CA1101923A

CA1101923A - Injection laser operation

Info

Publication number: CA1101923A
Application number: CA311,758A
Authority: CA
Inventors: Joseph Straus
Original assignee: Northern Telecom Ltd
Current assignee: Nortel Networks Ltd
Priority date: 1978-09-21
Filing date: 1978-09-21
Publication date: 1981-05-26

Abstract

IMPROVEMENTS IN INJECTION LASER OPERATION
Abstract of the Disclosure Compensation for variation in lasing threshold caused by temperature change, laser aging and other degrading phenomena is based on utilization of the rectifying property of semiconductor lasers near threshold. A laser is modulated by a small sinusoidal test signal. According to the monitored position of a critical point of a preselected generated optical harmonic of the test signal fundamental frequency, laser bias current is adjusted to restore the laser threshold position;
in addition laser output at the fundamental frequency is used to preserve constant power output if the slope efficiency of the laser alters.

- i -

Description

This invention relates to a method and apparatus for operating an injection laser so as to compensate for variation of laser threshold caused by temperature change, laser aging and other degrading phenomena.
Semiconductor lasers for use in fibre optic communication systems typically show a variation of light output with applied current having a current threshold above which lasing is produced and below which a minimal amount of spontaneous emission occurs. Such lasers are typically operated by applying a constant d.c. bias at the lasing threshold level and applying a modulating current to obtain an output representative of a signal to be transmitted.
There are several reasons why it is important to stabilize the laser threshold against temperature variation.
Firstly, an increase in threshold current causes the laser output at fixed d.c. bias current to decrease and turn-on delay to increase, leading to an increase in intersymbol interference in a digital system. In an analogue system variation in slope efficiency due to device aging degrades signal to noise ratio.
Another reason for adjusting the laser bias to threshold is to keep the spectral width of the laser emission at a minimum, consistent with system requirements since maximum bit rate capacity of data transmission through dielectric optical wave-guides is reduced if spectral broadening occurs because of material dispersion in the waveguides.
Several possible bias stabilization techniques may be found in an article by R.E. Epworth, entitled "Subsystems For High Speed Optical Links", Proceedings of the second European Conference on Optical Fibre Communication, pp 377-382, Paris, France, 1976. In the earlier techniques, laser light output level ~k ~

:: :

is monitored and compared to a preset reference level to provide an error signal for subsequent generation of a feedback signal. A
disadvantage of some of the more feasible biasing schemes is that they require wide bandwidth components in the feedback path.
An alternative compensation method is described by A. Albanese in an article entitled "Automatic Bias Control (ABC) for injection lasers, Digest of Technical Papers at Topical Meeting on Optical Fibres Transmission II, pp WB5-1, WB5-4, Feb. 22-24, 1977, Williamsburg, Virginia. The method is based on voltage saturation principles of lasers at threshold. An advantage of the automatic biasing circuit described is that it does not require optical detection to stabilize the laser bias. In practice, it is considered that a loss of voltage saturation can occur in lasers with aging or degradation. A partial loss of saturation always occurs in non-ideal devices. These facts to a certain degree then limit the dynamic range of the compensating method described.
A technique is now proposed based on the rectifying property of semiconductor lasers when operated under AC modulation near threshold.
According to one aspect of the invention a method of operating an injection laser comprises applying to the laser a modulating data signal, a d.c. bias current and a modulating test signal having a fundamental frequency, monitoring the level of a predetermined harmonic of the fundamental frequency generated by the laser and, according to said monitored level, adjusting the d.c. bias current so that the level of said harmonic is maintained at a critical point. The critical point may be an extrema or zero depending on the nature of the predetermined harmonic.

