor module output during only the data transmission period during the sequence of bursts; and a controller configured to receive desired values regarding optical signalling output power levels of the laser diode and to receive the output from the optical sensor module proportional to the optical output of the laser diode corresponding to the logical high value and the logical low value and to receive the output corresponding to the average value of the sensor module output during only the data transmission period during the sequence of bursts; wherein the controller is configured to use the outputs from the optical sensor module and the desired values to provide control values for the drive circuitry.

The optical sensor module may comprise a photodiode output power detector.

The optical sensor module may comprise an optical sensor and a transimpedance amplifier, the trans-impedance amplifier being configured to provide the sensor module output.

The control values may be configured to control at least one of: an average power of the optical output of the laser diode; a power of the optical output of the laser diode representing a logical high; a power of the optical output of the laser diode representing a logical low; and a modulation index of the optical output of the laser diode.

The current may comprise a steady element and a variable element.

The drive circuitry may be configured to set the current applied to the laser diode dependent on a combination of a bias control value and a modulation control value.

The control values may be configured to control the drive circuitry to set the at least one of a bias current and a modulation current applied to the laser diode.

The drive circuitry may comprise bias circuitry configured to provide a bias current to the laser diode.

The drive circuitry may comprise modulation circuitry configured to provide a modulation current to the laser diode.

The drive circuitry may be configured to set the current applied to the laser diode dependent on a combination of an average value and a modulation value.

The burst period may be gated by a burst enable signal.

The control values may control the drive circuitry to deliver the desired logical high and logical low optical output power levels.

The extension time period may be greater than a settling time of the sensor module output.

The selection circuitry may alternately select one of the logical high value and logical low value for each consecutive extension time period.

The selection circuitry may select the logical high value or the logical low value for each consecutive extension time period according to a pre-defined sequence.

The selection circuitry may select the logical low value immediately after an extension time period where the logical high value has been selected.

The selection circuitry may comprise a selector switch function. A bandwidth of the selection circuitry may be configured to be able to switch between the data input, the logical high value and the logical low value in a time significantly less than that of the extension time period.

The control values for the drive circuitry may be based on a combination of the average and high and low values from the optical sensor module each scaled by a coefficient.

The system may comprise substantially digital circuits.

The control values for the drive circuitry may be calculated by a digital calculation function.

The system may comprise substantially analogue circuits.

According to a second aspect there is provided a method for transmitting a sequence of at least two data bursts in a fibre optical communications system, the method comprising: selecting one of a data input value, a logical high value or a logical low value wherein selecting comprises selecting the data input value during a data transmission period during a defined burst period and one of the logical high value and the logical low value during an extension time period during the defined burst period and immediately following the data transmission period, such that for the sequence of at least two data bursts, at least one data burst is a logical low value burst and at least one data burst is a logical high value burst; applying a current to a laser diode, the current corresponding to the one of a data input value, a logical high value or a logical low value selected during the defined burst period or a zero value otherwise, the current being such that the laser diode is configured to provide an optical output; providing an output corresponding to the optical output of the laser diode, wherein providing the output corresponding to the optical output of the laser diode comprises providing electrical outputs proportional to the optical outputs of the laser diode corresponding to the logical high value and the logical low value in the sequence of bursts and providing an output corresponding to an average value of output corresponding to the optical output of the laser diode during only the data transmission period during the sequence of bursts; receiving desired values regarding optical signalling output power levels of the laser diode; combining the average and high and low values of the outputs corresponding to the optical outputs of the laser diode wherein each value is scaled by a coefficient and the desired values regarding optical signalling output power levels of the laser diode and using these combined values to provide control values to control a current applied to a laser diode.

The method may further comprise applying the control values to control at least one of: an average power of the optical output of the laser diode; a power of the optical output of the laser diode representing a logical high; a power of the optical output of the laser diode representing a logical low; and a modulation index of the optical output of the laser diode.

The current may comprise a steady element and a variable element.

The method may further comprise setting the current applied to the laser diode dependent on a combination of a bias control value and a modulation control value.

Setting the current applied to the laser diode may comprise setting at least one of a bias current and a modulation current applied to the laser diode based on the bias control value and modulation control value.

Applying the current may further comprise providing a bias current to the laser diode.

Applying the current may further comprise providing a modulation current to the laser diode.

Setting the current applied to the laser diode may comprise setting the current dependent on a combination of an average value and a modulation value.

