Patent Application
20080225381
1. Field of the Invention
The present invention relates to apparatus and methods for adjusting constructive and destructive transfer functions of a differentially encoded phase shift keyed receiver for reducing inter-symbol interference in optical systems.
2. Description of the Prior Art
For an optical system with filters, the effective concatenated bandwidth of the filters induces intersymbol interference (ISI). The ISI causes distortion of the signal and reduces the decision quality (the ability to accurately detect if a bit is a logical “1” or “0”) at a receiver. This decision quality may be quantified by counting the number of error bits and dividing it by the total number of transmitted bits. The resulting ratio is called bit error ratio (BER). Another way of discussing the quality of the signal at the receiver involves translating the BER to a parameter called Q using the equation Q=20 log└√{square root over (2)}erfc−1(2BER)┘ where erfc−1 is the inverse complementary error function. The distortion effect of ISI on signal quality may be viewed in a general way in a baseband eye diagram of the modulated signal where ISI causes the space between “1” and “0” symbol levels to be partially filled by the trailing and leading edges of the symbols.
Optical signals commonly use binary phase shift keyed (BPSK) modulation where a carrier is modulated for data bits for logical “0” and “1” with phase shifts of 0 and π radians. The logical “0” or “1” is decoded at the receiver by determining whether the detected signal is to the left or right of a vertical imaginary axis of a signal vector diagram, sometimes called an IQ diagram. A detector viewed as a polar detector determines whether the absolute value of the received phase is greater than π/2 for “0” and less than π/2 for “1”. A detector viewed as a rectangular detector determines whether the cosine of the phase of the signal is negative or positive for “0” or “1”.
The BPSK optical signals may use a differentially-encoded phase shift keyed (DeBPSK, or DPSK) modulation format. The DPSK modulation format encodes input data as the difference between two consecutive transmitted symbols. The input data is differentially pre-coded using the preceding symbol as a reference with an electrical “delay+add” function so that an input data bit of logical “0” or “1” is encoded as a change of carrier phase of 0 or π radians relative to the preceding bit. At the detector the process is reversed by comparing a current bit to the preceding bit.
The DPSK decoding function may be performed using a delay line interferometer (DLI) and a balanced detector. The interferometer works on the principle that two waves that coincide with the same phase will add to each other while two waves that have opposite phases will tend to cancel each other. The interferometer has an input port for receiving the optical signal and two output ports—a constructive output port for issuing the waves that add and a destructive output for issuing the waves that tend to cancel.
The delay line interferometer (DLI) for DPSK signals has an additional element of an internal delay difference between the two waves that is about equal to the symbol time T of the DPSK modulation. The constructive output port issues a signal Ec=E(t)+E(t−T) and the destructive output port issues a signal Ed=E(t)−E(t−T). The effect of the time T is to reverse the signals at the two output ports so that the waves add at the destructive output port and cancel at the constructive output port when consecutive bits differ by π radians. The difference between Ec and Ed can be detected with a direct detection intensity receiver to determine when there is a change in phase in the signal between two consecutive bits and thereby estimate the logical bits carried by the DPSK modulation.
It is an effect of this delay difference to impose a transfer function having a sinusoidal amplitude response (in the frequency domain) from the input port to each output port. The spectral period of a cycle of the transfer function, equal to 1/T, is termed the free spectral range (FSR). The sinusoidal width proportional to the FSR effectively limits the frequency band of the signals that can be passed from the DLI input to the constructive and destructive outputs. The phase of the frequency domain cycle of the transfer function is termed the FSR phase.
It is commonly believed that a DLI delay difference equal to the symbol time T, and an FSR equal to the inverse of the symbol time T, is desired in order to provide the best system performance (fewest data estimation errors) by maximizing the difference between the signals Ec and Ed at the constructive and destructive outputs. Considered by itself, a differential delay not equal to the symbol time T would be expected to degrade system performance because the current and preceding symbols are not exactly differentially compared.
The present invention provides an optical receiver and methods for mitigating intersymbol interference (ISI) in a differentially-encoded modulation transmission system by controlling constructive and destructive transfer functions.
Briefly, an optical receiver of the present invention includes a signal processor having constructive and destructive transfer functions for receiving and demodulating an optical signal having differential modulation. In a preferred embodiment the signal processor includes a delay line interferometer (DLI), a free spectral range (FSR) phase controller, and a gain imbalancer. The DLI has a transit time difference Y between two signal paths for demodulating the differential modulation signal and defining a free spectral range (FSR) bandwidth of constructive and destructive transfer functions. The FSR is calculated or adjusted so that the performance benefit obtained by controlling the transfer functions for reducing ISI distortion is greater than the performance that is lost by not maximizing the demodulated signals at constructive and destructive outputs when the time difference Y is not equal to the symbol time of the modulated signal. The FSR phase controller adjusts the phases of the constructive and destructive transfer functions to tune the FSR transfer functions relative to the carrier of the modulated optical signal. The gain imbalancer applies a calculated or adjusted unequal gain to the signals in the constructive and destructive paths for determining or modifying the constructive and destructive transfer functions.
In a preferred embodiment, the present invention is a delay line interferometer for differentially demodulating an optical input signal, comprising: an optical splitter for splitting the input signal into two signal paths having a transit time difference for providing a differentially demodulated signal to at least one of constructive and destructive outputs; a positionable delay element for delaying a signal along a first direction in one of the signal paths with a selected optical delay, the optical delay selected according to a position of the delay element in a second direction; and a positioning device for positioning the delay element in the second direction for controlling the transit time difference.
In another preferred embodiment, the present invention is a method in a delay line interferometer for differentially demodulating an optical input signal, comprising: splitting the input signal into two signal paths having a transit time difference for providing a differentially demodulated signal to at least one of constructive and destructive outputs; delaying a signal traversing a positionable delay element along a first direction in one of the signal paths with a selected optical delay dependent on a position of the delay element in a second direction; and positioning the delay element in the second direction for controlling the transit time difference.
In another preferred embodiment, the present invention is a delay line interferometer for differentially demodulating an optical input signal, comprising: an optical splitter for splitting the input signal into two signal paths having a transit time difference for providing a differentially demodulated signal to at least one of constructive and destructive outputs; a movable mirror for reflecting a signal in one of the signal paths; and a positioning device for positioning the mirror to a selectable position for controlling the transit time difference.
In another preferred embodiment, the present invention is a method in a delay line interferometer for differentially demodulating an optical input signal, comprising: splitting the input signal into two signal paths having a transit time difference for providing a differentially demodulated signal to at least one of constructive and destructive outputs; reflecting a signal in one of the signal paths with a movable mirror; and positioning the mirror to a selectable position for controlling the transit time difference.
In a preferred embodiment, the present invention is an optical receiver, comprising: a signal processor having constructive and destructive transfer functions for receiving a modulated optical input signal and issuing signals at constructive and destructive outputs, respectively; at least one transfer phase element disposed in the signal processor, the transfer phase element for providing a controllable transfer function phase for at least one of the transfer functions with respect to a frequency of the input signal; and a transfer phase controller coupled to the transfer phase element for controlling the transfer function phase for maximizing a difference between signal powers for the constructive and destructive outputs.
In another preferred embodiment, the present invention is a method for receiving an optical signal, comprising: applying constructive and destructive transfer functions to a modulated optical input signal for providing signals at constructive and destructive outputs, respectively, at least one of the transfer functions having a controllable transfer function phase; and controlling the transfer function phase with respect to a frequency of the optical signal for maximizing a difference between signal powers for the constructive and destructive outputs.
In another preferred embodiment, the present invention is an optical receiver, comprising: a signal processor having constructive and destructive transfer functions for processing a modulated optical input signal for providing signals at constructive and destructive outputs, respectively, at least one of the constructive and destructive transfer functions having a controllable bandwidth; and a bandwidth control element disposed in the signal processor for selecting the bandwidth based on an effective bandwidth of the input signal for compensating for signal impairments in the input signal.
