A high-power DFB array light source equalization method

Abstract

The application provides a high-power DFB array light source equalization method, relates to the technical field of semiconductor lasers, generates coherent reference light corresponding to a plurality of target light frequencies respectively, and injects each coherent reference light into a corresponding high-power DFB laser channel; the effective output optical power and the optical field phase of each channel at the input end of a data modulator are collected, common equalization power and migrated photon flux are determined according to the optical power, the target light frequency, the channel transmission efficiency and the frequency conversion path efficiency; a corresponding power slice is extracted from a power surplus channel, the power slice is converted into the target light frequency of a power insufficient channel, a conversion carrier phase is corrected and is coherently combined into the main light path of the power insufficient channel; and the proportion, the pumping power and the phase correction amount are extracted according to the equalized optical power, the conversion efficiency and the phase error correction.

Description

Technical Field

[0001] This invention relates to the field of semiconductor laser technology, specifically to a high-power DFB array light source equalization method. Background Technology

[0002] Distributed feedback semiconductor lasers (DFBs) achieve distributed feedback of longitudinal modes satisfying the Bragg condition by setting periodic refractive index modulation gratings along an active waveguide. They feature single longitudinal mode output, compact structure, and ease of array packaging. Multiple DFB lasers arranged according to different target light frequencies form a multi-wavelength DFB array, which can be connected to wavelength division multiplexing modulators, silicon photonics switching chips, coherent measurement links, parallel lidar transmission links, and multi-wavelength spectral detection links.

[0003] Due to manufacturing variations in epitaxial material composition, grating period, grating coupling coefficient, electrode contact resistance, end-face reflection, chip thermal resistance, fiber coupling position, and output waveguide insertion loss, different channels in the same DFB array have different output optical powers under corresponding rated driving conditions. After long-term operation, each channel will also experience different threshold current increments, slope efficiency decreases, and coupling loss increments, causing channel power differences to vary with operating time.

[0004] The end-use of high-power DFB arrays is typically not at the DFB chip output, but rather at the data modulator input, wavelength division multiplexer output, or fiber optic interface. Each channel has different propagation losses from the DFB chip output to the final usage location, thus making the optical power at each DFB chip output consistent cannot guarantee that each channel will have consistent effective carrier power at the actual usage location.

[0005] Existing array equalization methods typically increase the DFB drive current in underpowered channels, decrease the DFB drive current in overpowered channels, or place semiconductor optical amplifiers in underpowered channels and variable optical attenuators in overpowered channels. Increasing the DFB drive current alters the active region carrier concentration and junction temperature, causing changes in emitted light frequency, output linewidth, chirp, and relative intensity noise. Decreasing the DFB drive current in overpowered channels reduces the optical power already generated by the array. Variable optical attenuators convert excess optical power into absorbed heat. Semiconductor optical amplifiers increase additional power consumption, thermal load, and amplify spontaneous emission components.

[0006] When the drive current of one channel changes, the resulting additional heat is transferred to adjacent channels via the chip substrate, bonding layer, and package heat sink, causing changes in the emission frequency and coupler insertion loss of adjacent channels. Power control and temperature control need to be repeatedly adjusted according to the thermal time constant. As the number of array channels increases, multiple control actions become interdependent.

[0007] Existing solutions lack a mechanism for reusing existing optical energy in channels with excess power. When a channel ages or stops outputting, a backup DFB laser at the same frequency is typically used, increasing the number of chips, package area, laser drive channels, and heat dissipation load. Therefore, a method is needed to perform optical domain migration on the continuous wave carrier before independent data modulation, so that the surplus carrier in the channel with excess power is converted into an effective carrier at the target optical frequency of the channel with insufficient power, and to avoid using the DFB drive current as the array equalization execution quantity. Summary of the Invention

[0008] To address the shortcomings of existing technologies, this invention provides a high-power DFB array light source equalization method, which solves the problems that the surplus carrier of the power-excess channel is difficult to use to compensate the power-deficient channel and the inconsistency of the effective carrier power at the actual usage location.