The burst period may be gated by a burst enable signal.

The method may further comprise applying the control values to deliver the desired logical high and logical low optical output power levels.

The extension time period may be greater than a settling time of providing the output.

Selecting one of a data input value, a logical high value or a logical low value may comprise alternately selecting one of the logical high value and logical low value for each consecutive extension time period.

Selecting one of a data input value, a logical high value or a logical low value may comprise selecting the logical high value or the logical low value for each consecutive extension time period according to a pre-defined sequence.

Selecting one of a data input value, a logical high value or a logical low value may comprise selecting the logical low value immediately after an extension time period where the logical high value has been selected.

Selecting one of a data input value, a logical high value or a logical low value may comprise selecting based on selector switch function.

Selecting one of a data input value, a logical high value or a logical low value may comprise switching between the data input, the logical high value and the logical low value in a time significantly less than that of the extension time period.

Using the output corresponding to the optical output of the laser diode and the desired values to provide control values for the drive circuitry may comprise providing control values based on a combination of the average and high and low values from the optical sensor module each scaled by a coefficient.

The invention will now be described solely by way of example and with reference to the accompanying drawings in which:

Figure 1 shows typical arrangements for a transmitter in a burst-mode optical fibre link using uni-directional or bi-directional modulation current.

Figure 2 shows a representation of a laser diode output characteristic and temperature effects.

Figure 3 shows the limitations of conventional estimation methods where there is curvature in the laser characteristic.

Figure 4 shows the structure of a typical data burst with typical allowable laser turn off time.

Figure 5 shows a burst mode optical signal with high and low reference levels embedded within valid data packets.

Figure 6 shows a burst mode optical signal with low reference levels embedded within valid data burst periods

Figure 7 shows a burst mode optical signal with high reference levels embedded within valid data burst periods.

Figure 8 shows an embodiment of the invention using uni-directional modulation current.

Figure 9 shows an embodiment of the invention using uni-directional modulation current.

Figure 10 shows a further embodiment of means to sense the reference levels.

Figure 11 shows a further embodiment of means to obtain a gated average value from an input.

The description is not to be taken in a limiting sense but is made merely for the purposes of describing the general principles of the embodiments of the invention. For example, operations that are illustrated as being performed using digital signals and digital circuits may also be achieved using substantially analogue signals and analogue circuits. FIG. 1 shows typical arrangements in a transmitter suitable for an optical communications system. A laser diode 101 is provided with a current by drive circuitry having a steady element and a variable element. This may be in the form of a smaller steady bias current 115 with a modulation current 116 that is disconnected by means of a switching function 110 to indicate a logical low level in the modulation data input 107. Alternatively, this may be in the form of an average current 125 with a bidirectional modulation currents 126 and 127 adding and subtracting current following selection by a switch 120 under control of the incoming data stream 107 to create the maxima and minima in the optical output. These currents may be provided by digital-to-analogue converters (DACs) 111 and 112 in the case of uni-directional modulation or 121 and 122 in the case of bi-directional modulation current, these DACs having current outputs controlled by digital values 113 and 114 or 123 and 124 respectively, whose values are set by the controller function 117 or 128 respectively.

When operating in a burst mode, these currents may be gated in a manner corresponding to the active transmission periods in a data burst by means of a further signal or signals 108 corresponding to the prescribed length of the transmission burst.

The optical output of the laser diode 101 is sensed by an optical sensor, such as a monitor photodiode 102, to create a current proportional to the sensed optical level. This said current may be sensed directly, but is more commonly converted to a voltage 105 with a trans-impedance amplifier 103. The combination of the monitor diode 102 and amplifier 103 typically have a bandwidth that is substantially less than that of the main data channel bandwidth. This monitor value 105 may be converted to digital form 106 by means of an analogue-to-digital converter 104 and these data may be used by the controller 117 or 128 to calculate and set the current levels for the laser diode according to some algorithm. The monitor diode 102 and its associated amplifier 103 typically have a signal bandwidth that is much less than the bandwidth of the data transmitted by the laser 101, and this limitation of the monitor channel bandwidth is very significant in the implementation of any transmit optical level control mechanism since it restricts the observability of the maximum and minimum peak and trough values of the transmitted optical signal. FIG. 2 is a diagrammatic representation of the characteristics of a typical laser diode as is used in optical communications systems. When used to generate a modulated optical signal, the current through the laser diode is modulated such that the minimum current is above the threshold value 203 for the laser 101, and the maximum current is below the manufacturer’s ratings for the device. When a laser diode is cold, or the current levels are relatively low, a simple linear model 201 may suffice. However, when the laser diode heats up, or as its characteristics change with age, the threshold current may change 204 and the relationship may exhibit a more curved shape 202. Thus, maintaining the desired optical output and the desired modulation depth during operation over a system’s lifetime is not considered trivial.