In another preferred embodiment, the present invention is a method for receiving a modulated optical signal, comprising: processing a modulated optical input signal according to constructive and destructive transfer functions for issuing signals at constructive and destructive outputs, respectively, at least one of the constructive and destructive transfer functions having a controllable bandwidth; and controlling the bandwidth based on an effective bandwidth of the input signal for compensating for signal impairments in the input signal.
In another preferred embodiment, the present invention is an optical receiver for receiving a modulated optical signal, comprising: a signal processor for separating a modulated optical input signal into constructive and destructive signal paths; and an optical gain imbalancer disposed in at least one of the signal paths for selecting an optical gain imbalance between the signal paths based on an effective bandwidth of the input signal for compensating for signal impairments in the input signal.
In another preferred embodiment, the present invention is a method of receiving a modulated optical signal, comprising: separating a modulated optical input signal into optical constructive and destructive signal paths; and selecting an optical gain imbalance between the signal paths based on an effective bandwidth of the input signal for compensating for signal impairments in the input signal.
Various preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings.
The details of preferred embodiments and best mode for carrying out the ideas of the invention will now be presented. It should be understood that it is not necessary to employ all of the details of the preferred embodiments in order to carry out the idea of the invention. It should be further understood that the details of the preferred embodiments may be mixed and matched for carrying out the invention. Therefore, these details should be viewed for understanding the idea of the invention but should not to be read as limitations of the idea that is expressed in the below listed claims.
The preferred embodiments are described in terms of binary phase shift keyed (BPSK) signals using a differentially-encoded BPSK (DeBPSK, or DPSK) modulation format. However, the idea of the invention may be carried out with higher order modulation formats such as quadrature phase shift keyed (QPSK), 4QAM, 8PSK, 16QAM and so on. For example, the idea can be carried out with differentially-encoded QPSK (DQPSK) and so on.
The DLI has a transit time difference Y for demodulating a differentially modulated signal. The transit time difference Y (
G(f)=[1+cos(2πfY)]/2 (1
H(f)=[1−cos(2πfY)]/2 (2
It can be seen that the FSR transfer functions G(f) and H(f) are periodic in the frequency domain. Phase of the periodic transfer function (offset in the frequency domain) is known as an FSR phase. In an optical system using differential modulation, best signal quality may be obtained when the FSR phase is adjusted so that the transfer functions G(f) and H(f) have a maximum ratio or normalized difference (difference scaled by the sum) at the carrier frequency of the optical signal or the center of the energy in the spectrum of the modulated optical signal.
The transmitter 12 transmits an optical signal using a differentially-encoded phase shift keyed (DPSK) modulation format where logical 1's and 0's of input data are encoded as phase differences between adjacent (consecutive in time) phase states. For example for DPSK, adjacent phase states of 0 radians and adjacent phase states of π radians both carry a data bit having a logical “0”; and a phase state of 0 radians following a phase state of π radians and a phase state of π radians following a phase state of 0 radians both carry a data bit having a logical “1”. Of course, the logical “0” and logical “1” may be reversed without loss of generality. It should also be noted that any two phase states that are separated by π radians may be used for the DPSK modulation.
The transmitter 12 illuminates one end of the link 16 with a modulated optical signal 22 having differentially-encoded phase shift keyed (DPSK) modulation for the logical bits of input data. The signal 22 passes through the link 16 and emerges at the other end of the link 16 as a modulated optical signal 24 to be received by the receiver 20. The link 16 has a frequency response having an effective optical bandwidth caused by one or more filters represented by filters 26. The optical bandwidth of the link 16 results in an effective optical bandwidth of the spectrum of the input signal 24.
The receiver 20 demodulates the signal 24 for providing output data that is its best estimate of the input data. The output data is desired to be an exact replica of the input data. However, the transmission link 16 degrades or impairs the quality of the received signal 24 and this degradation or impairment in signal quality causes the receiver 20 to occasionally make errors in the output data that it provides. One of the primary causes of the signal degradation is intersymbol interference (ISI) in the link 16 induced by the filters 26. The receiver 20 of the present invention has apparatus and methods, described below, for compensating for the quality degradation in the link 16, especially the ISI, in order to reduce the errors in the output data.
The apparatus and methods of the receiver 20 use measurements of signal quality and calculations based on the effective optical bandwidth of the link 16 and/or the effective optical bandwidth of the input signal 24 for compensating for one or more signal degradations or impairments in the input signal that may include but are not limited to ISI, signal-dependent noise and signal independent noise. The signal quality measurements may be bit error ratio (BER) measurements or eye opening ratio measurements. In some cases the signal quality measurements may use signal-to-noise measurements taken from optical or electrical constructive and destructive path signals in the receiver 20. In a preferred embodiment, the receiver 20 uses calculations based on the effective optical bandwidth of the link 16 for minimizing the BER for the received input signal 24.
The receiver 20 includes a demodulator 30 and a data estimator 32. The receiver 20 or an external computer includes a bandwidth control algorithm 33. The demodulator 30 demodulates the optical input signal 24 and issues an electrical baseband signal. The data estimator 32 processes the baseband signal and issues the output data. The receiver 20 may also include an input optical filter for filtering the optical signal 24 into a channel when the optical signal 24 is wavelength division multiplexed (WDM) and contains multiple channels.
The demodulator 30 includes a signal processor 34, a detector apparatus 35, a combiner 36, and a transfer phase controller 37. The signal processor 34 has two parts, an optical signal processor 34A and an electrical signal processor 34B. The optical signal processor 34A receives the signal 24 at an input port 42; separates the signal 24 into optical constructive and destructive interference signals; differentially demodulates the signal 24 with a differential transit time Y; and issues the signals at constructive and destructive output ports 43A and 44A, respectively. The detector apparatus 35 receives the optical constructive and destructive paths signals from the ports 43A and 44A and converts photons to electrons for providing electrical constructive and destructive path signals shown as electrical currents iG and iH for the modulations on the optical signals.
The signal processor 34B processes the electrical signals and passes the processed electrical signals through constructive and destructive output ports 43B and 44B, respectively, to the combiner 36. The combiner 36 takes a difference between the instantaneous signal level of the constructive path signal and the instantaneous signal level of the destructive path signal for providing the baseband signal. In a variation of the receiver 20, the data estimator 32 connects to the ports 43B and 44B for receiving differential electrical signals.
The separation of the input signal 24 using optical interference into the constructive and destructive paths provides the constructive and destructive transfer functions G(f) and H(f), respectively, in the signal processor 34A. The transfer functions G(f) and H(f) are a part of the constructive and destructive transfer functions provided by the signal processor 34 and the detector apparatus 35 from the input port 42 to the constructive and destructive output ports 43B and 44B, respectively. However in one preferred embodiment the constructive and destructive transfer functions are primarily determined within the signal processor 34A to the output ports 43A and 44A.
The transfer phase controller 37 includes a detector 45 for measuring and averaging power-related levels for the signals at the output ports 43A and 44A (or 43B and 44B). The power-related levels that are measured are indicative of, or have a monotonic relationship to, the signal powers at the output ports 43A and 44A (or the output ports 43B and 44B). For example, the measurements may be signal power, average signal magnitude, squared signal level, or absolute value of signal level with an arbitrary exponent. The transfer phase controller 37 uses the measurements for providing a feedback signal that maximizes the ratio of the signal power for the port 43A to the signal power for the port 44A (or the signal power for the port 43B to the signal power for the port 44B). The idea may also be used in an inverted mode for maximizing the ratio of the signal power for the port 44A to the signal power for the port 43A (or the signal power for the port 44B to the signal power for the port 43B).