[0009] To achieve the above objectives, the present invention provides the following technical solution: a high-power DFB array light source equalization method, comprising the following steps: generating coherent reference beams corresponding to multiple target optical frequencies; injecting each coherent reference beam into a corresponding high-power DFB laser channel, such that each high-power DFB laser channel is in an injection-locked state corresponding to the target optical frequency; inputting the unmodulated continuous wave carrier output from each high-power DFB laser channel into the corresponding channel main optical path; acquiring the effective output optical power of each channel main optical path at the corresponding data modulator input and the optical field phase of each unmodulated continuous wave carrier relative to the corresponding coherent reference beam; determining the common equalization power, the power excess channel, the power deficiency channel, and the source-side photon flux migrating from the power excess channel to the power deficiency channel based on the effective output optical power, the target optical frequency, the output transmission efficiency of each channel main optical path, and the photon transmission efficiency of the frequency conversion path; extracting the power corresponding to the source-side photon flux from the channel main optical path of the power excess channel. The power slice is sliced, and the unextracted continuous wave carrier is retained in the main optical path of the power excess channel. Based on the frequency difference between the target light frequency of the power excess channel and the target light frequency of the power deficiency channel, the power slice is converted into a converted carrier with the same frequency as the target light frequency of the power deficiency channel. Based on the phase error between the optical field phase of the converted carrier and the optical field phase of the original continuous wave carrier of the power deficiency channel, the optical field phase of the converted carrier is adjusted, and the adjusted converted carrier is incorporated into the main optical path of the power deficiency channel. Based on the effective output optical power, frequency conversion efficiency, and phase error of each main optical path after incorporating the converted carrier, the extraction ratio, frequency conversion pump power, and phase correction amount of the power slice are corrected until the effective output optical power of each main optical path is within the allowable power range corresponding to the common equalization power. The driving current of each high-power DFB laser channel is kept within the corresponding injection lock current range during the extraction, conversion, and incorporation of the power slice.

[0010] Preferably, the coherent reference light is generated by a narrow-linewidth reference laser and an optical frequency comb generation structure. The target light frequency of the coherent reference light satisfies In the formula, Indicates the first The target light frequency of a high-power DFB laser channel. This indicates the reference light frequency corresponding to the narrow linewidth reference laser. Indicates the first The integer frequency slot number corresponding to each high-power DFB laser channel. This indicates the basic frequency interval formed by the optical frequency comb generation structure. The high-power DFB laser channel number is indicated; each coherent reference beam is injected into the corresponding high-power DFB laser channel after wavelength demultiplexing and seed beam power adjustment, and the injection lock state is determined based on the beat frequency residual, phase error variance and output spectrum between the corresponding high-power DFB laser channel and the coherent reference beam.

[0011] Preferably, acquiring the effective output optical power and optical field phase of each of the above includes: extracting monitoring light from the input terminal of each of the data modulators, and then... The monitoring light from each channel is input to a 90-degree optical mixer along with the corresponding coherent reference light. The in-phase detection component is obtained through balanced photoelectric detection. and orthogonal detection components According to the in-phase detection component and the orthogonal detection components Construct the first The estimated complex optical field value of the first channel is determined based on the squared amplitude of the estimated complex optical field value, the monitoring splitting ratio, the corresponding coherent reference optical power, the photodetector responsivity, and the analog front-end gain. The effective output optical power of the first channel is determined based on the complex argument of the estimated complex optical field. The optical field phase of each channel.

[0012] Preferably, determining the common equalization power and source-side photon flux includes: defining each target optical frequency as a frequency node, defining the connection between the source frequency node and the target frequency node connected by the frequency conversion unit as a migration edge, and recording the maximum allowable source-side photon flux, power slice extraction efficiency, frequency conversion efficiency, migration bus transmission efficiency, coherent injection efficiency, pump power coefficient, and phase correction state for each migration edge; dividing the effective output optical power of the main optical path of each channel by the product of the corresponding target optical frequency and Planck's constant to obtain the existing photon flux of each frequency node; and determining the common equalization power and the source-side photon flux corresponding to each migration edge under the constraints that the source-side photon flux of each migration edge does not exceed the corresponding maximum allowable source-side photon flux, the photon flux retained by each power excess channel is not lower than the corresponding minimum retained photon flux, the total frequency conversion pump power does not exceed the pump power upper limit, and the effective output optical power of each data modulator input is within the same allowable power range.

[0013] Preferably, extracting the power slice from the main optical path of the power-excess channel includes: inputting the main optical path of the power-excess channel into an adjustable Mach-Zehnder coupler; determining a target extraction ratio based on the ratio of the sum of source-side photon fluxes migrating from the power-excess channel to each power-deficient channel to the existing photon flux of the power-excess channel; adjusting the adjustable Mach-Zehnder coupler according to a pre-calibrated correspondence between the driving voltage and the extraction ratio; inputting the power slice corresponding to the target extraction ratio into a photon migration bus; and outputting the remaining continuous wave carrier along the main optical path of the power-excess channel; when the maximum allowable source-side photon flux of a single frequency conversion path is less than the source-side photon flux migrating from the power-excess channel to the power-deficient channel, dividing the source-side photon flux into multiple sub-photon fluxes, and inputting the multiple sub-photon fluxes into parallel frequency conversion paths respectively.

