OPTICAL FREE-SPACE SIGNALLING SYSTEM
This invention relates to a signalling system and components thereof. In particular, this invention relates to a signalling system in which data is conveyed by modulating a free-space light beam.
International Patent Application WO 98/35328, which is incorporated herein by reference, describes a point-to- multipoint communication system utilising free-space light beams. In particular, WO 98/35328 describes a system in which a plurality of user stations emit light beams which are directed at a local distribution node. At the local distribution node, each of the incoming light beams is directed by a telecentric lens to a respective modulator element of an integrated array of modulator elements substantially located in the back focal plane of the telecentric' lens. The reflectivity of each modulator element is individually variable in accordance with a respective data signal and each incoming light beam is modulated and reflected back, via the telecentric lens, to the user station from which the light beam originated (i.e. the incoming light beam is retro-reflected). At the user station, the modulated light beam is detected and the corresponding data signal is regenerated.
The local distribution node described in WO 98/35328 is able to communicate with user stations within a horizontal field of view which is greater than +/-500 and
a vertical field of view of approximately +/-5°.
A problem with the signalling system described in WO 98/35328 is that the local distribution node is relatively complex and is expensive to manufacture. Therefore, if there are only a small number of user stations communicating with a local distribution node, the cost per user station is high.
According to a first aspect of the invention, there is provided a free-space optical signalling system for signalling between a distribution node and a plurality of remote terminals, the distribution node comprising a plurality of modulator units and a support operable to support the plurality of modulator units at respective different orientations. Each modulator unit comprises a respective housing which supports a retro-reflector and a modulator which are operable to modulate an incoming light beam from one of the remote terminals so that a modulated light beam is reflected back to the remote terminal. By using a plurality of separate modulator units in this way, each modulator unit is able to have a smaller field of view than the field of view of the local distribution node described in WO 98/35328. This reduction in the required field of view allows the modulator unit to have a relatively large collection aperture at reasonable expense. An increase in the collection aperture is advantageous because a larger proportion of the incoming light beam is collected by the modulator unit and therefore the power of the modulated light beam reflected back to the remote terminal is
increased.
Using separate modulator units also improves the flexibility of the signalling system because if the number of remote terminals is varied, then the number of modulator units can be correspondingly varied. If, therefore, only a small number of remote terminals are present, only a small number of modulator units need be used and therefore the cost of the distribution node can be kept comparatively low. This is particularly advantageous if the small number of remote terminals are dispersed within a 360° field of view because this would require a plurality of the local distribution node units described in WO 98/35328.
Preferably, the retro-reflector is formed by a reflector located substantially in a plane perpendicular to the optical axis of a telecentric lens . An advantage of such a retro-reflector is that the divergence of the reflected optical beam is adjustable by changing the relative positions of the telecentric lens and the reflector. This is advantageous because, in practice, the remote terminal does not generate a perfectly collimated optical beam and therefore, if no adjustment of the divergence is performed, the reflected light beam received by the remote terminal is generally significantly larger than the emitted light beam, leading to a low signal level being detected at the remote terminal. By moving the telecentric lens and the reflector relative to each other, the beam divergence can be reduced so that the size of the reflected light beam at the remote terminal
is reduced and the signal level is correspondingly increased.
The inventors have found that by making the field of view of the modulator unit a solid angle in the range of 0.2 x 10~3 to 5 x 10~3 steradians, the field of view is sufficiently large that a precise alignment of the modulator unit with the remote terminal is not required, while a telecentric lens with a large collection aperture is relatively cheap to manufacture. In other words, although the light beam emitted by the remote terminal must be precisely aligned on a modulator unit, the modulator unit need only be roughly aligned with the remote terminal. This is a significant advantage in comparison with a signalling system employing two optical units at respective ends of an optical link with each optical unit having a light source, because the light beam emitted by each optical unit must be precisely aligned with the other optical unit and therefore two precise alignments must be performed.
In an embodiment, the field of view of the modulator unit defines a cone with a half-angle of 2.5°. If the remote terminal is 100 metres away, then this field of view will correspond to a circle having a radius of approximately 5 metres at the remote terminal.
Preferably, the reflector and the modulator are formed by a common device including a quantum confined Stark effect (QCSE) device whose absorptivity varies in dependence upon an applied electric field. Advantages
of using a QCSE device include compactness and the ability to vary the absorptivity at high frequencies using conventional logic signals, for example HCMOS logic signals, as drive signals. Generally, a QCSE device is in the form of a p-i-n semiconductor diode with a plurality of quantum wells formed in the intrinsic region. When an electric voltage is applied across the quantum wells, the resulting electric field determines the optical absorption spectrum of the quantum wells.
As the size of the active area of the QCSE device (i.e. the area over which an incident optical beam is modulated and reflected) increases, a problem arises that, due to the resistivity of the p-doped and n-doped regions, the effective series resistance of the QCSE device increases. This increase in the effective series resistance reduces the maximum switching speed of the electric field across the multiple quantum wells. Further, at high switching frequencies the electric field across the multiple quantum wells is not uniform which causes the level of modulation to vary in dependence upon the position of incidence of a light beam within the active area.
According to a second aspect of the invention, there is provided a quantum confined Stark effect device having a face operable to transmit an incident light beam, and a plurality of quantum wells which absorb a portion of the incident light beam, the amount of the incident light beam absorbed varying in dependence upon an electric field applied across the plurality of quantum wells using first and second electrodes. One of the first and second
electrodes is provided adjacent the transmitting face and comprises a plurality of strip conductors which at least partially cross the face. In this way, the effective series resistance of the QCSE device is reduced leading to an increase in the maximum switching speed of the electric field applied across the plurality of quantum wells, and a more uniform electric field is achieved at high frequencies .
In an embodiment, the QCSE device is formed by a mesa- structure projecting from a substrate, with the first electrode being formed around the base of the mesa- structure and the second electrode being formed at the top surface of the mesa-structure, the strip conductors forming part of the second electrode and extending over the top surface of the mesa-structure. The first electrode makes an edge connection with the QCSE device and therefore current flowing through the centre of the mesa-structure must travel out to the edge of the mesa- structure to reach the first electrode. Preferably, the p-doped layer is provided adjacent the second electrode at the top of the mesa-structure and the n-doped layer is provided adjacent the first electrode because the conductivity of the n-doped layer is typically higher than the conductivity of the p-doped layer, leading to a further reduction in the effective series resistance.
Preferably, a push-pull driver is used to charge and discharge the capacitance of the QCSE device so that both the charging and discharging of the capacitance of the QCSE device are both actively driven. The maximum data
transmission rate depends upon how quickly the electric field across the plurality of quantum wells changes, which mainly depends upon the output impedance of the push-pull driver, the effective series resistance of the QCSE device and the capacitance of the QCSE device (which increases with size) .
In an embodiment, in order to reduce the output impedance of the driver circuit (and thereby increase the maximum data transmission rate) , a plurality of push-pull drivers are driven in parallel. However, a problem with connecting the push-pull drivers in parallel is that, due to manufacturing variations, each push-pull driver switches at a slightly different timing causing temporary current surges when two or more of the push-pull drivers are in different states. In order to reduce the power drain caused by these current surges, in a preferred embodiment a resistor is provided in series with each push-pull driver, the resistor having a resistance which is sufficiently small that the total output impedance of the driver circuit is less than the output impedance of a single push-pull driver used in the driver circuit.
In an embodiment of the signalling system, in order to convey information from a remote terminal to a modulator unit, an amplitude modulation is applied to the optical beam emitted by the remote terminal. At the distribution node, the optical beam is detected and converted into a corresponding electrical signal by a detector, and the corresponding electrical signal is processed to recover the information from the remote
terminal. Typically, this processing involves amplifying the electrical signal and demodulating the amplified signal.
A problem with amplifying the electrical signal is that background light forms a DC signal component. Further, variations in the transmission path of the optical beam through the atmosphere, which could originate from vibrational movement of the devices at either end of the transmission path or changes in the optical properties of the transmission path caused by smoke and the like, introduces a low-frequency AC signal component in the frequency range up to about 100kHz. The DC signal component and the low-frequency AC signal component effectively form a noise signal.