- 2 -In an operational system, the harmonic can be monitored by detecting and analyzing light from one facet of the inject laser, light from another facet coupled into a dielectric optical waveguide of the system. Alternatively, and especially if only one laser facet is accessible a coupler, for example, a 20dB coupler can be so inserted as to extract a proportion of light before it is launched into a dielectric optical waveguide. For an analogue modulating data signal, both the fundamental frequency and the frequency of said predetermined harmonic are chosen to be outside, preferably below the frequency band of the analogue modulating data signal and in normal operation the d.c. bias current is maintained close to and greater than a threshold current at which the laser lases. On the other hand if the modulating data signal is a digital signal, than the test modulating signal is inhibited while pulses of the digital signal are applied to the laser and, in normal operation, the d.c. bias current is maintained close to and lower than a threshold current at which the laser lases.
The laser output component at the fundamental frequency can also be utilized to compensate for variation in device slope efficiency caused by device aging. The level of this component is monitored and compared against a preset level, and according to the difference between said monitored level and said preset level, the level of the applied modulating current is adjusted to ensure normalized optical output levels.
According to another aspect of the invention apparatus for operating an injection laser as described previously comprises a signal means for applying to the laser a modulating data signal, a d.c. bias current, and a modulating test signal having a fundamental frequency, monitoring means for monitoring the level of a predetermined harmonic of the fundamental frequency generated by the laser, and feedback means under the control of said monitoring means for adjusting the d.c. bias current so that the level of said harmonic is maintained at a critical point.
The apparatus can include a photodiode arranged to receive light from one facet of the laser, a filter network arranged to pass a narrow band of frequencies, including said predetermined harmonic frequency, from the photodiode output and analyzing means for analyzing the content of the passed band of frequencies.
The analyzing means can include a slope detector to detect variation from extrema in a light output-frequency plot of said predetermined harmonic and a zero crossing detector to detect a phase change in the plot.
The apparatus can include a bias control network having an input from said analyzing means and an output to a d.c.
bias current genrating network whose output forms one input to a laser driver circuit. Specifically for an analogue system the output of an adder at which an analogue data modulating signal and ; 20 the test modulating signal are summed, can form a second input to the laser driver circuit. Alternatively, and specifically for a r digital system, the apparatus can further include switch means to inhibit application of the modulating test signal when a pulse is present in an applied digital data modulating signal, an output from the switch means forming the second input to the laser driver circuit.
For irnproving laser slope characteristic as described previously, the apparatus can further include a second feedback means having a low pass filter for filtering the fundamental frequency component generated by the laser, level 92~

detector means for comparing the level of said component against a preset level, control means for generating an error signal corresponding to the difference in the two levels and signal input means to which said error signal and data signal are applied for modifying said data signal precedent to the data signal being applied to a laser driver circuit whereby to maintain the laser output constant for a predetermined level of the applied data signal.
An embodiment of the invention will now be described by way of example with reference to the accompanying drawings, in which:-Figure 1 is a graphical representation of thevariation of lasing light output with applied current for a typical injection laser showing in addition, the effects of temperature variation and laser aging;
Figure 2 is a block schematic diagram of apparatus for operating an injection laser specifically for analogue systems;
Figure 3 is a block schematic diagram of apparatus for operating an injection laser specifically for digital systems;
Figure 4 shows graphically the manner in which the third harmonic of the fundamental frequency depends on modulation level; and Figure 5 is a graphical representation of power observed in the third harmonic when a fundamental frequency is applied to an injection laser; also shown is the effect of a temperature change.
Referring in detail to Figure 1, there is illustrated the combined effects of aging and temperature variation on the light output - applied current characteristics of a semiconductor injection laser. Significantly temperature variation changes the threshold current which need be applied before the device lases, while the lasing efficiency of the device, which is measured by the slope of zones A of the graphical plot, is reduced by aging.
As explained previously, variation in the threshold position is undesirable since it increases the turn-on delay which is a measure of the extent to which an applied data modulating signal is used in bringing the device to threshold before lasing proper is instituted to transmit intelligence contained in the data modulating signal. As a corollary output power is decreased if the initial data modulating signal amplitude is maintained. In addition variation in slope efficiency is undesirable since it produces a variable output level in response to a data modulating signal having a constant input level.
Practical circuitry for compensating for laser degradation is shown in Figures 2 and 3, the circuit of Figure 2 being particularly adapted for analogue data modulating signals while that of Figure

3 is particularly adapted for use in a digital system.
However before these circuits are described in detail an outline of the theory involved will be given.
It can be shown that neglecting the contribution of spontaneous emission to light output and considering the ideal case of a linear laser response, the L-I characteristic of the laser can be written as:
L = ~ Ith) I > Ith L = 0 I < Ith where n denotes the current to light conversion efficiency, L
is the light output, I is the applied current and Ith is the applied current at lasing threshold.

Under ac modulation of a type:

tot dc Ip sin ~t the light output near Ith will be rectified with Fourier spectrum given below:

n ~ [aO al sin x + ~ ~a2k(a) cos 2k x + a2k+1(~)cos(2k+1)x}]

Eq. ..... (1) with aO = m cos xl + (a~ -2xl) al = m (z -xl- sin22xl) +2(~-l)cos x a2k = m ~COs(2k+l)xl cos(2k-l)xl ~ + 2(1-a)sin 2kx \ (2k+1) (2k-1) ¦ 2k 2k 1 ( 2(k~ ) 2k ) 2k+1 and x ~ arcsin (1~~ ) n = IP = Idc The central idea behind the bias stabilization scheme of the invention is the fact that selected characteristic or critical points of various harmonics of an ac test signal can serve as reference bias points. Such points, for instance, could be extrema or zeros of given harmonics. As an illustration, it can be seen in (1) that the amplitudes of all harmonics at ~ m (IdC = Ith - Ip) are equal to zero. In addition, 2~