[0014] Preferably, converting the power slice into the converted carrier includes: inputting the power slice, the first coherent pump light, and the second coherent pump light into a Bragg scattering four-wave mixer waveguide, wherein the frequency of the first coherent pump light is... The frequency of the second coherent pump light is The target optical frequency of the power excess channel is The target light frequency of the underpowered channel is and satisfy The frequency generated by the Bragg scattering four-wave mixing waveguide is... The converted carrier, the unconverted carrier, and the pump light are guided to their respective output ports by a wavelength selective filtering structure.

[0015] Preferably, adjusting the optical field phase of the converted carrier and combining the adjusted converted carrier into the main optical path of the underpowered channel includes: in an initial photon flux state lower than the photon flux of the target source side, inputting the converted carrier and the original continuous wave carrier of the underpowered channel into an adjustable coherent multiplexer, scanning the drive quantity of the phase corrector set in the propagation path of the converted carrier, acquiring the combined optical power at the output of the adjustable coherent multiplexer, and recording the phase corrector drive quantity corresponding to when the combined optical power reaches a local maximum value as the initial phase-construction drive quantity; during photon migration, determining the phase error based on the phase of the converted carrier relative to the coherent reference light corresponding to the underpowered channel and the phase of the original continuous wave carrier of the underpowered channel relative to the coherent reference light, and correcting the phase corrector drive quantity based on the phase error.

[0016] Preferably, determining the source-side photon flux corresponding to each migration edge includes: constructing a solution objective based on the common equalization power, the conversion loss of each migration edge, the pump power consumption of each migration edge, and the change in source-side photon flux between the current control cycle and the previous control cycle; using the output photon flux of each frequency node after photon migration satisfying the target photon flux corresponding to that frequency node as a node constraint; and using the upper and lower limits of the source-side photon flux of each migration edge, the minimum retained photon flux of each power excess channel, the upper limit of the total pump power, and the locking state of each phase correction branch as boundary constraints to obtain the source-side photon flux of each migration edge. Based on the source-side photon flux, the target driving quantity of the tunable Mach-Zehnder coupler, the target power of the first coherent pump light and the second coherent pump light, and the target driving quantity of the phase corrector are generated respectively.

[0017] Preferably, before executing the high-power DFB array light source equalization method, the following calibrations are performed sequentially: with all power slice extraction units kept in direct mode and all frequency conversion pumps turned off, the output transmission efficiency of each high-power DFB laser channel from the frequency lattice equalization unit inlet to the corresponding data modulator input is collected; the driving voltage of each tunable Mach-Zehnder coupler is adjusted step by step, and a calibration relationship between the driving voltage and the extraction ratio is formed based on the main optical path power of the corresponding channel and the photon migration bus input power; a continuous wave carrier with known optical power is input to each frequency conversion path, and the power and phase of the first coherent pump light and the second coherent pump light are adjusted to form a calibration relationship between the input optical power, pump power, frequency conversion unit temperature and frequency conversion efficiency; a low-power conversion carrier is input to the tunable coherent combiner of the corresponding underpowered channel, the phase corrector drive quantity is scanned and the synthesized optical power is recorded to form a calibration relationship between the phase corrector drive quantity and the synthesized optical power.

[0018] Preferably, during the execution of the high-power DFB array light source equalization method, the injection lock status of each high-power DFB laser channel, the frequency conversion efficiency of each frequency conversion path, the temperature of each frequency conversion unit, and the phase error of each phase correction branch are detected. When any high-power DFB laser channel loses injection lock, the extraction of power slices from that high-power DFB laser channel is stopped, and the addition of conversion carriers to that high-power DFB laser channel is stopped. When the frequency conversion efficiency of any frequency conversion path is lower than the corresponding lower limit of conversion efficiency or the temperature of the frequency conversion unit is higher than the corresponding upper limit of temperature, the frequency conversion path is deleted from the frequency migration map, and the source-side photon flux corresponding to each migration edge is re-determined. When the phase error of any phase correction branch continuously exceeds the corresponding upper limit of phase error, the power slice extraction ratio corresponding to that phase correction branch is reduced, and the initial phase construction drive is re-determined. When any high-power DFB laser channel stops outputting, the existing photon flux of the frequency node corresponding to that high-power DFB laser channel is set to zero, and the common equalization power is re-determined based on the existing photon flux of the remaining high-power DFB laser channels.