The signal processor 34A has controllable transfer phase elements 46G and 46H for providing adjustable phase shifts ΦG and ΦH for the constructive and destructive transfer functions. The elements 46G and 46H may be the same physical element 46 and the phase shifts ΦG and ΦH may be the same phase shift Φ. The transfer phase controller 37 uses the power-related measurements from the detection 45 for controlling the elements 46G and 46H, or the element 46, for adjusting the phases ΦG and ΦH, or the phase Φ, for shifting the phases of the transfer functions for a maximum normalized signal power difference between the signals at the constructive port 43A (or 43B) and the destructive port 44A (or 44B). This process may be used to tune the transfer functions G(f) and H(f) relative to the carrier frequency of the modulated optical signal 24 and at the center frequency of the energy in the modulated optical signal 24.
The signal processor 34A has a transfer bandwidth element 48 for providing a selectable or controllable bandwidth (BW). At least one of the constructive and destructive transfer functions depends, at least in part, upon this bandwidth. In a preferred embodiment the optical signal processor 34A includes a delay line interferometer (DLI). In this case the bandwidth is defined or modified by the inverse of the time Y.
During the design or installation of the receiver 20, or when the receiver 20 is in operation, a calculation or test is made, or active feedback is provided, for signal quality or a bit error ratio of the output data. A primary degradation of the signal quality in the system 10 is intersymbol interference (ISI) caused by the filters 26. The bandwidth control algorithm 33 calculates or provides feedback for determining or controlling the transfer bandwidth element 48 as shown in the chart of
An imbalance control algorithm 64 may be included for calculating a gain imbalance or providing feedback from signal quality data to the signal processor 34 to either the optical processor 34A or the electrical processor 34B or both for optimizing signal quality. The signal processor 34 uses the gain imbalance calculations or feedback to imbalance the gains between the constructive and destructive path signals. The gain imbalance calculations may be based on the effective optical bandwidth of the link 16 and the input signal 24.
A side effect of changing the selection of the transit time difference Y is that the transfer function phase or FSR phase of the transfer functions G(f) and H(f) may slide many cycles with respect to the frequency of the input signal 24. In a general rule, whenever the FSR delay is changed, the transfer function phase shift Φ, or phase shifts ΦG and ΦH, must be re-adjusted by the transfer (FSR) phase controller 37 by adjusting the transfer (FSR) phase element 46, or 46G and 46H, for re-centering the transfer functions G(f) and H(f) to its optimal frequency position. When the received optical spectrum is symmetrical, the optimal position coincides with the carrier frequency of the input optical signal 24. On the other hand the effect of changing the phase shift Φ, or phase shifts ΦG and ΦH, on the FSR bandwidth is so small that is insignificant.
The receiver 20 may also include a path for signal quality feedback 92. Data for signal quality is processed through the signal quality feedback 92 and passed to the transfer phase controller 37. The transfer phase controller 37 uses the processed signal quality data for fine tuning the phase delay of the transfer phase element 46 for improving and optimizing the signal quality. Preferably, the element 46 is first tuned in a feedback loop according to the power-related measurements and then fine tuned in a second feedback loop for minimizing a bit error ratio (BER). The signal quality data may be obtained by measuring BER directly, by measuring an eye opening ratio of a baseband signal, and/or by measuring a signal to noise ratio (SNR) of the optical or electrical constructive and destructive path signals.
The receiver 120 includes a demodulator 130, a data estimator 132 and a bit error ratio (BER) detector 138. The receiver 120, or an external computer, also includes a bandwidth (FSR) control algorithm 133, and an imbalance control algorithm 164. The demodulator 130 demodulates the optical signal 24 and passes the demodulated electrical signal to the data estimator 132. The data estimator 132 processes the electrical signal for making a best estimate of the original input data and issues its best estimated input data as output data. The BER detector 138 estimates a BER for the output data. The BER may be used as signal quality data. The demodulator 130 uses the signal quality data through the algorithms 133, 164 and 192.
The demodulator 130 includes a signal processor 134, a detector apparatus 135, a combiner 136 and a transfer free spectral range (FSR) phase controller 137. The signal processor 134 includes an optical signal processor 134A and an electrical signal processor 134B. The optical signal processor 134A receives the optical input signal 24 at an input signal port 142; separates the signal 24 into optical constructive and destructive interference signals; differentially demodulates the signal 24 with the differential time Y; and issues signals from constructive and destructive output ports 143A and 144A, respectively, to the detector apparatus 135.
The detector apparatus 135 converts the modulations on the optical constructive and destructive path signals to electrical signals and passes the electrical signals to the electrical signal processor 134B. The electrical signal processor 134B processes the electrical signals and issues the processed electrical signals at constructive and destructive output ports 143B and 144B, respectively, to the combiner 136. The combiner 136 takes a difference between the instantaneous signal level of the constructive path signal and the instantaneous signal level of the destructive path signal for providing the baseband signal. In a variation of the receiver 120, the data estimator 132 connects to the ports 143B and 144B for receiving differential electrical signals.
The optical signal processor 134A includes a delay line interferometer (DLI) 150 and an optical imbalancer 152. The electrical signal processor 134B includes an electrical imbalancer 156. The DLI 150 has an input port 165 connected to the input port 142 of the demodulator 130 for receiving the signal 24. The constructive transfer function of the DLI 150 between the input port 165 and its constructive output port 166 includes the transfer function G(f) of the equation 1. The destructive transfer function of the DLI 150 between the input port 165 and its destructive output port 168 includes the transfer function H(f) of the equation 2.
The constructive transfer function of the signal processor 134 between the input port 142 and the constructive output port 143B includes the constructive transfer function of the DLI 150 and the transfer functions in the constructive signal path of the optical imbalancer 152, the detector apparatus 135 and the electrical imbalancer 156. Similarly, the destructive transfer function of the signal processor 134 between the input port 142 and the destructive output port 144B includes the destructive transfer function of the DLI 150 and the transfer functions in the destructive signal path of the optical imbalancer 152, the detector apparatus 135 and the electrical imbalancer 156.
The signals at the constructive and destructive output ports 166 and 168 may be created with optical interference by separating the input signal at the port 165 into two paths and then recombining the signals. The DLI 150 has a first signal delay element referred to as a transfer free spectral range (FSR) bandwidth element 148 and a second signal delay element referred to as a transfer (FSR) phase element 146. The FSR phase element 146 provides a delay difference between the signal transit times in the signal paths in the DLI 150 and also provides a transfer function phase shift Φ to the constructive and destructive free spectral range transfer functions for the DLI 150. The FSR bandwidth element 148 provides a signal delay Z (
The signal delay Z provided by the FSR bandwidth element 148 is called an FSR delay to distinguish it from the signal delay difference provided by the FSR phase element 146 called an FSR phase delay. The reader should be aware that two different types of phases are being described here—the phases of the periodic signals and the phases of the periodic transfer functions G(f) and H(f). The FSR delay Z is a major contributor to the signal transit time difference Y for differentially demodulating the input signal 24. It should be noted that for the receiver 120, the time difference Y will not, in general, be the same as the symbol time T of the modulated signal 24. In a typical system 10, the time difference Y of the receiver 120 is less than about 83% of the symbol time T.
