SPECIFICATIONA telecommunications system employing optical signals for transmissionThis invention relates to a telecommunication system which employs optical signals for transmission and which can provide a broad transmission band to enable diverse services such as video and hi-fidelity sound to be transmitted with data and conventional telephonic speech.
The system may find use in a local network or a junction network, that is between a local exchange and a subscriber or group of subscribers, or between two local eh changes or a local exchange and a trunk exchange, where costs are allocated in a greater proportion per customer, and hence a simple and economic system is necessary.
Digital techniques are attractive because of their compatibility with a digital telecommunications network, but the cost of the necessary codecs and multiplexing equipment for video signals may be prohibitive.
The basic analogue approach is direct intensity modulation which offers the desired simplicity and economy, but only at the expense of transmission range.
The present invention relies on the use of PulseTime Modulation techniques and preferred embodiments of the invention offer economical implementation.
According to the present invention, there is provided a transmission system utilising optical signals transmitted along optical fibres, wherein the required base band is broad and the transmitted waveform comprises pulses modulated in time by the base band signal.
The pulses are preferably frequency modulated by the base band signal.
The base band may be of the order of 1OMHz making possible the transmission of an 8M bits/s data stream either in ternary or binary form.
The form of a pulse time modulation technique implies that the transmission capacity of the system for a data steam is high, and hence the system may be modified to remove the modem components, if for example, the network in which it was located changed use. This gives great flexibility to network planners. Indeed the circuitry with modems may be advantageous used ab initio in an appropriate system of data stream transmission.
The system may, in one application, be used to transmit a broad band video signal together with one or more narrow band services.
In one embodiment, the transmitted PFM signal includes no base band component, the base band signals being regenerated in the receiver.
In this embodiment of the invention a preferred form of receiver may be provided including as a demodulator a monostable circuit and a low pass filter.
The monostable circuit may comprise a delay line, the signals along which are gated with those along a direct path to produce constant width pulses from the received PFM pulses.
The delay line and gate element may compriseNOR gates.
A preferred form of transmitter for the regenerated base band embodiment of the invention includes a voltage controlled multivibrator as the modulator, and this, in a preferred form, comprises a single integrated circuit.
Embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawings.
In the drawings: Figure lisa block diagram of a transmitter for aPFM regenerated base band system;Figure 2 is a block diagram of a receiverforthe signals produced by the transmitter of Figure 1;Figure 3 is a circuit diagram of the transmitter ofFigure 1;Figure 4 is a circuit diagram of the optical receiver, amplifier and equaliser of the receiver of Figure 2;Figure 5 is a circuit diagram of the threshold detector and demodulator components of the receiver;Figure 6 is a circuit diagram of an output amplifier of the receiver; andFigure 7 is a block diagram of a receiver and transmitterfora direct base band PFM transmission system.
Referring to Figure la a broad band transmission system of optical signals is based on a pulse frequency modulation technique using regenerated base band recovery. The base band input terminal 1 in the transmitter is connected to a pre-emphasis network 2 which introduces attenuation to certain components of the base band signal, particularly video test signals. Insertion loss resulting from the pre-emphasis network 2 is compensated by a preemphasis amplifier 3 at its output. The resultant output which comprises a pre-emphasised base band signal is passed to an input terminal of a modulator 4 which comprises a voltage controlled multivibrator.
A low-pass filter may be inserted between the amplifier 3 and the modulator 4 if required. The mod u lator 4 comprises a voltage controlled multivibrator (VCM) on a single ECL chip which generates a pulse train of fixed duty cycle (50%). The repetition rate of this pulse train is linearly varied in accordance with the input base band signal voltage. As an example, a repetition rate of 30 MHz with a 10 MHz deviation has been found acceptable. A driver 6 for the optical source 7 comprises a voltage to current converter and drives an optical source such as an edgeemitting LED. The resultant signal is then transmitted along an optical fibre.
The components of the transmitter will now be described in more detail with reference to Figure 3.
The output voltage from the pre-emphasis amplifier 3 is applied to an input terminal 8 and then to the multivibrator4. The repetition rate is set bytwo adjustments, VC1 and VR1. A 1K ohm resistor 9 is included to avoid any DC loading at the input to the multivibrator. The frequency deviation is set by the amplitude ofthe modulating signal itself; and a 0.56Vp-p swing applied to the control pin 2 of the chip has been found suitable to give a 1OMHz deviation about the centre frequency.
To set up the modulator, resistor VR1 is varied to present a mid-range voltage to the multivibrator.