[0019] This invention provides a high-power DFB array light source equalization method. It has the following beneficial effects: 1. Compared with the equalization method of changing the DFB drive current, the present invention adjusts the optical domain carrier allocation while keeping each high-power DFB laser channel in an injection-locked state, thereby reducing junction temperature changes, emission frequency changes and thermal crosstalk between adjacent channels caused by changes in drive current.

[0020] 2. Compared with the equalization method of reducing the power of strong channels by using a variable optical attenuator, the present invention converts the surplus continuous wave carrier extracted from the power excess channel into the target frequency carrier of the power deficiency channel, so that the extracted photon flux enters the input terminal of the data modulator of the power deficiency channel.

[0021] 3. In this invention, only the power slice corresponding to the power difference is input to the frequency conversion unit. The main high-power carrier that does not participate in the migration propagates along the main optical path of the original channel, reducing the input power capacity and pump power requirements of the frequency conversion unit. Photon flux migration is completed before independent data modulation. Each conversion branch processes the continuous wave carrier that is not loaded with independent service data. After equalization, the supply channel and the receiving channel enter the corresponding data modulators respectively.

[0022] 4. This invention records the capacity, efficiency, pump cost and phase state of each migration edge through a frequency migration diagram. When a migration edge exits operation, the migration path is re-determined. When a DFB laser channel stops outputting, the supplementary carrier power at the target frequency is determined based on the surplus photon flux of the remaining channels. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the overall structure of the high-power DFB array light source equalization system of the present invention.

[0024] Figure 2 This is a schematic diagram of the connection structure of a single high-power DFB laser channel, a power slice extraction unit, and the main optical path of the channel.

[0025] Figure 3 A schematic diagram of the optical path structure for performing carrier photon migration from a power-excess channel to a power-insufficient channel.

[0026] Figure 4 This is a schematic diagram of the structure of a Bragg scattering four-wave mixing frequency conversion unit.

[0027] Figure 5 A schematic diagram of the acquisition structure for the effective output optical power and optical field phase of the channel.

[0028] Figure 6 This is a schematic diagram of the frequency shift graph.

[0029] Figure 7 This is a schematic diagram of the carrier phase acquisition and phase holding process.

[0030] Figure 8 This is a schematic diagram illustrating the startup, calibration, operation, and abnormal reconstruction process of the high-power DFB array light source equalization method of the present invention.

[0031] Figure 9 This is the process for generating coherent pump light according to the present invention.

[0032] The components include: 1. Coherent reference generation unit; 2. Seed light distribution unit; 3. High-power DFB array; 31. High-power DFB laser channel; 4. Power slice extraction unit; 5. Photon migration bus; 6. Frequency conversion unit; 61. Bragg scattering four-wave mixing waveguide; 62. Pump generation unit; 63. Wavelength selective filtering structure; 7. Phase correction unit; 8. Coherent injection unit at the same frequency; 9. Channel status monitoring unit; 10. Equalization controller; 11. Data modulator array; 12. Output optical network; 13. Actual usage end; 14. Channel main optical path; 15. Monitoring beam splitter. Detailed Implementation

[0033] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention. Throughout the text, channel numbers correspond one-to-one with corresponding frequency node numbers; English phrases in superscripts and subscripts are category identifiers, and do not represent exponentiation unless explicitly indicated as squares. Specific Implementation Example 1: like Figures 1 to 9 As shown, a high-power DFB array light source equalization method is proposed. The method consists of coherent reference establishment, DFB injection locking, light field acquisition at the actual usage location, frequency migration map establishment, common equalization power solution, power slice extraction, target frequency conversion, coherent merging at the same frequency, closed-loop correction, and abnormal path reconstruction.

[0035] A coherent reference generation unit generates multiple coherent reference beams with defined frequency intervals and phase relationships, and each coherent reference beam is injected into a corresponding high-power DFB laser channel. The free operating frequency of each high-power DFB laser channel is tuned to the locking range of the corresponding coherent reference beam, so that each high-power DFB laser channel outputs an unmodulated continuous wave carrier at the target light frequency.

[0036] Each unmodulated continuous wave carrier is input into an independent channel main optical path, and a power slicing extraction unit, a coherent injection unit, and a channel state monitoring branch are set before entering the corresponding data modulator. The channel state monitoring branch measures the effective output optical power at the input of each data modulator and the optical field phase of each continuous wave carrier relative to the corresponding coherent reference light.

[0037] The equalization controller calculates the existing photon flux based on the effective output optical power, target optical frequency, and output transmission efficiency of each channel, and defines each target optical frequency as a frequency node, and defines the area between two frequency nodes connected by the frequency conversion unit as a migration edge. Under the constraints of migration edge capacity, total pump power, minimum retained photon flux of power-excess channels, and phase-locked state, the equalization controller solves for the common equalization power and the source-side photon flux of each migration edge.

