Optical Amplifier Circuit
The present invention relates to optical communication systems, and in particular to the control of optical power levels output from components within optical communication systems .
Optical communication systems utilise ever increasing data rates and require increasing functionality. Each of the additional functional blocks, such as optical switches, optical routing components, optical wavelength converters, optical regenerators or optical signal processing elements can degrade the power level of an optical signal and adversely affect the signal- to-noise ratio of the optical signal. Each of the aforementioned functional blocks needs to amplify the optical signal as it passes through the device. Various types of optical amplifiers, such as fibre amplifiers or semiconductor optical amplifiers (SOAs) , are typically utilised to perform this task.
Optical amplifiers ideally have an optical output power that can be controlled in several ways. Firstly, the optimum optical output power level is achieved when the difference between the zero level logical state output power and one level logical state output power is maximized. This is desirable because it increases the likelihood of the signal being successfully received at a subsequent optical component.
Secondly, control of optical output power on short (for example, nanosecond) time-scales is required because optical signals are grouped into packets with durations of hundreds of nanoseconds and the guard bands separating the packets have durations in the order of nanoseconds. The optical power level of consecutive
packets can vary considerably, where they originate from different transmitters and have propagated through the communication system with differing losses prior to aggregation. The relatively long response time of fibre amplifiers makes them unsuitable for use in components in packet-switched communication systems. In contrast, the response time of SOAs is in the order of nanoseconds .
Thirdly, it is desirable to control the one level logical state output power of the component so that it is at the maximum tolerable level and also at a constant level, irrespective of the optical input signal to the component. This aspect of the power control ensures that a high signal-to-noise ratio is achieved.
Ideally, optical output power can be controlled irrespective of the state of polarisation of the optical input signal. Furthermore, in order to achieve an improved overall performance in a communications system, it is preferable to implement optical output power control at numerous different optical components (which may be multisignal, multiple fibre routing components) within the communication system. Thus, a means of controlling optical output power which requires a minimum of components and can be utilised in a variety of optical components is preferable.
Figure 1 shows a simplified cross-section through a conventional SOA 10 which comprises an active guide layer 12 positioned between an n-doped substrate 14 below and a p-doped substrate 16 above. The active guide layer 12 is comprised of an undoped quaternary layer 18. A strongly p-doped ternary contact layer 24 is located above the p-doped substrate 16, and a top
electrode 26 and a bottom electrode 28 form the outermost SOA surfaces. An input facet 30 and an output facet 32 are usually produced by cleaving a semiconductor wafer which incorporates a plurality of SOAs, and then the facets 30, 32 receive an anti- reflection treatment.
In operation, an electrical pump current I is passed between the top and bottom electrodes 26 28, in order to inject charge carriers of both types into the active guide layer 12. This application of the electrical pump current I serves to provide the device with gain and the amount of gain is determined by the type of semiconductor material utilised in the device. An optical input signal is incident on the input face 30 and an amplified optical signal emerges from the output face 32.
A conventional technique for optical output power control in SOAs operates by monitoring electrical signals at the multiple electrodes of an SOA during a period when an optical signal is passing through it. Specifically, when a change occurs in the optical input signal to the SOA (generally resulting from the modulation of the optical input signal ie . a data carrying signal) , a change in the charge carrier density is induced. This causes alteration in the junction voltage or in the electrical pump current or both. Therefore, this fluctuation in the junction voltage or the electrical pump current is representative of the modulation in the optical input signal and can be used in both the amplification and detection of the input optical signal. For example, see IEEE Journal of Lightwave Technology 8(4) 610 1990.
A variation on the multiple electrode technique enables optical output power and SOA gain to be estimated. The modulated electrical signal at the SOA electrodes are monitored and compared (see, for example, IEEE Photonics Technology Letters 4 (11) 1258 1992) .
However, such multiple electrode techniques increase the complexity of an optical component and therefore, of the communication system.
It is also known to control optical output power through the use of a low frequency pilot signal to modulate the optical input signal. In fibre amplifiers this technique must be used in conjunction with external photodiodes in order to appropriately adjust gain (see European Patent No. 0637 148) . In SOAs this technique does not require external photodiodes because the SOA electrode voltage or current variation can directly represent the required optical input signal including the pilot signal . The pilot signal is isolated by use of a bias tee to separate the pilot signal from the high frequency data modulation of the optical input signal represented in the derived electrical signal . The derived electrical signal can be processed to adjust the electrical pump current in order to control the SOA gain and optical output power (see, for example, Japanese Patent No. 4 107 429) . However, such pilot signal techniques are incompatible with high operational speeds required by packet switched communication systems, and increasing the frequency of the pilot signal would lead to optical signal distortion.
It is, therefore, desirable to provide a system for controlling optical power levels output from optical signal amplification elements of components within an
optical communication system in which the problems associated with the prior art are at least mitigated.
According to a first aspect of the present invention, there is provided an optical amplifier device comprising an optical amplifier for amplifying a modulated light signal, and a control circuit, wherein the control circuit comprises; means for supplying a pump current to the optical amplifier, a detection means for detecting modulation of an optical amplifier electrode signal caused by modulation of the light signal at a data rate thereof, a measuring means for measuring a parameter of the modulated electrode signal within a predetermined range of the frequency spectrum of the modulated electrode signal, and a feedback means which utilises the parameter measurement to control a pump current to the optical amplifier.