According to a third aspect of the invention, there is provided a signalling system comprising first and second signalling devices, the first signalling device comprising: a receiver operable to receive a free-space optical beam conveying information within a transmission frequency band from the second signalling device; an optical-to-electrical converter for converting a portion of the received optical beam into a corresponding electric signal; and a processor for processing the corresponding electric signal to retrieve the conveyed information. The processor includes a cancellation circuit operable to remove part of the electric signal generated by the optical to electric converter in a frequency range below the transmission frequency band.
In an embodiment, the second signalling device applies a small signal amplitude modulation in which case the cancellation circuit of the first signalling device is advantageous because the unmodulated part of the optical beam received by the first signalling device is substantially removed by the cancellation circuit.
A problem with mounting a plurality of detachable mounting units, or other types of optical unit, to a support structure is how to add or remove a modulator unit without disturbing other modulator units already mounted to the support structure, particularly when the other modulator units are in the process of data communication.
According to a fourth aspect of the invention there is provided a free-space optical signalling system for signalling between a distribution node and a plurality of remote terminals, the distribution node comprising a plurality of optical units, each optical unit operable either to detect information conveyed by an incoming modulated light beam or to transmit information by producing a modulated outgoing light beam, and a support operable to support the plurality of modulator units at respective different orientations. The support comprises one or more elongate mounting rods and each, optical unit includes a clamping device having a first part and a second part, each of the first and second parts having a bearing surface. The first and second parts are relatively movable between a first configuration, in which the bearing surfaces of the first and second parts
define a passage for receiving the rod of the support, and a second configuration, in which the first and second parts define a gap through which the mounting rod is insertable into the clamping device by moving the clamping device and the mounting rod relative to each other in a radial direction with respect to the axis of the mounting rod. In this way, in the second configuration the mounting rod can be inserted into the clamping device without disturbing the orientation of any optical units previously mounted to the mounting rod, and when the mounting rod has been inserted into the clamping device, the first and second parts are movable into the first configuration in which the bearing surfaces abut the mounting rod.
Preferably, the clamping device and the mounting rod are both thermally conductive so that they act as a heat sink for the optical unit.
Preferably, the mounting rod has a circular cross-section and the clamping device is rotatable about the axis of the mounting rod, and the main body of the optical unit is rotatable relative to the clamping device about an axis of rotation transverse to the axis of the mounting rod. In this way the optical unit can be aligned by rotating the optical unit about the axis of rotation of the clamping device, and by rotating the clamping device about the axis of the mounting rod.
Preferably, the clamping device includes " a fixing mechanism including a single actuator which is operable
to fix both the orientation of the main body of the optical unit relative to the clamping device and the orientation of the clamping device relative to the mounting rod. This facilitates the installation of an optical unit onto the support.
Exemplary embodiments of the invention will now be described with reference to the accompanying drawings in which:
Figure 1 is a schematic diagram of a point-to-multipoint communication system for distributing data between an Ethernet network and a plurality of user devices;
Figure 2 is a schematic diagram of a user terminal which forms part of the communication system shown in Figure 1;
Figure 3 shows a perspective view of the user terminal illustrated in Figure 2;
Figure 4 is a schematic diagram showing the detection surface of a detector forming part of the user terminal illustrated in Figure 2;
Figure 5 is a plot illustrating the way that the power of a laser beam emitted by the user terminal shown in Figure 2 is varied to achieve a small signal modulation for transmitting uplink data from the user terminal to a local distribution node forming part of the communication system shown in Figure 1;
Figure 6 is an eye diagram schematically illustrating the effect of the small-signal modulation on the detection by the user terminal of downlink data transmitted from the local distribution node to the user terminal;
Figure 7 is a schematic diagram of a microhub forming part of the local distribution node of the communication system illustrated in Figure 1;
Figure 8 is a schematic perspective view of a modulator which forms part of the microhub illustrated in Figure 7;
Figure 9A is a plan view showing the layout of a first electrode forming part of the modulator shown in Figure 8;
Figure 9B is a plan view showing the layout of a second electrode forming part of the modulator shown in Figure 8;
Figure 9C schematically shows a first sectional view of the modulator illustrated in Figure 8;
Figure 9D schematically shows a second sectional view of the modulator illustrated in Figure 8;
Figure 10 is a signal diagram which schematically illustrates the way in which the light incident on the modulator shown in Figure 8 is modulated in dependence upon the bias voltage applied to the electrodes thereof;
Figure 11 is a schematic block diagram showing in more detail the components of a modulator drive circuit which forms part of the microhub illustrated in Figure 7;
Figure 12 schematically shows in more detail a single driver circuit which forms part of the modulator drive circuit illustrated in Figure 11;
Figure 13 is a schematic block diagram showing in more detail a detection circuit which forms part of the microhub illustrated in Figure 7;
Figure 14 is a circuit diagram showing in more detail a DC cancellation unit and an amplifier which form part of the detection circuit shown in Figure 13;
Figure 15 is a perspective view of the microhub schematically illustrated in Figure 7;
Figure 16 is a cross-sectional view of a mounting device forming part of the microhub illustrated in Figure 15 in a clamped configuration;
Figure 17 is a cross-sectional view of the mounting device of the microhub illustrated in Figure 15 in a mounting configuration;
Figure 18 is a perspective view of a support structure supporting a plurality of microhubs as illustrated in Figure 15;
Figure 19 is a perspective view of an alternative mounting device to the mounting device illustrated in Figures 17 and 18;
Figure 20 is a cross-sectional view of the alternative mounting device shown in Figure 19 in a clamped configuration;
Figure 21 is a cross-sectional view of the alternative mounting device shown in Figure 19 in a mounting configuration; and
Figure 22 is a plan view showing the layout of an alternative second electrode to the second electrode illustrated in Figure 9B.
Figure 1 schematically illustrates a communication system which employs a point-to-multipoint signalling system to transmit data between a lOOMbit/s Ethernet network 1 and a plurality of user devices 3a to 3c. The point-to- multipoint signalling system utilises free-space optical links 5a to 5c to transmit data between a local distribution node 7 , which is connected to the Ethernet network 1, and a plurality of user terminals 9a to 9c which are each connected to a respective user device 3. As shown in Figure 1 , the local distribution node 7 has a plurality of microhubs 11a to lie with each microhub 11 communicating with a respective one of the user terminals 9 via a respective optical link 5.
For illustrative purposes, three user devices 3, three
user terminals 9 and three microhubs 11 are shown in Figure 1. However, the number of user devices 3, and correspondingly the number of user terminals 9 and microhubs 11, is variable so that more or fewer user devices 3 can communicate with the Ethernet network 1 via the point-to-multipoint signalling system.
Each user terminal 9 emits a low divergence free-space optical beam, which is modulated in accordance with uplink data to be conveyed to the Ethernet network 1, and directs the emitted optical beam at the corresponding microhub 11. Each microhub 11 has a detector (not shown in Figure 1) for detecting part of the optical beam from the corresponding user terminal 9 and recovering the uplink data, and a modulator (not shown in Figure 1) which modulates and retro-reflects part of the optical beam from the corresponding user terminal 9 to convey downlink data from the Ethernet network 1 to the user device 3.
The local distribution node 7 also includes an Ethernet switch 13 which is connected between the Ethernet network 1 and each of the microhubs 11. The Ethernet switch 13 receives downlink data for all the user devices 3 from the Ethernet network 1 in the form of data blocks, each data block having an address portion, identifying the desired recipient, and an information portion. Each user device 3 has an associated address, and for each received data block the Ethernet switch 13 checks if the address in the address portion corresponds to the address of any of the user devices 3, and if so transmits the data block
to the microhub 11 corresponding to the identified user device 3. Similarly, the Ethernet switch 13 receives uplink data from the user devices 3, via the corresponding microhubs 11, and transmits the received uplink data to the Ethernet network 1. The local distribution node 7 also includes a network management unit 15 which is connected to each of the microhubs 11 and monitors the operation of the free-space optical links 5 between the microhubs 11 and the user terminals 9.