[0038] The power slice extraction unit extracts a portion of the continuous wave carrier corresponding to the source-side photon flux from the power-excess channel, while the unextracted main continuous wave carrier remains in the original channel's main optical path. The extracted power slice enters the photon migration bus and the frequency conversion unit. The frequency conversion unit generates a target frequency conversion carrier based on the frequency difference between the source and target channels. The phase correction unit adjusts the phase of the conversion carrier, and the coherent injection unit merges the conversion carrier into the original continuous wave carrier of the power-deficient channel.

[0039] The equalization controller operates on four time scales: a phase hold loop, a conversion efficiency loop, a power distribution loop, and a parameter update loop. The phase hold loop adjusts the phase correction, the conversion efficiency loop adjusts the frequency conversion pump power, the power distribution loop adjusts the power slice extraction ratio, and the parameter update loop corrects the output transmission efficiency, migration side capacity, and calibration data.

[0040] like Figure 1 As shown, the high-power DFB array light source implementing the method of the present invention includes a coherent reference generation unit 1, a seed light distribution unit 2, a high-power DFB array 3, a power slice extraction unit 4, a photon migration bus 5, a frequency conversion unit 6, a phase correction unit 7, a coherent injection unit at the same frequency 8, a channel state monitoring unit 9, an equalization controller 10, a data modulator array 11, an output optical network 12, and an actual user terminal 13.

[0041] The coherent reference generation unit 1 includes a narrow-linewidth reference laser, a reference laser driving circuit, a reference laser temperature control circuit, and an optical frequency comb generation structure. In this embodiment, the center wavelength of the narrow-linewidth reference laser is located in the 1550nm band, and the output linewidth is no greater than 100kHz. The optical frequency comb generation structure adopts an electro-optic modulation type optical frequency comb structure, which consists of an intensity modulator, two phase modulators, an RF clock source, and an RF power amplifier. The RF clock source provides the same frequency reference to each modulator, and the basic frequency spacing of the optical frequency comb is set to 100GHz.

[0042] The seed light distribution unit 2 includes a wavelength demultiplexer, a seed light power regulator for each channel, and an optical isolator for each channel. The wavelength demultiplexer guides each reference comb tooth in the optical frequency comb to its corresponding high-power DFB laser channel 31. The seed light power regulator adjusts the injected seed light power entering a single high-power DFB laser channel 31 to 0.1mW to 5mW. The optical isolator restricts the output light of the high-power DFB laser channel 31 from being reversed and entering the coherent reference generation unit 1.

[0043] The high-power DFB array 3 includes eight high-power DFB laser channels 31. Each high-power DFB laser channel 31 includes a DFB laser chip, a constant current driver, a thermoelectric cooler, a temperature sensor, a seed light injection terminal, and a continuous wave carrier output terminal. The rated continuous wave output power of each high-power DFB laser channel 31 is 50mW to 300mW. The resolution of the constant current driver is no higher than 0.1mA, and the control resolution of the thermoelectric cooler is no higher than 0.01℃.

[0044] like Figure 2As shown, the continuous wave carrier output of each high-power DFB laser channel 31 is connected to the corresponding channel main optical path 14. The channel main optical path 14 contains, in sequence, a power slice extraction unit 4, a coherent injection unit 8, a monitoring beam splitter 15, and a corresponding data modulator. The main output of the power slice extraction unit 4 is connected to the first input of the coherent injection unit 8, and the migration output of the power slice extraction unit 4 is connected to the photon migration bus 5. The second input of the coherent injection unit 8 is connected to the output of the phase correction unit 7.

[0045] The power slice extraction unit 4 employs an adjustable Mach-Zehnder coupler. The adjustable Mach-Zehnder coupler includes a first 3dB coupler, a first branch phase modulator, a second branch phase modulator, and a second 3dB coupler. The first branch phase modulator uses a carrier depletion-type phase modulator, and the second branch phase modulator uses a thermo-optical phase modulator. The continuous adjustment range of the power slice extraction ratio is 0% to 30%.

[0046] The photonic migration bus 5 employs a low-loss silicon nitride waveguide, connecting the migration outputs of each power slice extraction unit 4 to multiple frequency conversion units 6. Each frequency conversion unit 6 includes a Bragg scattering four-wave mixing waveguide 61, a pump generation unit 62, and a wavelength selective filtering structure 63. The Bragg scattering four-wave mixing waveguide 61 is a silicon nitride helical waveguide with an effective waveguide length of 2cm to 10cm and a propagation loss not exceeding 1dB / cm.