because of symmetry, odd harmonics at ~ = 1, i.e., at IdC = Ith are equal to zero, while the amplitudes of even harmonics at this bias point are maximum.
Further characteristic points can be obtained by solving the following transcendental equations for ~:

d~ [a2k+1 (~)~ = _ ~a2k (~)~ = 0 k = 1, 2, 3 Since ~ = IdC/Ith is a function of dc bias it follows that for a given threshold current, Ith the locations of certain characteristic points of a given harmonic such as the third harmonic illustrated in Figure 5 can be related uniquely to a corresponding dc bias. As illustrated in Figure 4 for the third harmonic, the locations of the characteristic points also depend on the modulation level amplitude m = Ip/Ith. Taking account of these interdependencies, feedback circuits as shown in Figures 2 and 3 can be constructed to stabilize laser operation at a chosen point or relative to threshold.
A change in the threshold current at a fixed modulation level is reflected as shown in Figure 5 for the third harmonic in a shift of the reference bias point from the preset level. The feedback circuits of Figures 2 and 3 minimize these excursions from equilibrium. The presence of a fundamental frequency ; of the test modulating signal can also be used to provide stablization against slope efficiency variations of the laser as will be described with reference to Figures 2 and 3.
Because of symmetry the reference bias point can be set up either below or above the lasing threshold. Biasing at the threshold is not recommended because the noise current is at a maximum. For analog systems (Figure 2) when linear laser - .

2~

operation is required, the bias point must be above the threshold.
In digital systems (Figure 3) however, it is pre~erable to bias the laser just below the threshold to reduce the contribution of shot noise to total system noise. In choosing a bias point below the threshold however, intersymbol interference arising from turn-on delays must be avoided.
Referring in detail to Figure 2, a test signal having a fundamental frequency and a suitable amplitude is combined with a signal to be transmitted in an adder 3. The amplitude of the test signal is chosen to modulate a laser 5 for selected test signal excursions below the threshold. The frequency of the test signal together with frequencies of the major harmonics, are arranged to be below the frequency band of data to be transmitted. The output of the adder 3 is coupled to a laser driver 4 which drives injection laser 5. The laser is biased above threshold at a suitable preselected dc bias level generated by means of a biasing network 7. The light emitted from one facet of the laser S is coupled into a dielectric optical waveguicle 8. Light from the other facet is detected by a photodiode 9 and amplified by an amplifier 10. The output of the amplifier 10 is passed through a filter network 11 having outputs at both the fundamental frequency and at the selected harmonic, say the third, of the test signal. The third harmonic is further analyzed in a harmonic detection network 14 for either slope of phase reversals. This analysis is accomplished with the aid of either a slope or zero crossing detector depending on the particular characteristic or critical point selected in the third harmonic.
As was discussed previously, changes in threshold current to either side of the initial value are reflected in changes of location of these characteristic or critical points. This in turn produces either slope or phase reversals in the content of the third harmonic as shown in Figure 5. The output of the detection network 14 thus provides an error signal representing the change in location of the critical points, the error signal being fed to the biasing network 7 which sets the dc bias current relative to a new laser threshold.
For simultaneously normalizing the lasing output, the fundamental frequency component of the test signal generated by the laser is rectified in a rms converter 17 and compared with a preset reference level by a comparator 19. An output from the comparator forms an input to a signal control stage 20 which controls an input stage 21 which receives the incoming data signal.
Thus, in response to variation of the laser slope efficiency indicated in section A of the graphical representations of Figure 1, the modulating current is altered to normalize the laser output level. It will be apparent that the two feedback paths are interactive and therefore priority and compensation sub-systems must be incorporated. Such sub-systems are not described here since they are not material to the central invention and will readily be deductible by those skilled in the semiconductor laser drive circuitry art.
Turning now to Figure 3, the circuit shown in this block schematic diagram is particularly adapted for use in digital signal transmission systems. In a normal operating mode, the laser is biased slightly below threshold at a selected characteristic reference bias level in order to reduce system noise and to improve laser reliability. Drive pulses containing the information to be transmitted then drive the laser above threshold. In many respects the circuit resembles that of Figure 2 and as in the analogue system bias control is performed with a low frequency ~ `:
sinusoidal test signal. However, the system has a switching network 3a so that the test signal is gated off when an input pulse is present, such gating prevents any interference between the incoming data stream and the test signal.
From the bias stabilization point of view, the gating is equivalent to test signal sampling. According to sampling theory, the original frequency spectrum of the sample signal can be recovered by low pass filtering providing the sampling frequency is greater than twice the information bandwidth.
In practice, such a recovery, in case of a test signal for bias stabilization, can always be assured since the bit rate of digital data transmission is expected to be considerably higher than the test signal frequency. The penalty for such a stabilization scheme in digital systems is an increased transmitter circuit complexity as well as a slightly reduced detection margin between l-s and 0-s. Assuming a 200mA peak drive pulse for the digital signal and a test signal of a maximum 5mA p-p amplitude, the reduction in the detection margin in the worst case situation is equal to 1.25%.
In both the Figure 2 and 3 embodiments a test signal fundamental frequency of 0.8 MHz is proposed, although the frequency chosen in any practical system will depend on the data signal bandwidth or bit rate as the case may be.