Figure 2 schematically shows in more detail the main components of one of the user terminals 9 of the communication system shown in Figure 1. As shown, the user terminal 9 includes a laser diode 21 which outputs a beam 23 of coherent light. In this embodiment, the user terminals 9 are designed so that they can communicate with a local distribution node 7 within a range of 200 metres with a high link availability. To achieve this, the laser diode 21 is a 50mw laser diode which outputs a laser beam having a wavelength of 785nm.
The output light beam 23 is passed through a lens 25, hereafter called the collimating lens 25, which reduces the angle of divergence of the light beam 23 to form a substantially low divergence light beam 27. The divergence of the low divergence light beam can be varied by varying the distance between the collimated lens 25 and the laser diode 21. However, those skilled in the art will appreciate that a perfectly collimated light beam is not possible due to diffraction at the emitting
aperture of the laser diode 21. In this embodiment, the collimating lens 25 is a low aberration lens, so that the low divergence beam 27 has a relatively uniform wavefront, with a 50mm diameter and an F-number which is just large enough to collect all the light emitted by the laser diode 21.
Although the divergence of the light beam 27 is low, the size of the light beam 29 incident on the user terminal 9, after reflection by the corresponding microhub 11 at the local distribution node 7, is sufficiently large that, as shown in Figure 3, part of the returned light beam is incident on a lens, hereafter called the downlink detection lens 31, which is provided adjacent to the collimating lens 25.
Returning to Figure 2 , in which for clarity only the portion of the received light beam 29 incident on the downlink detection lens 31 is shown, the downlink detection lens 31 focusses light from the received light beam 29 onto a detector 33, which in this embodiment is an avalanche photodiode. The downlink detection lens 31 has a diameter of 100mm but is not required to be of as high quality as the collimating lens 25 because its primary purpose is simply to direct as much light as possible onto the detector 33. Figure 4 schematically shows the detection surface 61 of the detector 33 and the light spot 63 formed by the downlink detection lens 31 focussing light from the received light beam 29. In this embodiment, the diameter of the detection surface 61 is
500μm whereas the diameter of the light spot 63 is approximately 50μm.
The detector 33 converts the received light beam into a corresponding electrical signal which varies in accordance with the modulation provided at the local distribution node 7. The electrical signal is amplified by an amplifier 35 and then filtered by a filter 37. The filtered signals are input to a central control unit 39 which performs a clock recovery and data retrieval operation to regenerate the data from the Ethernet network 1. The retrieved data is then passed to an interface unit 41 which is connected to the corresponding user device 3.
The interface unit 41 also receives uplink data from the user device 3 and transmits the received data to the central control unit 39, which sends the uplink data to a laser driver 43 which modulates the light beam 23 output by the laser diode 21 in accordance with the uplink data. In this embodiment, the laser driver 43 applies a small signal modulation to the light beam 23 output by the laser diode 21. Figure 5 illustrates this modulation and shows the CW laser level 65 and the small signal modulation 67 applied to it. As those skilled in the art will appreciate, this uplink modulation data becomes an additional noise source for the downlink data. This is illustrated in Figure 6 which shows an eye diagram for the downlink data 69, which includes the interfering uplink data 67, and the consequent reduction
in the noise margin 71. However, if the uplink modulation depth is kept sufficiently low, then both the uplink and the downlink can operate with equal bandwidth. In this embodiment, the uplink modulation depth is approximately 10% of the CW laser level. Further details of this type of small signal modulation can be found in International Patent Application WO 01/05071 and US Patent Application No. 10/038576, the whole contents of which are incorporated herein by reference.
In this embodiment, the optical link 5 between a user terminal 9 and a microhub 11 essentially acts as a data pipe, i.e. data received by the user terminal 9 is transmitted without any further encoding ( for example without adding additional error detection and correction bits) to the microhub 11 and vice versa. In order to transmit information concerning the optical link 5 between the user terminal 9 and the microhub 11, a separate "operation and maintenance" (OAM) channel is formed by modulating the timing of the data clock signal in accordance with OAM data. In this embodiment, when performing the clock and data recovery operation, the central control unit 39 monitors the clock timing to recover OAM data sent by the microhub 11 at the other end of the optical links 5.
Returning to Figure 2, the central control unit 39 is also connected to a first motor driver 45a for supplying drive signals to a first stepper motor 47a, and to a second motor driver 45b for supplying drive signals to a second stepper motor 47b. The laser diode 21, the
collimating lens 25, the detector 33 and the downlink detection lens 31 are mounted together to form a single optical assembly 51, and the first and second stepper motors 47 are operable to rotate the optical assembly 51 about respective orthogonal axes. In this embodiment, when mounted the first and second stepper motor 47 typically rotate the optical assembly 51 about respective axes orientated at 45° either side of the vertical. In this way, the direction of the emitted light beam is variable .
Figure 7 schematically illustrates the main components of one of the microhubs 11. As shown, the microhub 11 includes an Ethernet interface 81, which is connected to the Ethernet switch 13 of the local distribution node 7. The Ethernet interface 81 is also connected to an input of a modulator drive circuit 83 and an output of a detection circuit 85. An output of the modulator drive circuit 85 is connected to an optical modulator 87, and an input of the detection circuit 85 is connected to a light detector, which in this embodiment is a photodiode 89.
The Ethernet interface 81 directs downlink data signals received from the Ethernet switch 13 to the modulator drive circuit 83, which generates corresponding drive signals for the optical modulator 87. The Ethernet interface 81 also receives uplink data signals generated by the detection circuit 85 from electrical signals formed by light incident on the photodiode' 89, and transmits the received uplink data signals to the
Ethernet switch 13. The detection circuit 85 also retrieves the OAM data signal from the remote terminal 9 and transmits the retrieved OAM data signal to the network management unit 15.
The modulator 87 is positioned approximately in the back focal plane of a telecentric lens 91, which is schematically represented in Figure 7 by a lens element 93 and a stop member 95 having a central aperture 97 positioned in the front focal plane. Those skilled in the art will appreciate that in practice more than one lens element is likely to be used in the telecentric lens 91 with the exact arrangement being a design choice depending upon the particular requirements of installation. The size of the aperture 97 is also a design choice, with a large aperture 97 transmitting more of the light from the corresponding user terminal 9 than a small aperture 97 but requiring a more complex and expensive lens arrangement to focus the light than is required with the small aperture 97.
In this embodiment, the telecentric lens 91 is axially symmetric and has a conical field of view with a half- angle of 1.25°.
The telecentric lens 91 focuses incident light at a position within the back focal plane which depends upon the angle of incidence of the light so that different angles of incidence within the field of view of the telecentric lens are mapped to different positions on the modulator 87. Further, the principal rays transmitted
through the telecentric lens 91 are incident perpendicular to the back focal plane and therefore the modulator 87 reflects light incident along a principal ray back along its path of incidence. In this way, the modulator 87 and the telecentric lens 91 act as a retro- reflector. An advantage of using the telecentric lens 91 is that the efficiency of modulation (i.e. the modulation depth) of existing optical modulators 87 generally depends upon the angle at which the light beam hits the modulator 87, and using the telecentric lens 91 ensures that the principal rays of the light beams are incident parallel to the optical axis of the modulator 87 regardless of the position of the user terminal 9 generating the incident light beam within the field of view of the telecentric lens 91. The dependency of the efficiency of modulation upon the position of the user terminal 9 relative to the local distribution node 7 is therefore substantially removed.