[0047] Pump generation unit 62 obtains a frequency reference from the reference clock of coherent reference generation unit 1 and outputs a first coherent pump light and a second coherent pump light. The two coherent pump lights are input to Bragg scattering four-wave mixing waveguide 61 via a power regulator and a polarization regulator, respectively. The optical power of a single pump light entering Bragg scattering four-wave mixing waveguide 61 is set to 100mW to 600mW. Wavelength selective filtering structure 63 uses a cascaded micro-ring filter and an arrayed waveguide grating to guide the converted carrier, unconverted carrier, and pump light to independent ports, with a pump isolation of not less than 40dB.

[0048] Phase correction unit 7 includes a thermo-optical coarse phase tuner and a carrier depletion-type fast phase tuner. The phase adjustment range of the thermo-optical coarse phase tuner covers... to The bandwidth of the fast phase converter is no less than 100MHz. The coherent injection unit 8 uses an adjustable Mach-Zehnder combiner to combine the converted carrier and the original continuous wave carrier of the underpowered channel into the same channel main optical path 14.

[0049] The beam splitter ratio of the monitoring beam splitter 15 is set to 0.2% to 2%. The channel status monitoring unit 9 includes a reference beam splitter, a 90-degree optical mixer, two sets of balanced photodetectors, a transimpedance amplifier, an analog-to-digital converter, and digital signal processing circuitry. The sampling clock of the analog-to-digital converter is synchronized with the radio frequency clock used by the coherent reference generation unit 1.

[0050] The equalization controller 10 employs a collaborative architecture of a field-programmable gate array (FPGA) and an embedded processor. The FPGA performs multi-channel synchronous sampling, digital down-conversion, in-phase quadrature component calculation, phase expansion, and fast phase loop closure. The embedded processor performs calibration data management, frequency migration map management, common equalization power solution, migration path allocation, state machine operation, and exception handling.

[0051] The output of coherent reference generation unit 1 is the first The target light frequency of each reference comb tooth satisfies: ; In the formula, Indicates the first The target light frequency of each high-power DFB laser channel 31 is in Hertz. This indicates the reference light frequency of a narrow-linewidth reference laser, measured in Hertz. Indicates the first The integer frequency slot number corresponding to each high-power DFB laser channel 31; The basic frequency interval of the optical frequency comb is represented by Hertz. This indicates the number of high-power DFB laser channel 31. The value range is 1 to ; This indicates the total number of high-power DFB laser channels 31.

[0052] No. The complex optical field of each reference comb tooth is represented as follows: ; In the formula, Indicates the first A reference comb tooth in time The complex optical field; Indicates the first The optical field amplitude of each reference comb tooth; Represents the natural exponential function; Represents the imaginary unit, and ; Represents pi; Indicates the first The target optical frequency of each reference comb tooth; Indicates time, in seconds; Indicates the first The reference comb teeth are in the reference phase relative to the common radio frequency clock, in radians. The optical field described above is represented using power normalization, and the amplitude of the optical field is in the square root of the watts; the superscript ref indicates only the coherent reference light.

[0053] The RF clock of the coherent reference generation unit 1 is simultaneously input to the pump generation unit 62 and the equalization controller 10. The frequency difference between each reference comb tooth, the frequency difference between each pump light, and the digital sampling clock are all derived from the same RF clock, forming a unified frequency reference and phase reference.

[0054] Seed light distribution unit 2 will... Each reference comb tooth input corresponds to a high-power DFB laser channel 31. The equalization controller 10 maintains the drive current of the high-power DFB laser channel 31 within a preset injection lock-in current range and adjusts the set temperature of the thermoelectric cooler so that the free operating frequency of the high-power DFB laser channel 31 enters the injection lock-in range of the corresponding reference comb tooth.

[0055] The first day after injection locking is completed The output light field of each high-power DFB laser channel 31 is represented as follows: ; In the formula, Indicates the first One high-power DFB laser channel 31 in time The output complex optical field; Indicates the first The output optical field amplitude of each high-power DFB laser channel 31; Indicates the first The target light frequency locked by a high-power DFB laser channel 31; Indicates the reference phase corresponding to the reference comb tooth; The symbol represents the seed light transmission path and the additional phase formed by the DFB resonator at the DFB output end, in radians; the remaining symbols have the same meaning as in the aforementioned formula.