The inverse of the time difference Y defines the free spectral range (FSR) and the bandwidth of the constructive and destructive transfer functions of the DLI 150. The free spectral range of the DLI 150 determines or is a contributor to the constructive and destructive transfer functions G(f) and H(f) for the DLI 150. The FSR delay Z of the FSR bandwidth element 148 is selected or adjusted based on known or measured characteristics of the link 16 to provide the time difference Y that provides a desired free spectral range (FSR) for the DLI 150 for improving the performance of the system 10, and especially for reducing the signal quality degradation due to intersymbol interference (ISI) caused by the filters 26. The bandwidth (FSR) control algorithm 133 calculates or provides feedback for determining or controlling the element 148 as shown in the chart of
The FSR phase element 146 is used for fine tuning the phase Φ of the cyclic frequency response of the transfer functions G(f) and H(f) to tune the transfer functions G(f) and H(f) relative to the carrier frequency of the modulated input signal 24. In general, the FSR phase must be re-adjusted each time a new FSR delay Z is selected or adjusted. The FSR phase element 146 may be controlled by a mechanism 174 included in the DLI 150 where the mechanism 174 is controlled by the FSR phase controller 137. The mechanism 174 may be an oven for controlling the temperature of the element 146.
The receiver 120 may include an input optical filter for filtering the optical signal 24 into a channel when the optical signal 24 has multiple channels that are wavelength division multiplexed (WDM). The input optical filter may be viewed as one of the filters 26 in the link 16. It is desirable for cost and convenience that the same processor 134, and the same DLI 150 be used for any channel.
In an exception to the general rule stated above, the FSR phase controller 137 and FSR phase element 146 may not be necessary when the FSR bandwidth element 148 is selected for providing the time difference Y exactly equal to the inverse of the frequency spacing of the channels. For example, for a channel spacing of 50 GHz and a symbol time of 23 picoseconds, the time difference Y might be 20 picoseconds. However, in this special case, the FSR of the DLI 150 may not be optimized for best BER. In the receiver 120, the FSR bandwidth element 148 is selected according to criteria of compensating for ISI in the transmission link 16 for providing the transit time difference Y and the FSR for best BER where the time difference Y is not the inverse of the channel spacing.
The optical imbalancer 152 includes constructive and destructive variable gain elements 176 and 178 for controlling the optical gains that are applied to the signals from the output ports 166 and 168, respectively, in order to provide a gain imbalance between the constructive and destructive signals to the output ports 143A and 144A. The gains of the elements 176 and 178 may be controlled by the imbalance control algorithm 164 for varying the ratio of the power gains for constructive and destructive paths for providing constructive and destructive transfer functions go(f) and ho(f) according to respective equations 3 and 4. In the equations 3 and 4, the optical gain imbalance, shown with symbol βo, varies from minus one to plus one.
g
o(f)=1−βo (3
h
o(f)=1+βo (4
The imbalance operation may be provided dynamically in a closed loop using active feedback for minimizing the BER from the BER detector 138. Or, the imbalance operation may be “set and forget” (until it is set and forgotten again) after measuring the BER. Or, the imbalance operation may be open loop based on calculations from known or measured characteristics of the link 16. The calculations are shown in a
The detector apparatus 135 includes a constructive photo-detector 182 and a destructive photo-detector 184 for detecting the optical signals for the ports 143A and 144A, respectively, by converting photons to electrons for providing electrical signals to the electrical imbalancer 156. Photodiodes may be used for the detectors 182 and 184. Each photodiode 182 and 184 produces an electrical signal proportional to detected optical power. The constructive and destructive transfer functions from the input port 165 to the electrical outputs of the detector apparatus 135 include the terms of respective equations 5 and 6.
G(f)*go(f)={[1+cos(2πfY)]/2}*(1−βo) (5
H(f)*ho(f)={[1−cos(2πfY)]/2)}*(1+βo) (6
The FSR phase controller 137 controls the phase delay of the FSR phase element 146 for maximizing a ratio of the optical powers in the constructive and destructive detectors 182 and 184. In a preferred embodiment, FSR phase controller 137 includes a detector 145 for making a power-related measurement for the signals in the constructive and destructive signal paths. The detector 145 measures and then averages the optical powers in the constructive and destructive detectors 182 and 184 by measuring photocurrents AC and AD, respectively. The photocurrents are the electrical currents in the detectors 182 and 184 that result from the conversions of photons to electrons. The photocurrents are measured by measuring the electrical currents passing through the detectors 182 and 184 and then averaging to remove high frequency components. The high frequency components can be removed with low pass electrical filters with passbands lower than the bandwidth of the optical modulation.
An algorithm in the FSR phase controller 137 controls the phase delay of the FSR phase element 146 in order to maximize a ratio, difference or normalized difference of the transfer functions. The normalized difference is the difference between the constructive and destructive signal path power-related measurements divided by the sum of the constructive and destructive signal path power-related measurement. The FSR phase controller 137 may be constructed in order to maximize the normalized difference ΔB measured from the average photocurrents as shown in an equation 7.
ΔB=(AC−AD)/(AC+AD) (7
The receiver 120 may also include a path for signal quality feedback 192. Data for signal quality is processed through the signal quality feedback 192 and passed to the FSR phase controller 137. The FSR phase controller 137 uses the processed signal quality data for fine tuning the phase delay of the FSR phase element 146 in order to improve and optimize the signal quality. Preferably, the FSR phase element 146 is first tuned in a feedback loop for maximizing a constructive—destructive normalized power difference and then fine tuned for minimizing a bit error ratio (BER). The signal quality data may be obtained by measuring BER directly, by measuring an eye opening ratio of a baseband signal and/or by measuring a signal to noise ratio (SNR) of the optical or electrical constructive and destructive path signals.
The electrical imbalancer 156 includes constructive and destructive variable gain elements 186 and 188 for controlling the electrical gains applied to the signals from the constructive and destructive detectors 182 and 184, respectively, and issuing signals from output ports 143B and 144B. The gains of the elements 186 and 188 may be controlled by the imbalance control algorithm 164 for varying the ratio of the gains for constructive and destructive paths for providing constructive and destructive transfer functions ge(f) and he(f) according to respective equations 8 and 9. In the equations 8 and 9, the electrical gain imbalance, shown with symbol βe, varies from minus one to plus one.
ge(f)=1−βe (8
he(f)=1+βe (9
The imbalance operation may be provided dynamically in a closed loop using active feedback for minimizing the BER from the BER detector 138. Or, the imbalance operation may be “set and forget” (until it is set and forgotten again) after measuring the BER. Or, the imbalance operation may be open loop provided based on calculations from known or measured characteristics of the link 16. The calculations are shown in a
The combiner 136 takes the difference between the electrical signals from the constructive and destructive output ports 143B and 144B and passes the difference as a baseband signal to the data estimator 132. The baseband signal is the demodulated signal corresponding to the input signal 24.
The baseband signal has instantaneous signal levels that in a system with no degradation would be exactly representative of the input data at sample times synchronized to a data clock. For example at the sample times, one signal level would represent a logical “1” and another signal level would represent a logical “0” for the input data. However, various signal degradations, especially intersymbol interference (ISI) due to the filters 26 in the link 16, cause the signal levels of the baseband signal at the sample times to have many levels and occasionally even have levels where a “1” appears to be a “0” and vice versa. The baseband signal synchronized to the data clock and shown over and over again on the same display appears as an eye diagram where the opening of the eye is a measure of the quality of the demodulated signal.
The data estimator 132 recovers frame and data clock signals and uses error detection and correction techniques for making its best estimate of the input data. Its best estimate of the input data is issued as output data. The BER detector 138 uses error detection and correction information from the date estimator 132 and/or programmed knowledge of expected data bits in the output data for estimating a bit error ratio (BER). For dynamic operation, the BER detector 138 passes the BER to the imbalance control algorithm 164 in the demodulator 130. The function of the BER detector 138 for providing BER measurements or feedback may be replaced or augmented with a device for measuring the signal quality of the baseband signal. The signal quality device and/or measurement may be internal to the receiver 120 or external. Test equipment may be used as an external device for measuring signal quality or BER.