Claims

THE EMBODIMENTS OF THE INVENTION IN WHICH AN
EXCLUSIVE PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:-

1. A method of operating an injection laser comprising applying to the laser a modulating data signal, a d.c.
bias current and a modulating test signal having a fundamental frequency, monitoring the level of a predetermined harmonic of the fundamental frequency generated by the laser and, according to the monitored level of said predetermined harmonic, adjusting the d c bias current so that the level of said harmonic is maintained at a critical point.

2. A method as claimed in claim 1, in which the harmonic is monitored by detecting and analyzing light from one facet of the laser and light from another facet of the laser is adapted for coupling into a dielectric optical waveguide.

3. A method as claimed in claim 2, in which the detected light is converted into an electrical signal, the electrical signal being passed through a filter having a passband including said predetermined harmonic.

4. A method as claimed in claim 3, in which the filter output is analyzed to determine whether the level of said harmonic occupies a critical point.

5. A method as claimed in claim 4, in which the analysed filter output is used to generate a feedback signal to modify the applied d.c. bias current, thereby to restore the level of said harmonic to said critical point.

6. A method as claimed in claim 2, in which said modulating data signal is an analogue signal and both the fundamental frequency and the frequency of said predetermined harmonic are outside the frequency band of the analogue modulating data signal.

7. A method as claimed in claim 6, in which the d.c. bias current is maintained close to and greater than a threshold current at which the laser lases.

8. A method as claimed in claim 2, in which said modulating data signal is a digital signal and said test modulating signal is inhibited while pulses of the digital signal are being applied to the laser.

9. A method as claimed in claim 8, in which the d.c. bias current is maintained close to and less than a threshold current at which the laser lases.

10. A method as claimed in claim 1, further comprising monitoring the level of the fundamental frequency generated by the laser, comparing the monitored level with a preset level, and according to the difference between said monitored level and said preset level, adjusting the level of the applied modulating data signal whereby to maintain the laser light output substantially constant for a predetermined modulating data signal input level.

11. Apparatus for operating an injection laser, said apparatus comprising:-signal means for applying to the laser a modulating data signal, a dc bias current and, a modulating test signal having a fundamental frequency;
monitoring means for monitoring the level of a predetermined harmonic of the fundamental frequency generated by the laser; and feedback means under the control of said monitoring means for adjusting the dc bias current so that the level of said harmonic is maintained at a critical point.

12. Apparatus as claimed in claim 11, in which said monitoring means includes a photodiode arranged to receive light from one facet of the laser.

13. Apparatus as claimed in claim 12, in which said monitoring means further includes a filter network arranged to pass a narrow band of frequencies including said predetermined harmonic frequency, from an output from the photodiode.

14. Apparatus as claimed in claim 13, in which the monitoring means further includes analyzing means for analyzing the content of said passed band of frequencies.

15. Apparatus as claimed in claim 14, in which said analyzing means includes a slope detector to detect variation from extrema in a light output-frequency plot centered on said predetermined harmonic.

16. Apparatus as claimed in claim 14, in which said analyzing means includes a zero crossing detector to detect a phase change in a light output-frequency plot centered on said predetermined harmonic.

17. Apparatus as claimed in claim 11, in which said feedback means includes a bias control network having an input from said analyzing means and an input from said analyzing means and an output to a network for generating said dc bias current.

18. Apparatus as claimed in claim 17, specifically adapted for use in an analogue system, said signal means including an adder stage at which an analogue modulating data signal and the modulating test signal are combined to provide an output .

19. Apparatus as claimed in claim 18, said signal means further including a laser driver stage at which the outputs from said adder stage and the dc bias current generating network are summed, said laser driver stage having an output directly controlling the laser.

20. Apparatus as claimed in claim 17, specifically adapted for use in a digital system, said signal means including a switching stage to inhibit application to the laser of the modulating test signal when a pulse is present in an applied digital data modulating signal.

21. Apparatus as claimed in claim 20, said signal means further including a laser driver stage at which the outputs from said switching stage and the dc bias generating network are summed, said laser driver stage having an output directly controlling the laser.

22. Apparatus as claimed in claim 11, further comprising a second feedback means having a low pass filter for filtering the fundamental frequency component generated by the laser, level detector means for comparing the level of said component against a preset level, control means for generating an error signal corresponding to the difference in the two levels and signal input means to which said error signal and data signal are applied for modifying said data signal precedent to the data signal being applied to a laser driver circuit whereby to maintain the laser output constant for a predetermined level of the applied data signal.