A further advantage of using a telecentric lens 91 is that, if required, the position of the modulator 87 can be moved slightly out of the back focal plane of the telecentric lens in order to provide a refocusing effect which reduces the spot size of the reflected light beam at the user terminal 9, thereby increasing the signal level at the user terminal 9. In this embodiment, the lens elements of the telecentric lens 91 are mounted to a cylindrical lens mount, and the lens mount is mounted within a cylindrical housing. A projection from the lens mount is sandwiched between the tip of a screw and a spring which provides a biassing force abutting the
projection against the screw tip. During installation, the screw is adjusted by an installer to move the telecentric lens 91 relative to the modulator 87 in order to adjust the divergence of the reflected light beam.
Preferably, the telecentric lens 91 has as large a collection aperture as possible in order to collect as much of the incoming light as possible. However, in order to increase the collection aperture, the focal length must also be increased in order to avoid the use of complex and expensive low f-number lenses. This increase in focal length increases the required size of the active area of the modulator 87 in order for all incoming light beams within the field of view of the telecentric lens 91 to be incident on the active area. As will be described in more detail hereafter, in this embodiment the active area of the modulator 87 is an octagon having sides with a length of approximately 0.75mm.
The photodiode 89 is placed in the back focal plane of an uplink detection lens 99. As schematically shown in Figure 7 , the collection aperture and focal length of the uplink detection lens 99 are less than those of the telecentric lens 91. This allows the active area of the photodiode 89 to be reduced in comparison with the optical modulator 87, which is advantageous because this reduces the capacitance (and therefore the response time) of the photodiode 89. The reason why the collection aperture of the uplink detection lens 99 can be made less than that of the telecentric lens 91 is that the
photodiode 89 need only detect sufficient light to be able to recover the uplink data signal and OAM signal, whereas the telecentric lens 91 must collect enough light to ensure that the retro-reflected light beam has enough energy for a sufficient signal level to be achieved at the detector 33 in the user terminal 9.
Figure 8 shows a schematic perspective view of the modulator 87. As shown, the modulator 87 is formed by a mesa-structure 101, in the form of an octagonal right prism, on a substrate 103. A first electrode 105 is formed by an octagonal ring around the base of the mesa- structure 101, and a second electrode 107 is formed at the top surface of the mesa-structure 101. The first electrode 105 includes six contact pads 109a to 109f which are shown most clearly on the plan view of the first electrode illustrated in Figure 9A.
Figure 9B shows a plan view of the second electrode formed on the top surface of the mesa-structure 101. As shown, the second electrode 107 is formed by first and second C-shaped conductors Ilia, 111b with eighteen strip conductors 113_1 to 113_18 extending from the first C- shaped conductor Ilia toward the second C-shaped conductor 111b, and a further eighteen strip conductors 113_19 to 113_36 extending from the second C-shaped conductor 111b toward the first C-shaped conductor. The strip conductors 113 are all parallel to each other with the eighteen strip conductors 113 connected to the first C-shaped electrode Ilia being respectively aligned with the eighteen strip conductors connected to the second C-
shaped conductor 111b to form eighteen pairs of strip conductors 113. Each pair of strip conductors 113 is separated by a small gap which allows an etchant to flow between the strip conductors 113 during fabrication of the modulator 87.
As shown for the second C-shaped conductor 111b in Figure 8, each C-shaped conductor 111 is connected to three connection pads 115a to 115e via eight conductive tracks 117a to 117h which extend down the side walls of the mesa-structure and over the first electrode 105. An insulating layer (not shown) is provided between the side wall of the mesa-structure 101 and the conductive tracks 117 and between the first electrode 105 and the conductive tracks 117. Figure 9C shows a cross-section of the modulator 87 in a plane perpendicular to the surface of the semiconductor substrate 103 and parallel with the strip electrodes 113, and Figure 9D shows a cross-section of the modulator 87 through a plane perpendicular to the surface of the semiconductor substrate 103 and perpendicular to the strip electrodes 113. As shown, the side walls of the mesa-structure 101 are not perpendicular to the surface of the semiconductor substrate 103. In particular, the side walls over which the conductive tracks 117 extend slope inwardly from the base to the top of the mesa-structure 101. This allows the conductive tracks to be deposited upon these sloping side walls. As shown in Figure 9D, the side walls of the mesa-structure over which no tracks extend slope outwardly from the base to the top of the mesa-structure 101 so that the top of the mesa-structure 101 partly
overhangs the base of the mesa-structure 101.
As shown in Figures 9C and 9D, the modulator 87 comprises five layers, three of which are formed in the mesa- structure 101. In this embodiment, the layers formed are based on Gallium Arsenide (GaAs) and Aluminium Gallium Arsenide (AlGaAs). In particular, the mesa-structure 101 is formed by a p-conductivity type GaAs layer 101_1 formed on an intrinsic AlGaAs layer 101_2 having a plurality of quantum wells formed therein, which is in turn formed on a n-conductivity type AlGaAs layer 101_3 having a Bragg reflector formed therein. The mesa- structure 101 is formed on an n-conductivity GaAs contact layer 103_1 which is in turn formed on an intrinsic GaAs substrate 103_2.
In operation, the light beam from the user terminal 9 passes through the p-conductivity GaAs layer 101__1, which is made thin so that only a small proportion of the incident light is absorbed. The amount of light absorbed by the intrinsic AlGaAs layer 101__2 depends upon the DC bias voltage applied to the second electrode 107. Ideally, when a DC bias voltage of approximately -5 volts is applied to the electrode 107, the incident light is totally absorbed within the intrinsic AlGaAs layer 101_2. Consequently, when there is a DC bias voltage of approximately -5V applied to the electrode 107, no light is reflected back to the corresponding user terminal 9. On the other hand, when no DC bias voltage is applied to the electrode 107, the light beam from the corresponding user terminal 9 passes through the intrinsic AlGaAs layer
101_2 and is reflected by the n-conductivity AlGaAs layer 101_3. Therefore, by changing the bias voltage applied to the second electrode 107 in accordance with the drive signals from the modulator drive circuit 83, the modulator 87 amplitude modulates the received light beam and reflects the modulated light beam back to the user terminal 9.
In the ideal case, as illustrated in Figure 10, a zero voltage bias, resulting in the light from the user terminal 9 being reflected back from the QCSE modulator 87, is applied to the electrode 107 to transmit a binary 1 and a DC bias voltage of -5 volts is applied to the second electrode 107, resulting in no reflected light, to transmit a binary 0. Typically, however, the QCSE modulator 87 reflects 70% of the light beam when no DC bias is applied to the second electrode 107 and 40% of the light beam when -5 volts DC bias is applied to the second electrode 107. Therefore, in practice, there will be a difference of about 45% between the amount of light which is reflected back to the user terminal 9 when a binary 0 is transmitted and when a binary 1 is transmitted.
In this embodiment, forty quantum wells are formed in the intrinsic AlGaAs layer 101_2. The amount of the received light beam absorbed by the intrinsic AlGaAs layer 101_2 can be increased by adding additional quantum wells to increase the depth of the intrinsic AlGaAs layer 101_2. However, if the depth of the intrinsic AlGaAs layer 101_2 is increased, then a higher voltage must be applied to
the second electrode 107 in order to produce the required electric field across the intrinsic AlGaAs layer 101_2. There is, therefore, a trade-off between the absorptivity of the intrinsic AlGaAs layer 101_2 and the voltage applied to the second electrode 107.
The strip conductors 113 distribute the drive signal from the modulator drive circuit 83 over the surface of the p-conductivity layer 101_1. This reduces the effective series resistance of the modulator 87 in comparison with the case where the strip electrodes 125 are not included (for example, if the second electrode is formed by only the C-shaped conductors 111) because current flowing through the centre of the mesa-structure 101 is able to flow along the strip conductors 113 to the centre of the top surface of the mesa-structure 101.
In this embodiment, the modulator 87 is fabricated by sequentially forming on the substrate 103_2 the n- conductivity GaAs contact layer 103_1, the n-conductivity type AlGaAs layer 101_3 including the Bragg reflector, the intrinsic AlGaAs layer 101_2 including the plurality of quantum wells and the p-conductivity type GaAs layer 101_1. The C-shaped conductors 111 and the strip conductors 113 are then deposited. Next, the mesa- structure 101 is formed by etching down to the n- conductivity contact layer 103_1. The first electrode 105 and the contact pads 115 are then deposited around the base of the mesa-structure 101, after which the insulating layer is deposited on the inwardly sloping sides of the mesa-structure 101 and the first electrode
105. Finally, the conductive tracks 117 are deposited on the insulating layer to form a connection between the C-shaped conductors 111 and the contact pads 115.