In any given practical system, the maximum current may be set so that the average operating power of the laser is set to a defined level with regard to the required signal level for reliable communications to be established. A critical parameter in such a system is the ratio of the maximum to minimum optical output, usually referred to as the Extinction Ratio (ER), as this affects the signal to noise levels for the receiver. The ER is a function of the minimum and maximum laser diode current values, and is sometimes represented as a simple linear relationship, but in reality this is not an accurate representation. FIG 3 shows how average optical power 303 of a laser diode at an elevated temperature is not suitable as the basis for an accurate estimate of the minimum 305 and maximum 306 laser drive current levels necessary to generate the specified minimum 301 and maximum 302 optical output levels and hence and hence maintain the desired ER. The drive current 304 corresponding the observed average optical output will not be the average of the actual minimum 305 and maximum 306 current levels.

Where a system operates with a continuous data stream, the laser can reach a steady state temperature that is relatively easy to monitor. Further, there is sufficient time to gather data from a monitor diode system to estimate the peak and trough optical data levels with some kind of averaging of the measurements to provide a reliable estimate of the ER and average optical power. Systems for this purpose are known in prior art (for example, Smith et al, Electronics Letter Vol 14,1978, and similar derivative arrangements) often using a slow modulation of the absolute drive current levels.

Figure 4 shows the general form of optical signals intended to transmit data bursts in a system adhering to specifications for burst mode operation, (such as, for example, standard ITU-T Recommendation G.984.2). The bias current to the laser is gated by a burst enable signal 108 before data signals 107 are used to modulate the laser output. In such standards the duration T1 of the data burst 403 is precisely defined, and typically of the order of a few hundred nanoseconds. Note that at the end of a data burst, the logical value may be in a high state (a logical high value) or a low state (a logical low value). Such standards also typically define T2 a time interval 404 within which the laser output must return to zero. To allow for the bandwidth of practical bias control systems, this interval is of the order of 10ns (and in the specific example given 12.8ns is the defined value).

In such a burst mode system the problem of controlling the average power and ER is difficult. Before the start of a burst the laser will be in a relatively cool state. As soon as the data packets are transmitted, the laser will begin to heat up and will continue to do so during a typical burst. It is a requirement of the standards that the system be operational after only a short number of training bursts, for example 5 or less, during which the system’s operating parameters should come under control. Means for establishing the operating parameters in a timely fashion have been disclosed in the patent GB2535553B wherein defined amplitude trial bursts are output in order to determine an estimate of the slope efficiency of the laser at the start of a train of data bursts. Maintaining the operating conditions after the operation of such a start-up system solely by monitoring the average optical output is not generally satisfactory due firstly to the intermittent bursts of optical power, and secondly, due to the reliance on there being data content in each burst that has a well-defined average value. This latter requirement demands that the numbers of data 0 and 1 values in the burst are substantially equal which may not be guaranteed.

There remains a further requirementto provide means for accurately controlling the ER of the laser output after the initial training bursts where the laser has substantially warmed up to an elevated average temperature. Any measurements of the peak and trough values have the same monitor channel bandwidth limitations as in a continuous mode system, but the demands are further complicated by the intermittent nature of the signal making the task more difficult.

In an embodiment of the invention means are provided to make rapid and accurate estimates of the instantaneous values of the optical output representing data ‘1’ and/or data Ό’ values, or other such values as may be defined, and means are also provided to make estimates of the average value of the optical output solely during the data content period of each data burst, all of said means operating in a manner that does not require modification of the data content of the said burst. Using the said estimates, further means are provided that are able to calculate the required values of the drive currents to deliver the desired output levels, and to maintain these levels notwithstanding changes in the laser characteristics due to short term heating and/or long term ageing. The currents may be in the form of a smaller bias current and a unidirectional modulation current, or an average current and bi-directional modulation current as appropriate for the system.