A side effect of changing the selection of the FSR delay Z is that the transfer function phase or FSR phase of the transfer functions G(f) and H(f) may slide many cycles with respect to the frequency of the input signal 24. In a general rule, whenever the FSR delay is changed, the transfer function phase shift Φ, or phase shifts ΦG and ΦH, must be re-adjusted by the transfer (FSR) phase controller 137 by adjusting the FSR phase element 146 for re-tuning the transfer functions G(f) and H(f) to the frequency of the input optical signal 24. On the other hand the effect of changing the phase shift Φ, or phase shifts ΦG and ΦH, on the FSR bandwidth is so small that it is insignificant
The receiver 20,120 includes a microprocessor system for operating the receiver 20,120 according to instructions stored in a memory. These instructions include the above described bandwidth (FSR) control algorithm 33,133, the imbalance control algorithm 64,164 and the signal quality feedback 92,192. Signal quality for the receiver 20,120 may be defined in terms of BER, ISI, eye opening ratio, and/or signal to noise ratio (SNR). Typically the minimum BER, the best compensation for ISI, the largest eye openings and the highest signal to noise ratios (SNR)s of the optical and electrical constructive and destructive path signals are optimized, or nearly optimized, for the same selections and adjustments within the receiver 20,120. The algorithm 192 may operate in a feedback loop for minimizing BER.
The structural elements of DLI 150A are disposed as follows. The input optical signal 24 illuminates the front side of the partially reflecting first mirror 202A. The first mirror 202A is set at an angle to the path of the optical signal 24 so that part of the signal 24 is reflected as a signal 212A and part of the signal 24 is passed through as a signal 214A. The signal 212A is reflected from the second mirror 204A as a signal 216A back to the front side of the first mirror 202A. The signal 214A illuminates the element 146A and emerges after a fine tune phase delay as a signal 218A. The signal 218A reflects from the third mirror 208A as a signal 222A.
The signal 222A illuminates the element 146A and emerges after the phase delay as a signal 224A. The signal 224A illuminates the back side of the first mirror 202A. Part of the signal 224A is reflected from the back side of the first mirror 202A to combine with part of the signal 216A passed through the front side of the first mirror 202A for providing a signal 226A at the constructive output port 166A. Part of the signal 224A passes through the back side of the first mirror 202A to combine with part of the signal 216A reflected from the front side of the first mirror 202A for providing a signal 228A at the destructive output port 168A.
The elements of the DLI 150A split the input signal 24 into a first path 232A and a second path 234A. The transit time of the first path 232A is the sum of the transit times of the signals 212A and 216A. The transit time of the second path 234A is the sum of the transit times of the signals 214A, 218A, 222A and 224A plus two times the phase delay of the element 146A. The difference between the first and second path transit times is the differential transit time Y that is used for demodulation of the input optical signal 24. The time Y is fine tuned by adjusting the signal phase delay in the element 146A in order to adjust the FSR phase of the DLI 150A for adjusting the transfer function phase of the constructive and destructive transfer functions G(f) and H(f) (see
The material for the element 146A is selected to have an optical index that depends upon temperature. The FSR phase controller 137A provides a control signal to adjust the temperature of the oven 174A in order to fine tune the delay of the element 146A for centering the constructive and destructive transfer functions G(f) and H(f) of the DLI 150A on the optical carrier frequency of the input optical signal 24.
The structural elements of DLI 150B are disposed as follows. The input optical signal 24 illuminates the front side of the partially reflecting first mirror 202B. The first mirror 202B is set at an angle to the path of the optical signal 24 so that part of the signal 24 is reflected as a signal 212B and part of the signal 24 is passed through as a signal 214B. The signal 212B is reflected from the second mirror 204B as a signal 216B back to the front side of the first mirror 202B. The signal 214B illuminates the element 148B and emerges after the delay Z as a signal 217B. The signal 217B illuminates the element 146B and emerges after a fine tune phase delay as a signal 218B. The signal 218B reflects from the third mirror 208B as a signal 222B.
The signal 222B illuminates the element 146B and emerges after the phase delay as a signal 223B. The signal 223B illuminates the element 148B and emerges after the delay Z as a signal 224B. The signal 224B illuminates the back side of the first mirror 202B. Part of the signal 224B is reflected from the back side of the first mirror 202B to combine with part of the signal 216B passed through the front side of the first mirror 202B for providing a signal 226B at the constructive output port 166B. Part of the signal 224B passes through the back side of the first mirror 202B to combine with part of the signal 216B reflected from the front side of the first mirror 202B for providing a signal 228B at the destructive output port 168B.
The elements of the DLI 150B split the input signal 24 into a first path 232B and a second path 234B. The transit time of the first path 232B is the sum of the transit times of the signals 212B and 216B. The transit time of the second path 234B is the sum of the transit times of the signals 214B, 217B, 218B, 222B, 223B and 224B plus two times the phase delay of the element 146B plus two times the delay Z. The difference between the first and second path transit times is the differential transit time Y that is used for demodulation of the input optical signal 24. The FSR delay Z is a part of the transit time difference Y. A bandwidth (FSR) control algorithm 133B (
The material for the element 146B is selected to have an optical index that depends upon temperature. The FSR phase controller 137B provides a control signal to adjust the temperature of the oven 174B in order to fine tune the delay of the element 146B for centering the constructive and destructive transfer functions G(f) and H(f) (
The structural elements of DLI 150C are disposed as follows. The input optical signal 24 illuminates the front side of the partially reflecting first mirror 202C. The first mirror 202C is set at an angle to the path of the optical signal 24 so that part of the signal 24 is reflected as a signal 212C and part of the signal 24 is passed through as a signal 214C. The signal 212C is reflected from the second mirror 204C as a signal 216C back to the front side of the first mirror 202C. The signal 214C illuminates the element 148C,146C and emerges after the delay Z and an adjustment by the fine tune phase delay as a signal 218C. The signal 218C reflects from the third mirror 208C as a signal 222C.
The signal 222C illuminates the element 148C,146C and emerges after the delay Z and an adjustment by the phase delay as a signal 224C. The signal 224C illuminates the back side of the first mirror 202C. Part of the signal 224C is reflected from the back side of the first mirror 202C to combine with part of the signal 216C passed through the front side of the first mirror 202C for providing a signal 226C at the constructive output port 166C. Part of the signal 224C passes through the back side of the first mirror 202C to combine with part of the signal 216C reflected from the front side of the first mirror 202C for providing a signal 228C at the destructive output port 168C.
The elements of the DLI 150C split the input signal 24 into a first path 232C and a second path 234C. The transit time of the first path 232C is the sum of the transit times of the signals 212C and 216C. The transit time of the second path 234C is the sum of the transit times of the signals 214C, 218C, 222C and 224C plus two times the delay Z with the adjustment of the phase delay of the element 148C,146C. The difference between the first and second path transit times is the differential transit time Y that is used for demodulation of the input optical signal 24. The FSR delay Z is a part of the transit time difference Y. A bandwidth (FSR) control algorithm 133C (
The material for the element 148C,146C is selected to have an optical index that depends upon temperature. The FSR phase controller 137C provides a control signal to adjust the temperature of the oven 174C in order to fine tune the phase delay of the element 146C for centering the constructive and destructive transfer functions G(f) and H(f) (
The DLI 150D includes an input port 165D, a transfer (FSR) phase element 146D, the positionable delay element 148D, a mechanism or oven 174D, a positioning device 175D, a partially reflecting first mirror 202D, a second mirror 204D, a third mirror 208D, and constructive and destructive output ports 166D and 168D disposed as follows. The input optical signal 24 illuminates the front side of the partially reflecting first mirror 202D. The first mirror 202D is set at an angle to the path of the optical signal 24 so that part of the signal 24 is reflected as a signal 212D and part of the signal 24 is passed through as a signal 214D. The signal 212D is reflected from the second mirror 204D as a signal 216D back to the front side of the first mirror 202D. The signal 214D illuminates the element 148D and emerges after the FSR delay Z as a signal 217D. The signal 217D illuminates the element 146D and emerges after an adjustable fine tuning delay as a signal 218D. The signal 218D reflects from the third mirror 208D as a signal 222D.