The repetition rate is then set to 30MHz by the variable capacitor VCl. The voltage controlled multivibrator has two output, only one of which needs to beused. However, the unused output must be correctlybiased and terminated to avoid mismatch reflectionsback into the device. A resistor 10 (510 ohms) connected to terminal 6 provides the DC bias, while a de-coupled resistor 11 (51 ohms) provides the cor recttransmission termination. A reactive flexibility point 1 la may be provided if very small reactivemismatches are required to be adjusted to maintain transfer linearity.
The bias and termination technique described above with reference to the unused output from the multivibrator is appropriate only because of the very short transmission line from that terminal. The used outout from the multivibrator is fed via a resistive divider 12 whose series resistors give the correct DC bias, while their parallel value gives the correct termination.
A VMOS-FET 13 provides the modulated current to drive the optical source 7. For most optical sources, and with the voltage levels used in this system the current provided by the FET would be excessive and hence a PNP transister 14 is introduced to act as a current drain to reduce the absolute current levels provided to the optical source. The degree of currentbleed off is set by the division chain comprising theresistors 15 and 16.
Referring now to Figure 2, the receiving system isillustrated and its main components will now bedescribed with reference to that figure. An opticalreceiver 20 comprises a wide band PIN-FET packagedesigned for operation at 140 M bits/s. This packagehas bias control and will be described in detailbelow.
The output from the optical receiver is transmitted via a frequency compensator 21 to the input terminal of a wide-band amplifier 22. After equalisation through an equaliser 23 and subsequent detection by a threshold detector 24 the output electrical signal is substantially the same as that obtained directly at the output of the modulator in the transmitter, except that the receiver noise is now superimposed on the pulse chain in the form of pulse edge jitter.
This jitter constitutes phase errorwhich when de-modulated becomes base band noise.
The de-modulator in the receiver utilises a pulse counting technique. The fixed 50% duty cycle from the threshold detector 24 are converted to a pulse chain of fixed pulse width by an ECL monostable, incorporating a delay element to set the pulse width at longs. The resultant signal has a frequency spectrum which includes a base band signal and this is then filtered off by a low pass filter 26. The resultant signal is fed via a de-emphasis network 27 and an output amplifier 28 to give the base band output.
Referring now to Figure 4the optical receiver comprises a PIN photo detector 30, a GaAs FET31 and two bipolar transistor stages 32 and 33, one providing gain and the other acting as a buffer.
With the FETused as an example, its characteristics dictate that for lowest noise operation the source/drain current through this device should be about lama. Since the value ofthis current is influenced by the mean photo-currents through the loadresistance 34(100 M ohms), the FET bias is placedunderthe control of a bias control feedback loop constituted by a 741 operational amplifier 35. Given the correct FET current, the source potential is + 1 volts as a result of a resistor 36 (100 ohms) at pin 2 of the FET package. Consequently an AGC reference voltage of + 1 volts is used, and this is set by a variable resistor 37 at the non-inverting inputtermi- nal to the operational amplifier 35.The bias controlloop is in equilibrium when the differential voltages presented to the input to the operational amplifier 35 are equal. Afeedback resistor 38(2.2 M ohms) connected to the inverting input of the operational amplifier 35 ensures a high feedback gain.
The bias for the bipolartransistor stages 32 and 33 is provided at two positions; the fixed potential at pins 11 and 20 of the FET package and the variable current at pin 8. A variable resistor 39 is adjusted until the DC output at pin 9 is about + 8 volts.
Various capacitors of 10 mF and 0.1 MF are connected at the packing pins to decouple all the DC voltage and bias control points to prevent internal oscillations. In addition the photo-diode bias and package power rail are derived from LC low-pass filters to ensure that the voltage sources are as noise free as possible.
A CR time constant exists at the photo detection stage due to the 100 M ohm load resistor 34 and the finite FET input capacitance of 0.3pF. These figures constitute a first order low-pass response with a break point at about 5kHz. To achieve the required wide band response from the reception package it is necessary to follow it with the frequency compensator 21. This compensator 21 comprises a highpass circuit exhibiting two break points, one coincident with that of the optical receiver package and the other set to the upper limit of the required wide band width. It is not necessary to provide a DC signal path since the received frequency modulated pulse train contains no base band component component. In fact it is advantageous for receiver noise performance to limit the lower frequency of the pass band of the package as much as can be accepted.Consequently, the frequency compensator 21 need only exhibit the upper break point, i.e. a first order high pass filter. This function can be performed by a 22 pF capacitor 40 in series with the signal path to provide a compensated receiver response from 5MHz to about 100MHz.