The first electrode 105 and the connection pads 115 are fabricated by initially depositing very thin nickel layers, and then depositing gold on the nickel. The C- shaped conductors 111 and the strip conductors 113 are fabricated by initially depositing very thin platinum layers, and then depositing gold on the platinum layers. The nickel and platinum layers are used because gold does not adhere well to Gallium Arsenide based material. The nickel and platinum layers also prevent the gold diffusing into the Gallium Arsenide, which would degrade the performance of the modulator 87.
An advantage of forming the first electrode 105 and the connection pads 115 using gold is that they provide a reference reflector when measuring the reflectance performance of the modulator 87.
In this embodiment, the strip conductors 113 are formed in straight lines with a width of 2μm. This width is sufficient for the strip conductors 113 to have a low resistance in comparison with the p-conductivity type GaAs layer 101__1 without covering a significant proportion of the active area. In particular, the telecentric lens 91 forms spots on the modulator 87 with a diameter in the region of 20 to 80μm, and therefore the strip conductors 113 do not significantly "affect the retro-reflected light beam.
The capacitance of the p-i-n diode of the modulator 87 of this embodiment is approximately 400pF under depleted conditions. A 74AC CMOS line driver has an output impedance in the region of 20 ohms. Therefore, if a single 74AC CMOS line driver is used to drive the modulator 87, the RC time constant for varying the electric field across the quantum wells of the modulator 87 is in the region of 8ns. This is too slow for communicating at 155 Mbits/second.
As shown in Figure 11, the modulator drive circuit 83 includes eight CMOS line drivers 141a to 141h which are each connected in series with one of eight resistors 143a to 143h. Figure 12 schematically shows the contents of one of the CMOS drivers 141. As shown, the CMOS driver 141 includes a p-channel MOSFET 145 whose source is connected to a positive supply rail and whose drain is connected to the drain of a n-channel MOSFET 147. The source of the n-channel MOSFET 147 is connected to ground. An input port 148 is connected in common to the control electrode (i.e. the gate) of the p-channel MOSFET 145 and the control electrode of the n-channel MOSFET 147, and an output port 149 is connected in common to the drain of the p-channel MOSFET 145 and the drain of the n-channel MOSFET 147. In this way, if a high signal level is applied to the input port 148 the p-channel MOSFET 145 is switched off and the n-channel MOSFET 147 is switched on so that a low voltage level is seen at the output port 149, and if a low signal level is applied to the input port 148 then the p-channel MOSFET 145 is switched on and the n-channel MOSFET 147 is switched off
so that a high voltage level is seen at the output port 149.
In this embodiment, the eight CMOS line drivers 141 are integrally formed on a single chip. However, the switching times of the CMOS drivers still vary slightly and therefore during each switching operation momentarily one of the CMOS drivers 141 can be in a HIGH state while another of the CMOS drivers 141 is a LOW state. The resistors 143 are therefore included in the modulator drive circuit 83 to reduce the temporary current surges, sometimes referred to as shoot-through, which occur when two of the CMOS drivers 141 are in different states.
In this embodiment, each of the resistors 143 has a resistance of 10 ohms. The downlink data signal is connected to the respective inputs of all the CMOS drivers 141 and the drive signals output by the CMOS drivers 141, through the respective resistors 143, are applied to the anode of the modulator 87. In this way, the effective output impedance of the modulator drive circuit 83 is equivalent to eight resistors, each having a resistance of 30 ohms (corresponding to the series connection of the twenty ohms output impedance of a CMOS driver and the ten ohms resistance of the corresponding resistor 143), connected in parallel. The effective output impedance is therefore 3.75 ohms, giving an RC time constant of 1.5ns for varying the electric field across the quantum wells of the modulator 87. This reduction in output impedance, in comparison with a single CMOS driver, allows the electric field to be
switched at a fast enough rate for 155Mbits/s data communication.
The electric signal from the detector 89 consists of three main components: the modulated part of the incoming light beam from a corresponding user terminal 9 ; the unmodulated part of the incoming light beam from the corresponding user terminal 9; and background light. The respective amplitudes of these three components varies, due to changes in the transmission path, at a rate of up to approximately 100kHz.
Figure 13 shows the main components of the detection circuit 85 of the microhub 11. As shown, the electric signal from the photodiode 89 is input to a DC cancellation unit 151, which removes most of the components of the electrical signal relating to the unmodulated part of the incoming light beam and the background light, and also reduces any low frequency variations caused by changes to the transmission path. The remainder of the electrical signal is output by the DC cancellation circuit 151 and input to an amplifier 153. The amplified signal output by the amplifier 153 is input to a clock recovery unit 155, which recovers the uplink data signal and any OAM signal from the user terminal 9. In this embodiment, the DC cancellation unit 151 also outputs a signal representative of the total DC received signal strength, hereafter called the DC-RSSI signal, to the clock recovery unit 155 and the amplifier 153 outputs a signal representative of the signal strength at the data transmission frequency to the clock
recovery unit 155, hereafter called the AC-RSSI signal. The clock recovery unit 155 forwards the DC-RSSI signal and the AC-RSSI signal to the network management unit 15.
Figure 14 shows in more detail the DC cancellation unit 151 and the amplifier 153 of the detection circuit 85. As shown, the cathode of the photodiode 89 is connected to electrical ground via a capacitor Ci, which in this embodiment has a capacitance of 3.3nF," and to the input branch of a first current mirror 161. In this embodiment, the first current mirror 161 is formed by a conventional PNP double transistor with matched emitter resistors Rx and R2, each having a resistance of 470Ω.
The output branch of the first current mirror 161 is connected to one end of a third resistor R3, which in this embodiment has a resistance of 470Ω. A fourth resistor R4, which in this embodiment has a resistance of lOOkΩ, and a second capacitor C2, which in this embodiment has a capacitance of lOnF, are connected in parallel between the other end of the third resistor R3 and electrical ground. The other end of the third resistor R3 is also connected to the input branch of a second current mirror 165, which in this embodiment is a Wilson mirror formed by a conventional NPN double transistor 167 and a single NPN transistor 169. The output branch of the second current mirror 165 is connected, via an inductor L having an inductance of 47uH, to the anode of the photodiode 89. The anode of the photodiode 89 is also connected to the input to a pre-amplifier 171.
When light is incident on the photodiode 89, a current Ic flows through the cathode of the photodiode and a corresponding current Ia flows through the anode of the photodiode. The cathode current Ic consists of a high frequency part Ih, which flows to electrical ground via the capacitor Ci, and a low frequency part Ilf which is input to the first current mirror 161. In this embodiment, the 3dB cut-off frequency between the high frequency part Ih and the low frequency part Ix is 100kHz. In this way, the data traffic component, at 155Mbits/s, flows through the capacitor d rather than the first current mirror 161.
The first current mirror 161 causes a mirror current matching the low frequency part I
x to flow through the third resistor R
3. The voltage at the input of the second current mirror 165 is limited at approximately 1.4V
" by the Wilson mirror arrangement and therefore a current, hereafter called the bleed current I
b, of up to approximately 14μA flows to electrical ground through the fourth resistor R
4. A high frequency noise current I
n similarly flows to electrical ground through the second capacitor C
2. The current I
s flowing into the input branch of the second current mirror 165 is therefore substantially given by:
The second current mirror 165 causes a mirror current matching the current Is to flow through the output branch
of the second current mirror 165. The detection current Id flowing into the pre-amplifier 171 is therefore substantially given by:
(2)
In this way, the detection current Id corresponds primarily to the modulated component of the incoming light beam. However, some low frequency current, substantially equivalent to the bleed current Ib, also forms part of the detection current Id. This prevents current being sourced to the output branch of the second current mirror 165 from the input of the pre-amplifier 171.