In Figure 4 it will be noticed that the time to turn off the laser after a burst of data may not be a constant duration, but is likely to depend on the logical value present at the end of the data transmission period 401. The process of turning off the laser bias current (or average current) at the end of a data transmission period from a high state 405 is likely to be significantly greater than the laser bias turn-off time at the end of a data transmission period from a low state 406. This turn-on and turn-off time is typically determined by the response time of the internal circuitry that maintains the bias current 115 (or average current 125), and said circuitry is typically not designed to respond at the same rate as the data modulation of the laser. However, the bandwidth of the modulation circuit 110 or 120 in response to the modulation data signal 107 is necessarily very fast in order to switch the laser current at the data symbol rate. Hence rather than use the bias current (or average current) control to turn off from a high state, the modulation circuitry 110 or 120 may be used to reduce the laser output very rapidly to the low state first, typically in a time of the order of tens of picoseconds. Once the laser output is in said low state, the task of turning off to full extinction becomes much easier. Further, it is not a difficult task to ensure that the bias current 115 (or average current 125) responds to the burst enable signal 108 or a substantially equivalent signal in a time interval substantially less than the interval 404 required by the standard. This approach makes available a time interval that while not large, is nonetheless greater than the transient settling time typical of such monitor channel circuits. Using this knowledge it is possible to exploit the time available in the specified turn-off interval 404 to execute valuable measurements of the prevailing optical high and low output levels.

Concurrent with said measurements of the optical high and low values during a short time extension of the data burst, it is also possible to gather useful information on the average optical power present during the transmission of the data content of said bursts wherein the average output of the monitor photodiode circuitry 105 is taken only during the duration of the data in the burst and also outside of any settling time required by the monitor circuitry. Gating for such averaging is easily derived from the burst enable signal 108 in some combination with the data input 107 and other internal logical signals.

Figure 5 shows the optical levels associated with a burst mode system wherein subtle modifications have been made to the transmitted signal that facilitate measurements of the high and low levels. Said modifications are made such that they do not affect the normal transfer of data within the burst packets and do not transgress the specifications set by the relevant transmission standards.

To provide the framework for said modifications a time interval is first defined to satisfy the conditions that said time interval is substantially less than the laser turnoff time 405 allowed by the transmission standard but long enough to be substantially longer than the settling time of the monitor channel output 105 and at the same time allows sufficient remaining time within the period 405 for the bias current control circuitry to extinguish the laser completely. A feature of the invention is the replacement of the raw data signal 107 with a modified form of the laser modulation signal 501 wherein at the end of each burst a known logical value is held for an extended time period T3 502. At the same time, the bias current to the laser 115 (or average current 125) is controlled by a modified version of the burst enable signal (the laser current enable control signal 506) such that the bias (or average) and modulation circuitry remain active for a defined period after the data for that burst has ceased.

The logical value of this extension of the data burst may conveniently be made to alternate between a ‘1’ denoted 503 in Figure 5 and a Ό’ denoted 504 in Figure 5. Alternatively the logical value of this extension of the data burst may be set to be Ό’ for several successive data bursts as depicted in Figure 6. Alternatively the logical value of this extension of the data burst may be set to be ‘1’ for several successive data bursts as depicted in Figure 7. The duration of this logical value holding period 502 is made to be sufficiently long for the monitor channel output 105 to be able to settle to a substantially accurate measurement result. If the logical value held at the end of the data burst is ‘T, then the laser modulation current 115 (or 125 in the case of bi-directional modulation current) is returned to a Ό’ at the end of this extension period 502 by means of a command edge 505 of the laser current enable control signal 506 to the data modulation circuitry 110 (or 120 in the case of a bi-directional modulation). In this way, the laser current is reduced substantially towards its extinction state by means of the high bandwidth modulation circuit function typically in some tens of picoseconds, rather than by a possibly much slower bias current control. Immediately this state has been reached, the bias current 115 (or average current 125) and modulation current 116 (or 126 and 127 in the case of bi-directional modulation) are turned off by the laser current control signal 506 and the total current in the laser decays to zero before the end of the time permitted by the relevant standard. By these or substantially similar means it is therefore possible for the monitor output 105 to deliver substantially accurate estimates of the true prevailing optical outputs during both logical high ‘1’ and logical low Ό’ data states, without significant restrictions arising from particular data patterns and/or run lengths as is often the case in prior art.