The signal 222D illuminates the element 146D and emerges after the phase delay as a signal 223D. The signal 223D illuminates the element 148D and emerges after the FSR delay Z as a signal 224D. The signal 224D illuminates the back side of the first mirror 202D. Part of the signal 224D is reflected from the back side of the first mirror 202D to combine with part of the signal 216D passed through the front side of the first mirror 202D for providing a signal 226D at the constructive output port 166D. Part of the signal 224D passes through the back side of the first mirror 202D to combine with part of the signal 216D reflected from the front side of the first mirror 202D for providing a signal 228D at the destructive output port 168D. Typically, the elements 148D and 146D have bulk optical group indices (time delay equals group index multiplied by distance divided by the speed of light in a vacuum) that are much greater than the group indices experienced by the signals 212D, 216D, 214D, 217D, 218D, 222D, 223D and 224D outside of the elements 148D and 146D.
The elements of the DLI 150D split the input signal 24 into a first path 232D and a second path 234D. The transit time of the first path 232D is the sum of the transit times of the signals 212D and 216D. The transit time of the second path 234D is the sum of the transit times of the signals 214D, 217D, 218D, 222D, 223D and 224D plus two times the adjustable delay of the element 146D plus two times the FSR delay Z of the element 148D. The difference between the first and second path transit times is the differential transit time Y that is used for demodulation of the input optical signal 24. Either or both of the elements 146D and 148D may be constructed in two pieces, one in the signal path 232D and one in the signal path 234D, for providing a signal delay that is the difference between the signal delays of the two element pieces.
The element 148D has a stair step cross section having two or more stair risers 242D and stair treads 244D. Alternatively, the element 148D may have segments having different group indices. The positioning device 175D positions the element 148D so that the signal 214D enters the element 148D at one of the risers 242D and the signal 224D emerges from the element 148D at one of the risers 242D. Alternatively, the stairway is on the opposite side of the element 148D so the signal 223D enters and the signal 217D emerges from the element 148D at one of the risers 242D. The delay step sizes are proportional to the physical lengths of the treads 244D.
The positioning device 175D stepwise positions the element 148D in a direction approximately perpendicular to the signals 214D, 217D, 223D and 224D in order to increase or decrease the effective optical length of the element 148D in order to increase or decrease the FSR delay Z. The discrete steps of the delay Z provide discrete steps of the differential transit time Y, thereby providing discrete steps in the FSR bandwidths of the constructive and destructive transfer functions G(f) and H(f) for the DLI 150D. Discrete steps may be beneficial to provide immunity to jitter in the position of the element 148D.
A bandwidth (FSR) control algorithm 133D controls the positioning device 175D for positioning the element 148D. The control algorithm 133D may be external to the receiver 20,120 or included within the receiver 20,120. A technician uses information from the control algorithm 133D to operate the positioning device 175D or the information from the control algorithm 133D is part of a feedback loop for automatic operation of the positioning device 175D.
The material for the element 146D is selected to have a group index that depends upon temperature. The FSR phase controller 137D provides a control signal to adjust the temperature of the oven 174D in order to fine tune the delay of the element 146D for centering the constructive and destructive transfer functions G(f) and H(f) of the DLI 150D on the optical carrier frequency of the input optical signal 24. The functions of the transfer phase element 146D and the stepped FSR delay element 148D may be combined (as shown in the
The DLI 150E includes an input port 165E, a transfer (FSR) phase element 146E, the positionable delay element 148E, a mechanism or oven 174E, a positioning device 175E, a partially reflecting first mirror 202E, a second mirror 204E, a third mirror 208E, and constructive and destructive output ports 166E and 168E disposed as follows. The input optical signal 24 illuminates the front side of the partially reflecting first mirror 202E. The first mirror 202E is set at an angle to the path of the optical signal 24 so that part of the signal 24 is reflected as a signal 212E and part of the signal 24 is passed through as a signal 214E. The signal 212E is reflected from the second mirror 204E as a signal 216E back to the front side of the first mirror 202E. The signal 214E illuminates the element 148E and emerges after the FSR delay Z as a signal 217E. The signal 217E illuminates the element 146E and emerges after an adjustable fine tuning delay as a signal 218E. The signal 218E reflects from the third mirror 208E as a signal 222E.
The signal 222E illuminates the element 146E and emerges after the phase delay as a signal 223E. The signal 223E illuminates the element 148E and emerges after the FSR delay Z as a signal 224E. The signal 224E illuminates the back side of the first mirror 202E. Part of the signal 224E is reflected from the back side of the first mirror 202E to combine with part of the signal 216E passed through the front side of the first mirror 202E for providing a signal 226E at the constructive output port 166E. Part of the signal 224E passes through the back side of the first mirror 202E to combine with part of the signal 216E reflected from the front side of the first mirror 202E for providing a signal 228E at the destructive output port 168E. Typically, the elements 148E and 146E have bulk group indices that are much greater than the group indices experienced by the signals 212E, 216E, 214E, 217E, 218E, 222E, 223E and 224E outside the delay elements 148E and 146E.
The elements of the DLI 150E split the input signal 24 into a first path 232E and a second path 234E. The transit time of the first path 232E is the sum of the transit times of the signals 212E and 216E. The transit time of the second path 234E is the sum of the transit times of the signals 214E, 217E, 218E, 222E, 223E and 224E plus two times the adjustable delay of the element 146E plus two times the FSR delay Z of the element 148E. The difference between the first and second path transit times is the differential transit time Y that is used for demodulation of the input optical signal 24. Either or both of the elements 146E and 148E may be constructed in two pieces, one in the signal path 232E and one in the signal path 234E, for providing a signal delay that is the difference between the signal delays of the two element pieces.
The element 148E has a cross section having a smooth change or gradient of physical length in order to provide a continuously variable optical delay. Alternatively, the element 148E may have a smooth gradient of optical group index. The positioning device 175E moves the element 148E in a direction perpendicular to the signals 214E, 217E, 223E and 224E in order to increase or decrease the effective optical length of the element 148E in order to increase or decrease the FSR delay Z. The continuously variable FSR delay Z provides a continuously variable differential transit time Y, thereby providing a smooth, continuously variable FSR bandwidth for the constructive and destructive transfer functions G(f) and H(f) for the DLI 150E.
A bandwidth (FSR) control algorithm 133E controls the positioning device 175E for positioning the element 148E. The control algorithm 133E may be external to the receiver 20,120 or included within the receiver 20,120. A technician uses information from the control algorithm 133E to operate the positioning device 175E or the information from the control algorithm 133E operates the positioning device 175E automatically to move the element 148E more or less perpendicular to the optical signals 214E, 217E, 223E and 224E.
The material for the element 146E is selected to have an optical group index that depends upon temperature. The FSR phase controller 137E provides a control signal to adjust the temperature of the oven 174E in order to fine tune the delay of the element 146E for centering the constructive and destructive transfer functions G(f) and H(f) of the DLI 150E on the optical carrier frequency of the input optical signal 24.
The elements 148E and 146E may be combined (as shown in
The DLI 150F includes an input port 165F, a transfer (FSR) phase element 146F, a mechanism or oven 174F, a positioning device 175F, a partially reflecting first mirror 202F, a second mirror 204F, the movable third mirror 208F, and constructive and destructive output ports 166F and 168F disposed as follows. The input optical signal 24 illuminates the front side of the partially reflecting first mirror 202F. The first mirror 202F is set at an angle to the path of the optical signal 24 so that part of the signal 24 is reflected as a signal 212F and part of the signal 24 is passed through as a signal 214F. The signal 212F is reflected from the second mirror 204F as a signal 216F back to the front side of the first mirror 202F. The signal 214F illuminates the element 146F and emerges after a fine tune signal delay as a signal 218F. The signal 218F passes through the delay Z to reflect from the third mirror 208F as a signal 222F.