In this embodiment, the pre-amplifier 171 is a MAX3963 transimpedance pre-amplifier which is available from Maxim Integrated Products. The pre-amplifier 171 converts an input current signal into a corresponding differential voltage signal output from a non-inverting output port (0UT+) and an inverting output port (OUT-). The non-inverting output port (0UT+) of the pre-amplifier 171 is capacitively coupled, via a third capacitor C3, to an inverting input (IN-) of a limiting amplifier 173, which in this embodiment is a MAX3964 limiting amplifier available from Maxim Integrated Products. Similarly the inverting output (OUT-) of the pre-amplifier 171 is capacitively coupled, via a fourth capacitor C , to a non-inverting input (IN+) of the limiting amplifier 173. A fifth capacitor C5 is connected between the non-
inverting and inverting outputs of the pre-amplifier 171 to reduce high frequency noise.
The limiting amplifier 173 outputs, via non-inverting and inverting outputs (OUT+,OUT-), a PECL (positive emitter- coupled logic) differential data signal which is output to the clock recovery unit 155 via a first connector 175, which in this embodiment is a fifty ohm, co-axial cable connector.
The limiting amplifier 173 also outputs a signal, via complementary loss-of-signal output ports LOS+, LOS- indicating when the input power level falls below a threshold determined by fifth and sixth resistors Rs and R6. The signal output by the non-inverting loss-of- signal output LOS+ is output to the clock recovery unit 155 via a second connector 177, which in this embodiment is a ribbon cable connector.
As discussed above, the voltage at the input to the second current mirror 165 is limited to approximately 1.4V by the Wilson mirror arrangement and the current flowing through the third resistor R3 matches the low frequency part Ix of the cathode current Ic. The voltage level at the output of the first current mirror 161 is therefore representative of the detected low frequency signal level, hereafter called the DC signal level. This voltage level is sampled by connecting the interconnection between the output of the first current mirror 161 and the third resistor R3, via a" first low pass filter 179, to a first unity gain buffer 181. The
output of the first unity gain buffer 181 is connected, via a second low pass filter 183, to the second connector 177. In this way, the DC-RSSI signal is transmitted to the clock recovery unit 155.
The limiting amplifier 173 includes a RSSI port which outputs the AC-RSSI signal, via a second unity gain buffer 185, to the second connector 177 for transmission to the clock recovery unit 155.
The clock recovery unit 155 receives, via the first connector 175, the differential data signal output by the limiting amplifier 173 and performs a clock recovery and data regeneration operation during which the OAM signal is retrieved by monitoring variations in the clock timing. The uplink data signal regenerated by the clock recovery unit 155 is transmitted to the Ethernet switch 13, and the OAM signal is transmitted to the network management unit 15 along with the LOS signal, the DC-RSSI signal and the AC-RSSI signal.
Figure 15 shows a perspective view of one of the microhubs 11. As shown, the telecentric lens 91 and the uplink detection lens 99 are supported by a housing 201 which houses the Ethernet interface 81, the modulator drive circuit 83 and the detection circuit 85. A clamping device 203 is also mounted to the housing 201. Figures 16 and 17 show cross-sectional views of the clamping device 203 in a clamped configuration and a mounting configuration respectively.
The clamping device 203 includes first and second clamp pieces 205a, 205b which are generally symmetric. Each clamp piece 205 has a near semi-cylindrical shape with a transverse part-cylindrical bearing surface 207a, 207b formed in the axial planar surface. When the first and second clamp pieces 205 are aligned in the clamping configuration, as shown in Figure 16, the first and second clamp pieces 205 are separated by a diametric axial slot and define a generally cylindrical shape. Due to the symmetry of the first and second clamp pieces 205, in the clamped configuration the bearing surfaces 207 face each other forming a part-cylindrical passage, whose axis perpendicularly intersects the axis of the cylindrical shape defined by the clamp pieces 205, for receiving a circular mounting rod.
The first and second clamp pieces 205 each have a circumferential groove 209a, 209b positioned toward one axial end, hereafter called the base end, which align with each other when the clamp pieces 205 are in the clamping configuration. At the other axial end, hereafter called the mounting end, a threaded bore 211 and a screw hole 213 are respectively formed in the first and second clamp pieces 205a, 205b so that in the clamped configuration, a screw 215 can pass through the screw hole 213 of the second clamp piece 205b and engage the threaded bore 211 of the first clamp piece 205a. A part-conical surface 217a, 217b is also formed at the base end of each clamp piece 205 so that when the two clamp pieces 205 are in the clamping configuration, the two part-conical surfaces 217 define a conical frustum
whose base is in the plane of the base end and whose top defines first and second reaction surfaces 219a, 219b of the first and second clamp pieces 205 respectively.
As shown in Figures 16 and 17, the groove 209 and the conical surface 217 of each clamp piece 205 define a jaw portion 221a, 221b which fits within a lipped recess formed in the housing 201. In particular, the lipped recess is formed by a ring 223 which is screwed to the housing and which has a circular lip 225 that engages the groove 209 of each clamp piece 205. As shown, the thickness of the lip 225 is less than the axial extent of the groove 209 so that each clamp piece 205 can pivot about the lip 225. In order to allow this pivotal movement within the recess, a gap is formed between the base end of each clamp piece 205 and the housing 201 by making the axial distance between the base end of each clamp piece 205 and the groove 209 less than the axial distance between the housing 201 and the lip 225 of the ring 223.
The reaction surfaces 219 formed at the top of the part-conical surfaces 217 are axially aligned with the approximate centre of the circumferential grooves 209. A spring 227 is positioned between the housing 20i and the clamp pieces 205 and applies a biassing force, via a cover 229, to the reaction surface 219 of each clamp piece 205. As shown in Figure 17, this biassing force causes each clamp piece 205 to pivot about the lip 22.5 of the ring 223 and therefore causes the clamppieces 205 to separate at the mounting end. In particular, the
separation of the clamp pieces 205 creates a gap which is wider than the diameter of the part-cylindrical passage formed by the bearing surfaces 207 in the clamping configuration so that a mounting rod can be passed radially, with respect to the axis of the mounting rod, through the gap.
During a mounting operation, initially the screw 215 is removed so that the clamping device 203 is biassed into the mounting configuration by the spring 227. The clamping device 203 is then manually moved so that a mounted rod passes through the gap between the two mounting ends and is located substantially adjacent the two clamping surfaces 207. The mounting ends of the two clamp pieces 205 are then manually moved toward each other and the screw 215 is inserted through the screw hole 213 and engages the threaded bore 211. The screw 215 is then tightened until the bearing surfaces 207 loosely engage the mounting rod, at which position the circumferential surfaces of the jaw portions 221 loosely engage the inner circumferential surface of the ring 223. In this position, the clamping device 203 is rotatable about the axis of the mounting rod and the microhub 11 is rotatable around the axis of the clamping device 203. Therefore, an installer is able to align manually the microhub 11 with a remote terminal 9.
In this embodiment, in order to align the microhub 11 with a remote terminal 9, a sighting device (not shown) is attached to the microhub 11. The installer is then able to use the sighting device to align the microhub 11
with the remote terminal 9 by eye.
Once the microhub 11 has been aligned, the screw 215 is fully tightened which causes the bearing surfaces 207 to grip firmly the mounting rod, so that the clamping device 203 is unable to slide along the axis of the mounting rod or to rotate about the axis of the mounting rod, and the circumferential surface of each jaw portion 221 grips firmly the inner circumferential surface of the ring 223, so that the microhub 11 is unable to rotate about the axis of the clamping device 203. In this way, a fixing mechanism is provided which fixes the orientation of the microhub 11 by movement of a single actuator (i.e. the screw 215) .