During the data burst period the burst enable signal 108 in combination with the data input 117 and other internal signals are used to provide a gating signal 508 that is active for a period entirely within the data transmission period and typically active for most of the said data transmission period wherein said gating signal is used to enable an averaging function to compute the average value from the monitor channel output 105. Said gating signal 508 is configured to become active only after the expected settling time of the monitor channel output resulting from the start of the data burst and as a result of the restricted bandwidth of the monitor channel. This requirement for a delay from the start of the burst enable signal is typically of the order of several data symbol periods. The averaging function may in one embodiment hold the average value at the end of a burst and use this value in some proportion with the fresh input signal during the next burst in order to create a rolling average rather than an estimate based on a single burst.

From these measurements taken from a plurality of data bursts the analogue values representing the optical Ό’ optical‘1’ and optical average may be converted into digital form and a simple algorithm may be employed combining these values to determine the prevailing extinction ratio and the average optical power in an advantageous way that minimises susceptibility to noise and errors. Said algorithm may also determine any required adjustments to the modulation current 116 (or 126 and 127 in the case of bi-directional modulation) and the bias current 115 (or correspondingly average current 125) such that the ER and average power correspond with the desired target values for the system.

Figure 8 shows an arrangement according to an embodiment of the invention. The bias current 115 is set by a current output digital-to-analogue converter (DAC) 111 and the modulation current 116 is similarly set by another DAC 112. The controlling digital values for said DACs are determined by a digital calculation function 826 which takes its inputs from the system feedback values and the digital inputs corresponding the desired average power 131 and modulation depth 132 (or ER). The modulation circuitry 110 is no longer controlled directly by the incoming data input 107 but can now have its input switched between the data input 107, and logic ‘1’ or logic Ό’ by means of selection circuitry, for example a selector switch function 813. When the burst enable signal 108 is asserted to indicate the start of a data burst the logical control function 811 will set the modulation input path using selector 813 to pass the incoming data directly to the modulation circuitry 110. A modulated optical signal will be generated by the laser 101 and a band-limited representation of same 105 will be created by the monitor diode 102 and its associated amplifier 103. This monitor signal 105 is converted directly to a digital value 821 by an analogue-to-digital converter (ADC) 820. During the data transmission part of the data burst this output 821 may be used but it will be of limited value due to the bandwidth limitations of this channel. At the end of the data payload the burst enable signal 108 will indicate the end of this transmission period.

In a conventional system the de-assertion of the burst enable signal 108 would disable the modulation 116 and bias 115 currents completely. According to this embodiment of the present invention, the control logic 811 takes a defined delay time and holds the bias and modulation currents on. An additional burst status signal 810 is provided by the embodiment that can change the logical value as required with each data burst, effectively designating bursts as “HIGH” or “LOW”. As an example embodiment, if the burst is designated as “HIGH” then during the delay at the end of the burst, the modulation input selector 813 is set to a logical ‘T 503 such that the optical output is held at the high level 302. This modulation optical value is held for a time period 502 long enough for the monitor channel to make an accurate measurement despite the limited bandwidth of said monitor; but still short enough that there is time to fully extinguish the laser within the time allowed 404 by the transmission standard. The monitor channel output 105 is converted to digital form 821 and then passed at a suitable time instant to a first register 824 via a logical gate 822 enabled by the burst status signal 810. This register then provides the measured optical high value to the calculation function 826.

At the end of this delay period 503 the modulation selector may remain selecting a logical ‘1’ or may be set to a logical Ό’ to remove the laser modulation current 116 using the normal modulation circuitry 110 and hence reduce the optical output very rapidly. At the same instant 505, the control logic 811 commands the bias current DAC 111 and the modulation current DAC 112 to cease outputting current, such that the laser 101 becomes completely extinguished within the period 404 required by the relevant communication standard.

If the burst is designated as “LOW” by the burst status signal 810 then at the end of the data payload the modulation selector 813 is set to a logical Ό’ 504 such that the laser output is at the low level 301. Even if the last symbol in the burst data payload required a logical ‘1’ at the end of the burst, then the transition to a logical Ό’ can be effected with great speed by using the normal modulation circuitry 110. Again, this modulation optical value is held for a time period 502 long enough for the monitor channel to make an accurate measurement despite its limited bandwidth; but still short enough that there is time to fully extinguish the laser 101 within the period 404 required by the relevant communication standard. The monitor channel output 105 is then converted to digital form 821 and then passed at a suitable time instant to a second register 825 via a logical gate 823 enabled by the logical complement of the burst status signal 810. This register then provides the measured optical low value to the calculation function 826.