[0056] The equalization controller 10 determines the injection lock state by combining the beat frequency residual, phase error variance, and output spectrum. The beat frequency residual is obtained from the beat frequency signal of the monitoring light of the high-power DFB laser channel 31 and the corresponding reference comb on the balanced photodetector. The phase error variance is calculated in terms of length... Calculated within a sliding window of sampling points. This indicates the number of sampling points in the injection lock state determination window. The output spectrum is acquired by the spectral monitoring port connected during the startup phase.

[0057] When the absolute value of the beat frequency residual is not higher than the preset locked frequency threshold, the phase error variance is not higher than the preset locked phase variance threshold, and the main peak of the output spectrum is located within the target frequency window, the equalization controller 10 marks the high-power DFB laser channel 31 as a locked channel. Only locked channels are entered into the frequency shift map.

[0058] like Figure 5 As shown, the monitoring spectrometer 15 monitors the beam splitter from the first... The monitoring light is split at the input of each data modulator, and the corresponding reference comb teeth enter the 90-degree optical mixer via a reference light distributor. The four outputs of the 90-degree optical mixer are respectively fed into two sets of balanced photodetectors, and after transimpedance amplification and analog-to-digital conversion, they form in-phase detection components. and orthogonal detection components .

[0059] No. The estimated complex optical field values ​​for each channel are expressed as follows: ; In the formula, Indicates the first The estimated complex optical field values ​​of each channel at the input of the data modulator relative to the corresponding reference comb, before power amplitude calibration; Indicates the first The in-phase detection voltage components of each channel, in volts; Indicates the first The orthogonal detection voltage components of each channel, in volts; The unit represents the imaginary unit. Both detected components are voltage values ​​calculated from the analog-to-digital conversion results after bias, gain, and quadrature error correction. The unit of this complex optical field estimate is also volts.

[0060] No. The estimated phase value of the optical field for each channel is expressed as follows: ; In the formula, Indicates the first Each channel has its optical field phase relative to the corresponding reference comb at the data modulator input, in radians; Indicates based on orthogonal detection components and in-phase detection component A bivariate arctangent function that determines the phase of a complete quadrant.

[0061] No. The estimated effective output optical power of each channel is expressed as follows: ; In the formula, Indicates the first Estimated effective output optical power of each channel at the input of the data modulator, in watts; Indicates the first The optical power calibration factor for each channel is expressed in watts per square volt. Indicates the in-phase detection component The square of; Represents orthogonal detection components The square of. The estimated value is jointly calibrated by the monitoring splitting ratio, the reference comb optical power, the balanced photodetector responsivity, the transimpedance amplification gain, and the analog-to-digital conversion ratio. This estimated value corresponds to the optical power at the input of the data modulator as defined below. .

[0062] 10 pairs of equalization controllers and The process sequentially performs DC bias subtraction, channel gain correction, quadrature phase error correction, unified timestamp alignment, saturation sampling rejection, and low-pass filtering. The unified timestamp is derived from the synchronous sampling counter of the field-programmable gate array. The phase expansion algorithm separates adjacent sampling points that exceed [a certain threshold]. phase jump variable minus , will be less than The phase jump variable increases .

[0063] With all power slice extraction units 4 kept through, all frequency conversion pumps turned off, and all coherent injection branches turned off, the equalization controller 10 measures the optical power at the entrance of the frequency lattice equalization unit and the optical power at the corresponding data modulator input of each high-power DFB laser channel 31.

[0064] No. The output transmission efficiency of each channel is expressed as: ; In the formula, Indicates the first The output transmission efficiency of each channel from the entrance of the frequency lattice equalization unit to the corresponding data modulator input; Indicates the first The optical power measured at the corresponding data modulator input of each channel, in watts; Indicates the first The optical power measured at the entrance of the frequency lattice equalization unit of each channel is expressed in watts.

[0065] Output transmission efficiency calibration is performed at multiple input optical power points. The equalization controller 10 records the operating points corresponding to 30%, 50%, 70%, and 100% of the rated output power, respectively. This forms a piecewise linear calibration table between the input optical power and the output transmission efficiency. During normal operation, the equalization controller 10 performs linear interpolation in the calibration table based on the current input optical power.

[0066] No. The output transmission efficiency of each channel is expressed as: ; In the formula, Indicates the first The output transmission efficiency of each channel from the entrance of the frequency lattice equalization unit to the corresponding data modulator input; Indicates the first The optical power measured at the corresponding data modulator input of each channel, in watts; Indicates the first The optical power measured at the entrance of the frequency lattice equalization unit of each channel is expressed in watts. This formula is calculated only under the above-mentioned no-migration calibration condition, with both optical power values ​​measured at the same operating point; the output transmission efficiency is dimensionless.