Figure 18 shows eight microhubs ll! to 118 mounted to a support structure 233. The support structure 233 is formed by a base member 235 and a top member 237 separated by upstanding support members 239a to 239c to form a partial enclosure. The support structure 233 defines two windows through which light beams from the user stations 9 pass and are incident on their respective microhubs 11, and the reflected light beams from the microhubs 11 pass on the way back to respective user stations 9. In this embodiment, the base member 235 and the top member 237 include hinged portions via which one of the microhubs 11 can be manually adjusted without disturbing the optical beams communicating with the other microhubs 11.
Four mounting rods 241a to 24Id are mounted within the
partial enclosure, with each mounting rod 241 supporting two microhubs 11. In particular, the four mounting rods 241 are configured as two pairs of two parallel mounting rods 241 with each pair of mounting rods 241 being separated by respective spacer elements 243a to 243c. As shown in Figure 17, each pair of mounting rods 241 is aligned approximately with the centre of one of the windows defined by the support structure 233. In this way, four microhubs 11 communicate via each of the two windows .
The mounting rods 241 are arranged approximately horizontally with each pair of mounting rods 241 being positioned one above the other. For each pair of mounting rods 241, the microhubs 11 supported by the lower mounting rod 241 hang below the lower mounting rod 241, while the microhubs 11 mounted to the upper mounting rod 241 are supported above the upper mounting rod 241.
The housing 201 of the microhubs 11, the ring 223, the first and second clamping pieces 205 and the mounting rods 241 are fabricated from aluminium. In this way, a heat conduction path is formed from the microhub 11 to the support structure 233 so that the support structure 233 acts as a heat sink. In this embodiment, the mounted rods 241 are solid cylinders in order to improve this heat sinking capability.
The clamping device 203 of the first embodiment has two main advantages. Firstly, the clamping device 203 allows a microhub 11 to be mounted to a mounting rod 241 without
disturbing the alignment of any previously mounted microhub 11. This is achieved by having first and second clamp pieces 205a, 205b which define an aperture for receiving the mounting rod 241 and are movable relative to each other to generate a gap through which the mounting rod 241 can enter the aperture in a radial fashion. Secondly, the clamping device 203 of the first embodiment has a single fixing mechanism for fixing both the orientation of the clamping device 203 relative to the mounting rod 241 and the orientation of the microhub 11 relative to the clamping device 203. This facilitates the installation of the microhub 11.
A second embodiment will now be described with reference to Figures 19 to 21 in which the clamping device 203 of the first embodiment is replaced by an alternative clamping device 261 which also exhibits the two advantages outlined above. Figure 19 shows a perspective view of the clamping device 261 of the second embodiment, while Figures 20 and 21 show cross-sectional views of the clamping device 261 in a clamped configuration and a mounting configuration respectively. Apart from the clamping device 261, all the remaining components of the second embodiment are identical to the first embodiment and will not therefore be described again.
As shown in Figure 19, the clamping device 261 includes a lower part 263 and an upper part 265 which, in the clamped configuration, form a generally rectangular box. The lower part 263 includes a downwardly facing (in the orientation shown in Figures 19 to 21) cylindrical recess
267 and an upwardly facing part-cylindrical transverse bearing surface 269. At one end of the lower part 263, an upstanding- wall 271 is formed having a transverse opening 273 for receiving a tongue 275 which forms one end of the upper part 265.
The upper part 265 includes a downwardly facing part- cylindrical transverse bearing surface 277. As shown in Figure 20, in the clamped configuration the upwardly facing bearing surface 269 of the lower part 263 and the downwardly facing bearing surface 277 of the upper part 265 are aligned and define a generally cylindrical passage for receiving a circular mounting rod 279 (indicated by dashed lines in Figure 20).
As shown in Figures 20 and 21, the housing 201 includes a hollow cylindrical projection 281 formed by a distal portion having a flanged base portion. The cylindrical recess 267 in the lower part 263 is shaped to receive the distal portion of the hollow cylindrical projection 281. A first circular hole is formed through the top of the cylindrical projection 281, with the first circular hole centred on the axis of the cylindrical projection 281. The diameter of the first circular hole is less than the diameter of the cylindrical projection 281 so that a downwardly facing annular surface 283 is formed within the cylindrical projection 281. A second circular hole, having the same diameter as the first circular hole, is formed through the lower part 263 along the axis of the cylindrical recess 267 so that when the cylindrical projection 281 is inserted in the cylindrical recess 267,
the first circular hole is aligned with the second circular hole. As shown in Figure 20, in the clamped configuration the first and second circular holes are aligned with one end of a slot 285 formed along the majority of the length of the upper part 265.
A clamping pin 287 passes through the slot 285 in the upper part 265 and through the first and second circular holes so that one end of the clamping pin 287, hereafter called the first end 289, is within the hollow cylindrical projection 281 and the other end of the clamping pin 287, hereafter called the second end 291, protrudes through the upper part 265. A flange 293 is provided at the first end 289 of the clamping pin 287, and a low rate spring washer 295 and two high rate spring washers 297a, 297b are positioned between the flange 293 and the annular surface 285 of the cylindrical projection 281.
A clamping cam 299 is rotatably mounted, using a pivot pin 301, to the second end 291 of the clamping pin 287 so that rotation of the clamping cam 299 causes the clamping pin 287 to move up and down along the axis of the cylindrical projection 281. In particular, the clamping cam 299 has a cam surface which abuts a planar protective pad 307 fixed to the upper part 265, the cam surface having three substantially planar portions 303a, 303b and 303c so that the clamping pin 287 has three stable positions. The clamping cam 299 also has a splined portion 305 which facilitates rotation of the clamping cam 299 by a ratchet (not shown).
In the clamping configuration, as shown in Figure 20, the first planar portion 303a of the cam surface contacts the pad 307 and the clamping pin 287 is at the uppermost position in its range of travel. The low rate spring washer 295 and the two high rate spring washers 297 are compressed by the flange 293 to provide a biassing force which presses the first planar portion 303a of the cam surface firmly against the pad 307, causing the mounting rod 279 to be firmly gripped between the bearing surfaces 269, 277 of the lower part 263 and the upper part 277, and causing the lower part 263 to be firmly fixed relative to the cylindrical projection 281 due to friction between the lower part 263 and the flanged base portion of the cylindrical projection 281.
In an alignment configuration, the second planar portion 303b of the cam surface contacts the pad 307 and the clamping pin is in the middle of its range of travel. The flange 293 compresses the low rate spring washer 295, but not the high rate spring washers 297, to provide a biassing force which loosely presses the second planar portion 303b of the cam surface against the pad 307, causing the lower part 263 to be loosely fixed relative to the cylindrical projection 281 and the mounting rod 277 to be loosely held between the bearing surfaces 269, 277 of the lower part 263 and the upper part 265. In this way, in the alignment configuration the microhub 11 is alignable with a remote terminal 9 by rotating the clamping device 261 about the axis of the mounting rod 279 and by rotating the microhub 11 about the axis of the cylindrical projection 281.
In the mounting configuration, as shown in Figure 21, the third planar portion 303c of the cam surface is adjacent the pad 307 and the clamping pin 287 is at the lowermost position of its range of travel. Neither the low rate spring washer 295 nor the two high rate spring washers 297 are compressed. In the mounting configuration, the upper part 265 is movable away from the upstanding wall 271 of the lower part 263, with the clamping pin 287 remaining within the slot 285, to reveal the upwardly facing bearing surface 269 of the lower part 263. In this way, in the mounting configuration a mounting rod 279 is radially insertable into the clamping device 261.
As described in the first embodiment, the RC time constant of the modulator drive circuit and the modulator limits the speed at which the electric field applied across the multiple quantum wells can be switched, and therefore limits the data transfer rate. In the first embodiment, the RC time constant is reduced by using plural CMOS drivers connected in parallel to drive a single modulator element. In a third embodiment, the RC time constant is reduced by separating the modulator into a plurality of smaller modulator elements which are separately driven. Apart from the modulator and the modulator drive circuit, the remaining components of the third embodiment are identical to the corresponding components of the first embodiment and will not therefore be described again.