One convenient arrangement will be to designate the bursts as “HIGH” and “LOW’ in an alternating manner. However, the invention may also employ some other sequence of “HIGH” and “LOW” states where there may be a need to obtain an estimate of one level faster than the other, or to take account of some other requirements of the system, for example where noise is more significant at one of the levels.

Also during each data burst the output 105 of the monitor diode 102 and associated circuitry 103 is passed to an averaging function 804 that operates solely when commanded by a gating signal 508. This said averaging function can provide an average of the signal presented to its input over the time period enable by the gating signal 508, and hold the result when the gating signal indicates that the averaging should stop. Note that the end of the gating signal 508 ensures that the averaging function does not take account of the setting of the laser output to either a high or a low state during the duration of the extension of the data burst 504. The said averaging function may take account in each calculation of previous average values and may use weighting or decay rates that are defined to optimise the response time and the noise immunity, as would be common practice by persons skilled in the art when employing such functions. The output 805 of said averaging function 804 is passed to an ADC 806 and after the end of each data burst is converted to digital form and passed to a register 807. The timing of this conversion to digital form may be conveniently synchronised by the use of the control signal 801. The output of said register 807 is also passed to the calculation function 826.

The calculation function 826 then takes the estimates for the optical high and optical low and the optical average and also takes the required target value inputs for the average 131 and ER 132 and using a simple calculation derives the new bias current control value 113 and the new modulation current value 114. The calculation may take account of the optical high and optical low and the optical average using a number of scaling coefficients in the range from Ό’ to‘T for each input to the calculation depending on the achieved signal quality obtained from each channel in the practical application. The calculation is performed such that the errors between the calculated ER and average values and the corresponding required ER and average values are minimised and brought to a negligible or acceptable level. This process may take several iterations of “HIGH” and “LOW” bursts and several average operations and the precise rate of convergence of the system will depend on the scaling coefficients chosen for the said inputs and for other system variables in a particular application.

The numerical values for these scaling coefficients may be fixed or variable. For example the coefficients may be determined at the time of manufacture and testing and stored in the system. Alternatively, the user may determine the numerical values for the coefficients during testing or as a result of monitoring extended operation, and from these observations be able to optimise the values and then store them in the system. As another alternative, a controller function may be constructed with the capability to vary the coefficients while the system is in use in an adaptive manner using other performance information, possibly starting from some defined starting values.

Figure 9 shows an arrangement according to a second embodiment of the invention. In this arrangement, the calculation function 826 takes the estimates for the optical high and optical low and the optical average and also takes the required target value inputs for the average 131 and ER 132 and using a simple calculation derives the new average current control value 123 and the new bi-directional modulation current value 124. The calculation may similarly take account of the optical high and optical low and the optical average using a number of scaling coefficients in the range from Ό’ to ‘1’ for each input to the calculation depending on the achieved signal quality obtained from each channel in the practical application. The calculation is performed such that the errors between the calculated ER and average values and the corresponding required ER and average values are minimised and brought to a negligible or acceptable level. This process may take several iterations of “HIGH” and “LOW’ bursts several average operations and the precise rate of convergence of the system will depend on the scaling coefficients chosen for the said inputs and for other system variables in a particular application.

Figure 10 shows an embodiment of an arrangement that may be used within the system to provide digital information regarding the estimation of the optical high and optical low levels without employing a conventional ADC function. An analogue input 1007 is compared with some desired reference value 1003 in a comparator 1004. The output of said comparator may be gated in some logical function 1005 with a selection signal 1008 such that a counter 1006 is caused to increment or decrement when an enabling sampling clock signal 1009 is present. By these means, a digital output 1010 increases or decreases depending on the sign of the difference between the input signal 1007 and the reference value 1003. If this value 1010 is used in some closed loop system, such as the system presented in this specification, the value of the input signal 1007 will tend to approach and become equal to the reference value 1003. It may be convenient to use a DAC 1002 to convert a reference in digital form 1001 into an analogue reference 1003.