According to second aspect of the present invention, there is provided a method for amplification of an optical signal comprising the steps of; supplying a pump current to an optical amplifier; detecting modulation of an optical amplifier electrode signal caused by modulation of the optical signal at a data rate thereof, measuring a parameter of the modulated electrode signal within a predetermined range of the frequency spectrum of the modulated electrode signal, and utilising the parameter measurement to control the pump current to the optical amplifier.
For a better understanding of the present invention, and to show how it may be put into effect, reference will now be made, by way of example, to the accompanying drawings in which:
Figure 1 illustrates a cross-sectional view through a conventional SOA; and
Figure 2 is a schematic representation of an embodiment of the present invention.
Figure 2 illustrates an optical output power control circuit 34 for an SOA 36 (such as that illustrated in Figure 1) . An input optical fibre 38 is coupled to a SOA input facet 40 and an output optical fibre 42 is coupled to a SOA output facet 44. A control circuit 74 comprises a electrical coupler 46, an electronic signal detection and feedback control block 50, an electrical current source 58, and a SOA state control block 62.
The SOA state control block 62 decides if the SOA should be turned on by sending a control signal via a SOA state control block output 64 to a connection 60 of the electrical current source 58 which enables a pre- set DC current to flow from an electrical current source output 66 to an input 68 of the electrical coupler 46. The DC current is then output from the connection 70 of the electrical coupler 46 and injected into an SOA electrode 72.
In operation, a data modulated optical input signal in the input optical fibre 38 is received at the SOA input facet 40. When the DC current is injected into the SOA 36, the data modulated optical input signal induces a change in the charge carrier density in the active layer of the SOA 36, which in turn causes an alteration in the electrode voltage (or current) that is representative of the input data signal . It has been found that the change in amplitude of the electrode
voltage (or current) is proportional to the optical output power of the SOA 36. The relationship is independent of average electrical pump current, optical input power, and the state of polarisation of the optical input signal. Therefore, the electrode voltage (or current) change derived on the SOA electrode 72 can be used to control the output power of the SOA 36.
The electrical coupler 46 separates the electrode voltage (or current) change from the input optical signal, and outputs a derived voltage (or current) change via connection 48. An input 52 to the electronic signal detection and feedback control block 50 is connected to an output 48 of the electrical coupler 46. The electronic signal detection and feedback control block 50 detects at .least one of three parameters of the electrode signal within a predetermined spectral frequency range. The three parameters are the voltage amplitude of the electrode signal, the current amplitude of the electrode signal and the power of the electrode signal.
The electronic signal detection and feedback control block 50 filters the electrode signal to extract the wanted spectral range (and electronic amplification occurs if the signal is weak) . The signal is then passed through a suitable rectifier. Where the signal's voltage or current parameter is being detected a linear rectifier should be used, and where the signal's power is being detected a square law rectifier is used. The rectifiers are standard RF devices mainly consisting of diodes and capacitors. A control signal
is produced at an output 54 of the electronic signal detection and feedback control block 50.
The output 54 is connected to the input 56 of the current source block 58. The current source block 58 adjusts the DC current output at connection 66 according to the control signal, until the average output optical power stabilises at a desired level. Therefore, the output optical signal in the output optical fibre 42 is a controlled and amplified version of the input optical signal.
The electronic signal detection and feedback control block 50 includes an electronic signal detector and a feedback control unit. In practice, the electronic signal detector can be a radio frequency spectrum analyser or a wide-band diode detector. The feedback control unit can be implemented by digital signal processors (DSPs) to achieve a fast response time. In this way, output optical power control can be realised rapidly after the SOA 36 is turned on or the input optical signal appears at the SOA input facet 40.
The spectral frequency range within which the electronic signal detection and feedback control block 50 operates is primarily determined by the reaction speed required for the control loop. Theoretically, the block 50 must detect a component at a frequency of at least approximately 1/ (2πτ) , where τ is the required response time of the control loop. For example, if the control loop needs to react in less than 10 nanoseconds, the highest detected signal frequency is;
1/ ( 2πxl0~8) -16MHz
In practice, this frequency would be made higher, in particular when the signal format has limited energy below the calculated frequency.
Theoretically, there is no limitation on the lowest frequency component to be detected. However, many signal formats have very low energy levels in the lower frequency range and, in practice, only noise can be detected in this range. Therefore, the spectral frequency range can be adjusted to ensure sufficient signal power to enable effective detection.
Advantageously, the control circuit of the present invention exhibits a wide dynamic range and can function with various data modulation formats, whilst maintaining a good signal to noise ratio. Furthermore, the power control circuit functions at sufficiently high speed to allow for power equalisation and power optimisation in packet-switched communication systems. The circuit of the present invention avoids use of additional optical or opto-electronic components or pilot signals.
It will be apparent to the skilled person that the above described circuit architecture is not exhaustive and variations on this structure may be employed to achieve a similar result whilst employing the same inventive concept. For example, the electronic signal detector can measure at least one of the electrical power, current and voltage of the modulated electrical signal falling within the predetermined spectral range.
In a communication system utilising multiple optical signals (possibly of varying wavelengths) and multiple optical fibres, numerous optical power control circuits can be utilised in an array or matrix configuration.
It can therefore be seen that the present invention provides an optical amplifier control circuit which has significant advantages over conventional devices.