In the third embodiment, the multi-layer semiconductor structure of the first embodiment is separated into eight
different modulator elements by etching eight trenches through the p-conductivity type layer 103-1 and the intrinsic layer 103-2 so that they are electrically isolated. The eight*'" different modulators are respectively connected to eight CMOS drivers whose inputs are connected together to receive a common drive signal. By separating the modulator into eight separate modulator elements, for a fixed driver output impedance and a fixed effective series resistance of the modulator, the electric field across each modulator element can be switched approximately eight times faster because the capacitance of each modulator element is reduced by a factor of approximately eight.
MODIFICATIONS AND FURTHER EMBODIMENTS
In the first to third embodiments, the modulator includes a mesa-structure generally having the shape of an octagonal right prism. Other shapes of mesa-structure could be used, for example a cylinder with a circular cross-section. However, an octagonal cross-section is preferred because the mesa-structure is easier to manufacture compared with a circular cross-section, but still has an approximately circular shape so that the field of view of the telecentric lens maps efficiently onto the top surface of the mesa-structure. It is important that the field of view of the telecentric lens maps efficiently onto the active area because any portion of the active area which is not mapped to the- field of view serves no useful purpose but still has an associated
capacitance which slows the switching speed of the electric field across the modulator.
In the first embodiment, a plurality of parallel strip conductors traverse the active area to reduce the effective series resistance of the modulator. Those skilled in the art will appreciate that the exact layout of the strip conductors is not critical, and other strip conductor layouts are possible. Figure 22 shows an example of an alternative layout of the strip conductors of the second electrode. As shown, a first C-shaped conductor 311a has nine strip conductors 313_1 to 313_9 extending therefrom, and a second C-shaped conductor 311b has a further nine strip conductors 313__10 to 313_18 extending therefrom. The eighteen strip conductors 313 form an interdigitised pattern with each strip conductor 313 extending from the C-shaped conductor 311 to which it is connected to a point close to the other C-shaped conductor 311. Those skilled in the art will appreciate that the strip conductors need not be formed as rectilinear lines.
The modulator described in the first embodiment includes Gallium Arsenide (GaAs) and aluminium Gallium Arsenide (AlGaAs) layers. Those skilled in the art will appreciate that other semiconductor materials could be used. For example, other III-V semiconductor materials such as Indium Gallium Arsenide (InGaAs) could be used. Further, QCSE modulators have also been fabricated using II-VI semiconductor materials.
In the first embodiment, CMOS line drivers are used to switch the electric field across the multiple quantum wells of the modulator. It will be appreciated that other forms of push-pull driver could be used. For example, instead of using a p-channel MOSFET and an n- channel MOSFET, complementary bipolar transistors could be used. Those skilled in the art will also appreciate that the line drivers will typically include other circuitry, such as Schmidt triggers, to improve their performance.
Those skilled in the art will appreciate that many different types of detector could be used in the microhub and the user terminal. For example, a phototransistor could be used. In general, a detector is selected based on the required bit rate and sensitivity, the wavelength of the light beam, and cost considerations.
In the first embodiment, an integrated semiconductor device having a QCSE modulator formed on a Bragg mirror is used. As those skilled in the art will appreciate, other types of reflectors and modulators could be used. For example, a plane mirror may be used as the reflector and a transmissive modulator (such as liquid crystal) could be provided between the telecentric lens and the mirror. Further, those skilled in the art will appreciate that the retro-reflector need not be a telecentric lens. Alternatively, the retro-reflector could be formed by a corner-cube reflector or a cat's-eye reflector. However, use of a telecentric • lens is preferred due to the ability to adjust the divergence of
the retro-reflected beam.
In the first embodiment, a point-to-multipoint signalling system is connected to an Ethernet network. Those skilled in the art will appreciate that the point-to- multipoint signalling system could be connected to different types of network. For example, the point-to- multipoint signalling system could be used in a STM-1 telecommunications network operating at 155Mbits/s.
It is preferred that the field of view of the modulator units is conical with a half-angle between 0.5° and 2.5°. This enables the telecentric lens to be relatively cheaply manufactured with a large collection aperture, while still not requiring precision alignment with a user terminal .
The modulator units of the described embodiments have a telecentric lens and an uplink detection lens which are positioned side by side, and the user terminals have a collimating lens and a downlink detection lens which are positioned side by side. For such an arrangement, the divergence of the light beam travelling in each direction between a user terminal and a corresponding modulator unit must be sufficiently large so that the light beam expands to encompass both lenses of the receiving device. Alternatively, as described in International Patent Application WO 02/05459 (whose contents are hereby incorporated by reference) , a periscope arrangement could be employed to align optically the detector and the light source of the remote terminal, and/or to align optically
the modulator and the detector of the microhub.
In the first embodiment, the optical links between the microhubs and the user terminals act as data pipes. Alternatively, prior to transmission over the optical link data could be protocol encoded to include, for example, error detection and correction bits. Further, the OAM data could be multiplexed with the uplink data. In the first to third embodiments, full duplex transmission systems are described. Alternatively, a simplex transmission system, in which an unmodulated light beam is sent to a retro-reflector where it is modulated and reflected back to be detected by a detector. Alternatively, a half-duplex system could be used in which either the remote terminal sends an unmodulated light beam to the local distribution node where it is modulated and reflected back to the remote terminal, or a modulated light beam is emitted by the remote terminal to convey data to the local distribution node, could be used.
In the first and second embodiments, the mounting rod has a circular cross-section. In alternative embodiments, the mounting rod has a cross-section with a different shape, for example a polygon. If the cross-section of the mounting rod is a polygon, then by connecting the clamping device to the microhub via a ball-and-socket joint the microhub is rotatable about an axis parallel with the axis of the mounting rod as well as an axis perpendicular to the axis of the mounting rod. '
Those skilled in the art will appreciate that the mounting rod need not have a constant cross-section, and generally any form of elongate mounting member could be used. For example, an elongate mounting member could be used in which the cross-section varies to form shoulders for restricting movement of the microhub along the longitudinal axis of the mounting member.
In the first embodiment, the clamping device and the mounting rod are made of aluminium so that they act as a heat sink to remove heat generated by the microhub. Those skilled in the art will appreciate that other heat conductive materials could be used to make either or both of the clamping device and the mounting rod, for example brass.
In the described embodiments, the support structure includes two pairs of parallel mounting rods . Other arrangements of mounting rods could be used. For example, a single mounting rod could be used.
Those skilled in the art will appreciate that the support structure of the described embodiments can be adapted to encompass a 360° field of view. In particular, windows could be defined by the upstanding support members of the support structure through which a 360° field of view can be accessed.
In the first embodiment, an optical sight is attached to the microhub in order to align the microhub with a remote terminal. In an alternative embodiment, the microhub is
03 00951
54 detachably mountable, via a kinematic mount, to the clamping device. During alignment, the microhub is detached from the clamping device and replaced by a camera. The, camera is designed so that the field of view of the camera when mounted to the kinematic mount matches the field of view of the microhub when mounted to the kinematic mount. Alignment of the mounting device is performed using the camera and a display via which the installer can view the field of view -of the camera and position a remote terminal within the field of view. The camera is detached and the microhub is replaced. Alternatively, instead of a camera an alternative sighting device could be mounted, via the kinematic mount, to the clamping device.
The described support structures and clamping device could be used for mounting optical devices which do not include a reflective modulator. For example, the described support structure and mounting devices could be used to support a plurality of optical devices which each include a light source for communicating with a remote device having a detector.
Those skilled in the art will appreciate that the term "light" includes electromagnetic waves in the ultraviolet and infra-red regions of the electromagnetic spectrum as well as the visible region. Although the embodiments described above have used laser beams with a wavelength of about 785nm, other light beams could be used. In particular, a wavelength of about 1.5 microns is an attractive alternative because it is inherently
more eye-safe and emitters and detectors have been developed for this wavelength for optical fibre communications .
Although the lenses in the microhubs and the remote terminals have been schematically represented by a single lens, it will be appreciated that in practice each lens may have a plurality of lens elements .