The signal 222F passes through the delay Z to illuminate the element 146F and emerges after the phase delay as a signal 224F. Part of the signal 224F is reflected from the back side of the first mirror 202F to combine with part of the signal 216F passed through the front side of the first mirror 202F for providing a signal 226F at the constructive output port 166F. Part of the signal 224F passes through the back side of the first mirror 202F to combine with part of the signal 216F reflected from the front side of the first mirror 202F for providing a signal 228F at the destructive output port 168F. Typically, the element 146F has a group index much greater than the group indices experienced by the signals 212F, 216F, 214F, 218F, 222F and 224F outside the delay element 146F.
The elements of the DLI 150F split the input signal 24 into a first path 232F and a second path 234F. The transit time of the first path 232F is the sum of the transit times of the signals 212F and 216F. The transit time of the second path 234F is the sum of the transit times of the signals 214F, 218F, 222F and 224F plus two times the phase delay of the element 146F plus two times the FSR delay Z of the mechanical length adjustment of the movable mirror 208F. The difference between the first and second path transit times is the differential transit time Y that is used for demodulation of the input optical signal 24. The element 146F may have one piece in the signal path 232F and one piece in the signal path 234F for fine tuning a signal delay that is the difference between the signal delays in the two paths 232F and 234F. Either or both of the mirror 204F and 208F may be constructed as moving mirrors having a selectable position 246F.
The positioning device 175F moves the mirror 208F in the direction of the signals 218F and 222F in order to increase or decrease the effective optical length between the signal paths 232F and 234F of the DLI 150F in order to increase or decrease the FSR delay Z. The continuously variable FSR delay Z provides a continuously variable differential transit time Y, thereby providing a smooth, continuously variable FSR bandwidth for the constructive and destructive transfer functions G(f) and H(f) for the DLI 150F.
A bandwidth (FSR) control algorithm 133F controls the positioning device 175F for positioning the mirror 208F. The control algorithm 133F may be external to the receiver 20,120 or included within the receiver 20,120. A technician uses information from the control algorithm 133F to operate the positioning device 175F or the information from the control algorithm 133F operates the positioning device 175F automatically to move the element 148F to shorten or lengthen the distance traveled by the optical signals 218F and 222F. The positioning device 175F may be constructed in a manner similar to the construction described below for the positioning device 175D.
The material for the element 146F is selected to have an optical group index that depends upon temperature. The FSR phase controller 137F provides a control signal to adjust the temperature of the oven 174F in order to fine tune the delay of the element 146F for centering the constructive and destructive transfer functions G(f) and H(f) of the DLI 150F on the optical carrier frequency of the input optical signal 24. The movable mirror 208F may combine the functions for selecting the FSR delay Z and fine tuning the FSR phase.
In a step 301 constructive and destructive transfer functions are calculated, looked up in a table based on calculations, or actively tuned for minimizing the effect of intersymbol interference (ISI) for improving signal quality. The transfer functions may be implemented by selecting a delay Z in a signal path of a delay line interferometer (DLI) in order to select the free spectral range (FSR) of the DLI. The delay Z contributes to a differential time Y, in general not equal to a DPSK symbol time T, for providing differential demodulation. The signal quality may be determined in terms of bit error ratio (BER) for output data. In a first embodiment the delay Z is selected by dynamically adjusting the delay Z with feedback from a signal quality measurement in order to minimize the BER. In a second embodiment the delay Z is selected by trial and error in order to minimize a measured BER. In a third embodiment the delay Z is selected based upon a BER measurement on another optical transmission link channel where the other channel is known to have the same channel bandwidth. In a fourth embodiment the delay Z is selected by calculating from a known channel or spectrum bandwidth. In a fifth embodiment the delay Z is selected from a table having calculations based on channel bandwidth or spectrum for minimizing BER. The calculations for FSR are shown in the chart of
In a step 302 an optical gain imbalance between constructive and destructive output port signals is selected (as described above for the FSR delay Z) for best signal quality. The calculations for gain imbalance are shown in
In a step 303 the phase of the constructive and destructive transfer functions is adjusted for maximizing the signal power difference between optical constructive and destructive path signals. The transfer function phases may be adjusted as FSR phases while the system is in operation for providing output data without overly degrading the output data by fine tuning the delay of a signal delay element in a signal path in the DLI. Optionally, the FSR phase is further tuned for best signal quality. The FSR phase adjustment tunes the constructive and destructive transfer functions relative to the carrier frequency of the input optical signal.
Either during design, test or installation in a step 320 a free spectral range (FSR) of a delay line interferometer (DLI) is calculated based on characteristics, particularly the bandwidth of the link 16, for the transmission system 10 for obtaining the best signal quality and/or lowest bit error ratio (BER). In a step 322 optical and/or electrical gain imbalances are calculated based on the FSR of the DLI, the symbol rate R, and the characteristics of the transmission system 10, particularly the bandwidth of the filters 26, for obtaining the best signal quality and/or lowest bit error ratio (BER).
In operation the receiver 20,120 receives the modulated input signal 24 in a step 324. In a step 330 the DLI having the pre-calculated FSR differentially decodes the signal 24 and uses optical interference for separating the signal into constructive and destructive signal paths. In a step 332 the FSR phase is adjusted for tuning the FSR transfer functions relative to the carrier of the signal 24. In a step 334 the optical gain imbalance is applied to the signals in the constructive and destructive signal paths for providing optical constructive and destructive signal outputs.
The modulations of the signals at the optical constructive and destructive signal outputs are converted to electrical signals in a step 336. In a step 338 the electrical gain imbalance is applied to the signals in the constructive and destructive signal paths for providing electrical constructive and destructive signal outputs.
Power-related measurements are detected in a step 342 for the signals at the constructive and destructive signal outputs. When the gain imbalance is applied to the electrical signals, the electrical output signals are measured. When the gain imbalance is applied to the optical signals but not the electrical signals, either the optical or the electrical output signals may be measured. In one embodiment, the gain is applied to the optical signals and the power-related detections are measurements of the average photocurrents for converting the optical modulation to electrical signals. In a step 344 a normalized difference between the power-related measurements is applied to adjust the FSR phase for the step 332. In a step 352 the electrical constructive and destructive path signals are combined by taking the difference of the signals. The difference is issued as a baseband signal. Finally, in a step 354 the input data from the transmitter 12 is estimated from the baseband signal for providing output data.
The input signal 24 is received at the start in the step 324. In the step 330 the DLI differentially decodes the signal 24 and uses optical interference for separating the signal into constructive and destructive signal paths. In the step 332 the FSR phase is adjusted for tuning the FSR transfer functions relative to the carrier of the signal 24. For a symmetrical signal spectrum, the FSR phase is tuned for centering the FSR transfer functions to the carrier of the signal 24. In the step 334 the optical gain imbalance is applied to the signals in the constructive and destructive signal paths for providing optical constructive and destructive signal outputs.
The modulations of the signals at the optical constructive and destructive signal outputs are converted to electrical signals in the step 336. In the step 338 the electrical gain imbalance is applied to the signals in the constructive and destructive signal paths for providing electrical constructive and destructive signal outputs.
Power-related measurements are detected in the step 342 for the signals at the constructive and destructive signal outputs. When the gain imbalance is applied to the electrical signals, the electrical output signals are measured. When gain imbalance is applied to the optical signals but not the electrical signals, either the optical or the electrical output signals may be measured. In one embodiment, the gain is applied to the optical signals and the power-related detections are measurements of the average photocurrents for converting the optical modulation to electrical signals. In the step 344 a normalized difference between the power-related measurements is applied to adjust the FSR phase for the step 332. In a step 352 the electrical constructive and destructive path signals are combined by taking the difference of the signals. The difference is issued as a baseband signal.