Figure 11 shows an embodiment of a simple means of obtaining a gated average value from an analogue input signal 1101. Said analogue input signal 1101 is passed to an analogue integrator comprising amplifier 1103, resistor 1105 and capacitor 1104 via a switching function 1107 under the control of a gating signal 803. While the switch 1107 is closed the output 1102 of the integrating amplifier 1103 will rise or fall depending on the instantaneous sign and magnitude of the input signal 1101. When the switch 1107 is opened under the control of the gating signal 803 the integration and averaging operation will cease and the value will be held. Any drift in the output value will be due to electrical imperfections in the components used. It may be desirable to have some decaying function for the averaging operation. A simple and convenient method illustrated is to connect a resistance 1106 across the capacitance 1104 by means of switch 1108 that allows the output to decay at some rate set by the relative values of these components.

Over a number of data bursts, the system will adjust the currents so that the errors are minimised, and hence the laser will be operating at substantially the desired average optical output and with substantially the desired ER.

Claims

1. A system for transmitting a sequence of at least two data bursts in a fibre optical communications system, the system comprising: selection circuitry configured to select one of a data input value, a logical high value or a logical low value such that the selection circuitry is configured to select the data input value during a data transmission period during a defined burst period and one of the logical high value and the logical low value during an extension time period during the defined burst period and immediately following the data transmission period, such that for the sequence of at least two data bursts, at least one data burst is a logical low value burst and at least one data burst is a logical high value burst; drive circuitry configured to apply a current to a laser diode, the current corresponding to the one of a data input value, a logical high value or a logical low value selected by the selection circuitry during the defined burst period or a zero value otherwise, the current being such that the laser diode is configured to provide an optical output; an optical sensor module configured to provide a sensor module output corresponding to the optical output of the laser diode; wherein the sensor module output is configured to provide electrical outputs proportional to the optical outputs of the laser diode corresponding to the logical high value and the logical low value in the sequence of bursts; and further configured to provide an output corresponding to an average value of the sensor module output during only the data transmission period during the sequence of bursts; and a controller configured to receive desired values regarding optical signalling output power levels of the laser diode and to receive the output from the optical sensor module proportional to the optical output of the laser diode corresponding to the logical high value and the logical low value and to receive the output corresponding to the average value of the sensor module output during only the data transmission period during the sequence of bursts; wherein the controller is configured to use a combination of the average and high and low values of the outputs from the optical sensor module wherein each value is scaled by a coefficient and the desired values to provide control values for the drive circuitry.

2. A system as claimed in claim 1, wherein the optical sensor module comprises a photodiode output power detector.

3. A system as claimed any preceding claim, wherein the optical sensor module comprises an optical sensor and a trans-impedance amplifier, the transimpedance amplifier being configured to provide the sensor module output.

4. A system as claimed in any preceding claim, wherein the control values are configured to control at least one of: an average power of the optical output of the laser diode; a power of the optical output of the laser diode representing a logical high; a power of the optical output of the laser diode representing a logical low; and a modulation index of the optical output of the laser diode.

5. A system as claimed in any preceding claim, wherein the current comprises a steady element and a variable element.

6. A system as claimed in any preceding claim, wherein the drive circuitry is configured to set the current applied to the laser diode dependent on a combination of a bias control value and a modulation control value.

7. A system as claimed in claim 6, wherein the control values are configured to control the drive circuitry to set the at least one of a bias current and a modulation current applied to the laser diode.

8. A system as claimed in any of claims 1 to 5, wherein the drive circuitry comprises bias circuitry configured to provide a bias current to the laser diode.

9. A system as claimed in any of claims 1 to 5, wherein the drive circuitry comprises modulation circuitry configured to provide a modulation current to the laser diode.

10. A system as claimed in any of claims 1 to 5, wherein the drive circuitry is configured to set the current applied to the laser diode dependent on a combination of an average value and a modulation value.

11. A system as claimed in any preceding claim, wherein the burst period is gated by a burst enable signal.

12. A system as claimed in any preceding claim wherein the control values control the drive circuitry to deliver the desired logical high and logical low optical output power levels.

13. A system as claimed in any preceding claim wherein the extension time period is greater than a settling time of the sensor module output.

14. A system as claimed in any preceding claim wherein the selection circuitry alternately selects one of the logical high value and logical low value for each consecutive extension time period.

15. A system as claimed in any of claims 1 to 13, wherein the selection circuitry selects the logical high value or the logical low value for each consecutive extension time period according to a pre-defined sequence.