The wide band amplifier 22 comprises two cascaded AVANTEK amplifier packages 41 and 42.
These provide about 20 dB gain and their 5-200MHz frequency response introduces further low frequency attenuation in the signal path, reducing low frequency noise still further. It may be found desirable to include automatic gain control at this stage to ensure constant pulse amplitude.
The equaliser 23 comprises a 50 ohm 3rd order, maximally flat, Butterworth low-pass filter.
Referring now to Figure 5, the threshold detector 24 receives a 200 to 300mVp-p amplitude pulse train from the equaliser 23 which amplitude is sufficient to trigger the threshold detector 24. The threshold reference voltage is derived from two 3.6V zener diodes 50 and 51 producing a high degree of threshold stability. A voltage divider chain 52 and a 100 ohm potentiometer 53 are included to provide a fine adjustment facility.
The output from the threshold detector comprise aPFM pulse train virtually identical with that from the transmitter modulator except with the addition of pulse edge jitter.
The de-modulator utilises a pulse counting technique and is based on a monostable 25 comprising four NOR gates 54, 55, 56, and 57. The three gates 54-56 act as a delay line to give a delay of 1 Ons which sets the pulse width also to longs. The NOR gate 57 acts as a summer of the delayed signal and the original signal to produce a train of pulses of fixed pulse width. The individual gates are biased and terminated by 510 ohm and 51 ohm resistors respectively. The frequency spectrum of the waveform at the output of the monostable 25 contains a base band signal which is then filtered off to re-constitute the original base band signal. For video applications it is necessary to provide a DC transmission termination, so a resistive divider is used at the monostable output.However, instead of the previous 82 ohms toO volts arrangement the 82 ohm resistence is approximately derived from a 30 ohm series resistor 58 and 50 ohms provided by the following low-pass filter 26. The low-pass filter 26 is a 5th orderTchebychev 1 dB ripple design with a break point set to about 10 MHz.
For a 0.56 Vp-p base band signal applied to the modulator, giving the full 10MHz deviation, the corresponding amplitude of the base band signal from the de-modulator low-pass filter is about 80mVp-p.
To comply with standard video practice, sufficient base band gain must be introduced to produce an output signal of 1Vp-p into a 75 ohm resistor (sync- peak white). For system flexibility cascaded video amplifiers are used, each amplifier being capable of up to 14 dB gain with a 1OMHz band width.
Figure 6 shows a one such amplifier in which the gain is provided by two cascaded differential stages; the first of which utilises a matched pair of FETS 61, 62 to provide both a high impedence and a low noise front end. An output transistor 63 acts as a linear buffer to give sufficient current to drive a 75 ohm load at upto 2 volts p-p if so desired. The gain of the amplifier is set by the degree of feedback provided at two positions by a resistor 64, a resistor 65, the capacitance of a trimmer 66 (30pf), and the resistor 67.
The circuitry described above may be assembled on two boards of size lOOmm x 200 mm one each for the transmitter and receiver. This is a great simplification as compared with alternative approaches, and yet a very high performance may still be obtained. A minimum detectable power of-47 dBm can be achieved with a signal to unified weighted noise of 53dB. Assuming a launch power of O dBm, this means that a broadcast quality of colour vision channel could be supported over 7km using nominally 5dB/Km fibre with a reasonable system margin and allowance for connectors.
An additional significant attraction of the system is that is may be used forthe direct transmission of a range of independent services. For example, it could be used to transmit video with associated facilities, or CCITT 2Mbit/s or 8Mbit/s. Such flexibility is unusual and of great value to a network planner.
Furthermore the use of a 30MHz pulse repetition rate implies that the transmission capacity of the electrical/optical and optical/electrical translating elements could support a 34 Mbits/s data stream. This would entail the removal of the VCM at the transmitter and the demodulator at the receiver, and would require a modified equaliser and the inclusion of a 34Mbits/s regenerator. Nevertheless, this adaptability is economical compared with totai replacement.
Whilst the above described system relies on the regeneration of the base band signal from atrans- mitted PFM signal with no base band component, an alternative system comprises the direct transmission ofthe base band signal in the PFM transmission.
Figure 7 is a block diagram of such a system and it will be seen that the only major difference in the transmitter compared with that of Figure 1 is the inclusion of a monostable. This monostable may be of the delay line form as in the receiver of Figure 2 and produces constant width pulses for transmission. The difference between the receiver of this system and that in Figure 2 is more marked in that it requires a low band width detector, a narrow band video amplifier and a low-pass filter. The AGC circuit should however have a wide band width.