The difference of the constructive and destructive electrical signal outputs is determined in the step 352 for providing a baseband signal. In the step 354 the input data from the transmitter 12 is estimated from the baseband signal for providing output data.
A signal quality determined from the optical or electrical signals, or a bit error ratio (BER), is measured for the output data in a step 372. In a step 374, feedback for the signal quality or BER is applied to adjust the FSR used in the step 330. In a step 376 feedback for the signal quality is applied to adjust the optical and/or gain imbalance for the step 334. And optionally, in a step 378 feedback for the signal quality is applied to adjust the FSR phase for the step 332. The steps 330, 332 and/or 334 may be iterated until no further improvement in signal quality is detected. Whenever the FSR is changed due to a new selection or adjustment in the step 330, the FSR phase must be re-tuned in the step 332.
The elements 148D and 148E are positioned in a second direction 404 by the positioning devices 175D and 175E for selecting the delay Z for providing a desired transit time difference Y between the signal paths 232D and 234D in the DLI 150D or the signal paths 232E and 234E in the DLI 150E. The second direction 404 is about perpendicular to the first direction 402. The term “gradient” denotes the change of signal delay of the element 148D,148E with respect to a change in position of the element 148D,148E in the second direction 404. In various embodiments the delay Z can be varied over a range of one, two, five, ten or twenty picoseconds. The delay steps of the element 148D are typically about one-quarter to five picoseconds for a channel bandwidth of 50 GHz but may be as small as 20 femtoseconds (fs) or even less. In terms of a modulation symbol time, the delay steps are typically one to twenty percent of the modulation symbol time for the modulated optical input signal but may be as small as 0.025% or even less. With respect to the channel bandwidth or modulation bandwidth for the modulated optical input signal 24, the delay steps are typically one to twenty percent but may be as small as 0.025% or even less of the inverse of the bandwidth.
The positioning device 175D has a means of nudging or positioning the elements 148D and 148E in the second direction 404. For good position control, the positioning device 175D may have a screw 423. Manual operation by a technician or a stepping motor 424 controls a rotation 433 of the screw 423 to push or pull the element 148D in the second direction 404 based information from the bandwidth FSR control algorithm 133D. Brackets 426 retain the screw 423 and the motor 424 in a DLI housing with respect to the signals 400. The positioning device 175E may be constructed in a similar way. Some fixing means, such as tie down straps, fix the elements 148D and 148E once the elements 148D are 148E are properly positioned. The fixing means and/or brackets 426 may require shock absorption material to isolate the element 148D and 148E from mechanical vibration of the DLI housing.
The elements 409 and 410 induce wavelength dependent beam deviation angles 411 and 412, respectively, due to the refractive indices of the materials and the gradient angles between the sides of the materials and the signals 400. The materials and the gradient angles may be selected so that the wavelength dependence of the beam deviation angle 411 compensates for the wavelength dependence of the beam deviation angle 412 for providing a signal path that is largely independent of wavelength.
First and second sides of the fixed element 409 are denoted as sides 413 and 414 and first and second sides of the positionable element 410 are denoted as sides 415 and 416. For the same material for the elements 409 and 410, the sides 413 and 415 may be about parallel and the sides 414 and 416 may be about parallel. However, the element 410 may be allowed a small rotation with respect to the element 409. A continuous smooth variation of the position of the movable part 410 in the second direction 404 while the fixed part 409 remains stationary in the second direction 404 provides a continuous smooth variation of the delay Z.
A portion of one of the two signal paths 232A-F or 234A-F (
The adjustable tilt angle 448 is adjusted with respect to the directions of the signals 450 by a mechanical mechanism 474. The mechanism 474 is controlled by a transfer (FSR) phase controller 437 in the manner described above for the transfer FSR controllers 37, 137 and 137A-F. Adjusting the tilt angle 448 of the element 446 with respect to the signals 450 provides a fine adjustment to the delay of the signals 450 by changing the physical length traversed by the signals 450. The element 446 may be constructed with a material having an optical index having minimal temperature dependence.
The signal delay provided by the transfer (FSR) phase elements 46, 146, 146A-F and 446 must be adjustable over a range of at least one cycle period at the carrier frequency of the optical input signal 24 for providing the transfer function phase adjustment. Its tuning resolution and stability should be better than 1% of the carrier cycle period. If the FSR phase adjustment is tuned by temperature, the thermal expansion coefficient and the thermal group index coefficient will determine the scale factor between temperature change and FSR phase change. For example, a tuning plate made of LaSFN9 (by Schott A G of Mainz, Germany), the group index is approximately 1.8 and the sum of the thermal coefficients is approximately 9×10−6/K (Kelvin). The propagation delay through a plate of thickness 3 mm is approximately 18 picoseconds, and the thermal tuning range is 0.162 fs/K. At a carrier frequency of 200 terahertz (THz) the optical period is 5 femtoseconds (fs), so a change in FSR phase of one period would require a temperature change of 31 K, held to a stability of 0.31 K. This is a practical result.
In contrast, the desired differential transit time Y (controlled by selecting the signal delay Z provided by the FSR bandwidth elements 48, 148 and 148A-F is many cycles of the carrier frequency. In terms of cycles of the carrier frequency, the desired time Y may be calculated from the modulation system time divided by the carrier cycle time divided by the desired FSR/R, the FSR/R that correctly compensates for intersymbol interference (ISI) in the modulated optical input signal 24. For example for a modulation symbol time of 23.3 picoseconds, a carrier cycle time of 5 fs and a desired FSR/R of 1.01, the time Y is equivalent to 4613.86 cycles. For the same modulation symbol and carrier cycle times and a desired FSR/R of two, the time Y is equivalent to 2330 cycles. It is not be practical to combine the transfer (FSR) phase element and the FSR bandwidth element for the following reasons.
Taking the above example, for the thermally tuned transfer (FSR) phase element to provide the differential time Y, the phase element would have a delay range of about 2300 carrier cycle periods or 11.5 picoseconds (ps) in order to provide the FSR/R range from 1.01 to 2. This would require an impractical temperature range of 71000 K. The delay Z of the FSR bandwidth element 48, 148 and 148A-F need not provide all of the differential transit time Y. For example the time Y can be composed of the sum of the delay Z and a fixed differential transit time between the signal paths 232A-F and 234A-F, respectively. However, even if the range of the delay Z is limited to one picosecond, the temperature the required temperature tuning range is an impractical 7100 K.
It should be noted that the delay Z described throughout this application is the time for two transits (roundtrip time) through of the transfer function (FSR) bandwidth elements 48, 148 and 148A-F and the transfer (FSR) phase signal delay described throughout this application through the transfer (FSR) phase elements 46, 146, 146A-F and 446 is the time for two transits (roundtrip time). However, the receivers 20 and 120, and DLIs 150 and 150A-F could be constructed for single signal transits through either or both of the bandwidth and phase elements whereby the delay Z and/or the transfer (FSR) phase signal delay are provided by the times for single transits.
Although the present invention has been described in terms of the presently preferred embodiments, it is to be understood that such disclosure is not to be interpreted as limiting. Various alterations and modifications will no doubt become apparent to those skilled in the art after having read the above disclosure. Accordingly, it is intended that the appended claims be interpreted as covering all alterations and modifications as fall within the true spirit and scope of the present invention.
This application is a continuation-in-part of pending application Ser. No. 11/724,017 filed Mar. 14, 2007 by the same inventors for the same assignee.
| Number | Date | Country | |
|---|---|---|---|
| Parent | 11724017 | Mar 2007 | US |
| Child | 11799435 | US |