16. A system as claimed in any of claims 1 to 13, wherein the selection circuitry selects the logical low value immediately after an extension time period where the logical high value has been selected.

17. A system as claimed in any preceding claim wherein the selection circuitry comprises a selector switch function.

18. A system as claimed in any preceding claim wherein a bandwidth of the selection circuitry is configured to be able to switch between the data input, the logical high value and the logical low value in a time significantly less than that of the extension time period.

19. A system as claimed in any preceding claim wherein the system comprises substantially digital circuits.

20. A system as claimed in any preceding claim wherein the control values for the drive circuitry are calculated by a digital calculation function.

21. A system as claimed in any of claims 1 to 18 wherein the system comprises substantially analogue circuits.

22. A method for transmitting a sequence of at least two data bursts in a fibre optical communications system, the method comprising: selecting one of a data input value, a logical high value or a logical low value wherein selecting comprises selecting the data input value during a data transmission period during a defined burst period and one of the logical high value and the logical low value during an extension time period during the defined burst period and immediately following the data transmission period, such that for the sequence of at least two data bursts, at least one data burst is a logical low value burst and at least one data burst is a logical high value burst; applying a current to a laser diode, the current corresponding to the one of a data input value, a logical high value or a logical low value selected during the defined burst period or a zero value otherwise, the current being such that the laser diode is configured to provide an optical output; providing an output corresponding to the optical output of the laser diode, wherein providing the output corresponding to the optical output of the laser diode comprises providing electrical outputs proportional to the optical outputs of the laser diode corresponding to the logical high value and the logical low value in the sequence of bursts and providing an output corresponding to an average value of output corresponding to the optical output of the laser diode during only the data transmission period during the sequence of bursts; receiving desired values regarding optical signalling output power levels of the laser diode; Combining the average and high and low values of the outputs corresponding to the optical outputs of the laser diode wherein each value is scaled by a coefficient and the desired values regarding optical signalling output power levels of the laser diode and using these combined values to provide control values to control a current applied to a laser diode.

23. A method as claimed in claim 22, further comprising applying the control values to control at least one of: an average power of the optical output of the laser diode; a power of the optical output of the laser diode representing a logical high; a power of the optical output of the laser diode representing a logical low; and a modulation index of the optical output of the laser diode.

24. A method as claimed in any of claims 22 and 23, wherein the current comprises a steady element and a variable element.

25. A method as claimed in any of claims 22 to 24, further comprising setting the current applied to the laser diode dependent on a combination of a bias control value and a modulation control value.

26. A method as claimed in claim 25, wherein setting the current applied to the laser diode comprises setting at least one of a bias current and a modulation current applied to the laser diode based on the bias control value and modulation control value.

27. A method as claimed in any of claims 22 to 24, wherein applying the current further comprises providing a bias current to the laser diode.

28. A method as claimed in any of claims 22 to 24, wherein applying the current further comprises providing a modulation current to the laser diode.

29. A method as claimed in any of claims 22 to 24, wherein setting the current applied to the laser diode comprises setting the current dependent on a combination of an average value and a modulation value.

30. A method as claimed in any of claims 22 to 29, wherein the burst period is gated by a burst enable signal.

31. A method as claimed in any of claims 22 to 30, further comprising applying the control values to deliver the desired logical high and logical low optical output power levels.

32 A method as claimed in any of claims 22 to 31, wherein the extension time period is greater than a settling time of providing the output.

33. A method as claimed in any of claims 22 to 32, wherein selecting one of a data input value, a logical high value or a logical low value comprises alternately selecting one of the logical high value and logical low value for each consecutive extension time period.

34. A method as claimed in any of claims 22 to 32, wherein the selecting one of a data input value, a logical high value or a logical low value comprises selecting the logical high value or the logical low value for each consecutive extension time period according to a pre-defined sequence.

35. A method as claimed in any of claims 22 to 32, wherein the selecting one of a data input value, a logical high value or a logical low value comprises selecting the logical low value immediately after an extension time period where the logical high value has been selected.

36. A method as claimed in any of claims 22 to 35, wherein the selecting one of a data input value, a logical high value or a logical low value comprises selecting based on selector switch function.

37. A method as claimed in any of claims 22 to 36, wherein selecting one of a data input value, a logical high value or a logical low value comprises switching between the data input, the logical high value and the logical low value in a time significantly less than that of the extension time period.