[0067] No. The effective output optical power at the input of the data modulator after photon migration in each channel is expressed as: ; In the formula, Indicates the first The effective output optical power of each channel at the input of the data modulator is expressed in watts. Denotes Planck's constant; Indicates the first The target optical frequency of each channel; Indicates the first Output transmission efficiency of each channel; Indicates the first The equivalent output photon flux, measured in seconds, is calculated from the photon migration at each frequency node and referred to the reference plane at the entrance of the equalization unit of that channel. The output transmission efficiency is only considered once when converting this equivalent photon flux to the power at the input of the data modulator.

[0068] The equalization controller 10 records the data of each frequency node as node status. The node status includes node number, target optical frequency, current photon flux, minimum reserved photon flux, output transmission efficiency, injection lock status, current phase, current effective output optical power, and allowable power deviation.

[0069] like Figure 6As shown, the equalization controller 10 defines each target optical frequency as a frequency node, and defines a directed migration edge between the source frequency node and the target frequency node connected by a frequency conversion unit 6. When the same hardware frequency conversion unit 6 undertakes reverse conversion after pump frequency switching, another migration edge with the opposite direction is established in the frequency migration graph. Figure 6 In the diagram, v1 to v8 are merely graphical labels for the first to eighth frequency nodes and do not represent optical frequency variables; each bidirectional line represents two directed migration edges in opposite directions. This represents the capacity from the first node to the second node. for The abbreviation is used for the following numbers, and the remaining numerical subscripts are interpreted according to the same rules; the capacity and efficiency of the reverse migration edge are marked separately and are not considered equal because they share a graph connection.

[0070] The data record for each migration edge includes the source node number, target node number, maximum allowable source-side photon flux, power slice extraction efficiency, frequency conversion efficiency, migration bus transmission efficiency, coherent injection efficiency, pump power coefficient, conversion-related noise, frequency conversion unit temperature range, and phase correction status.

[0071] No. The frequency node to the first The total photon transmission efficiency during the migration of frequency nodes is expressed as: ; In the formula, Indicates the first The frequency node to the Total photon transmission efficiency of each frequency node; Indicates the power slice extraction efficiency; Indicates frequency conversion efficiency; This indicates the transmission efficiency of the converted carrier in the photonic migration bus 5; Indicates the conversion carrier is incorporated into the first Coherent injection efficiency of the main optical path of each channel at time 14; and All of these represent frequency node numbers, and their order in the migration direction determines the source node and the target node. The above efficiencies are all defined in terms of photon flux ratio, which is a dimensionless quantity; extraction, conversion, routing, and merging correspond to non-repeating calibration segments, and the total photon transmission efficiency is calibrated by the ratio of the equivalent arrival flux at the target node's entrance reference plane to the migration flux at the source node, without repeatedly including the output transmission efficiency of the target channel.

[0072] No. The output photon flux after photon migration at each frequency node is expressed as: ; In the formula, Indicates the first The equivalent output photon flux of each frequency node on the entrance reference plane after the migration is completed; Indicates the first Existing photon flux before the migration of each frequency node; Indicates the first The frequency node is for the first Photon flux subtracted from the source node's inlet reference plane during frequency node migration; Indicates by the first The frequency node to the first Total photon transmission efficiency during frequency node migration; Indicates by the first The frequency node migrated to the first Source-side photon flux at each frequency node; Represents the total number of frequency nodes; summation symbol This indicates that the values ​​within the listed node number range are summed. The unit for each photon flux is per second. When the source node and target node are the same, or there is no migration edge between the two nodes or the migration edge is unavailable, the corresponding source-side photon flux is zero.

[0073] The equalization controller 10 employs a deterministic solution algorithm that combines the search for the most feasible common power with the allocation of the minimum cost photon flux. This algorithm does not use training data and does not rely on neural network parameters. The algorithm inputs include the existing photon flux of each node, the output transmission efficiency, the minimum retained photon flux, the total photon transmission efficiency of each migration edge, the migration edge capacity, the pump power coefficient, and the phase correction state.

[0074] First, determine the lower bound of the common equalization power based on the minimum value among the current effective output optical power of all channels. The common equalization power upper bound is determined based on the average effective optical power that can be generated by all existing photon fluxes under the output transmission efficiency of each channel. . This represents the lower bound of the common equilibrium power search interval. This represents the upper bound of the common equilibrium power search interval.

[0075] In each search, the candidate common equalization power is taken: ; In the formula, This indicates the candidate common equalization power used in this feasibility assessment; Indicates the current lower bound of the search; Indicates the upper bound of the current search.

[0076] No. The target photon flux of each frequency node under the candidate common equalization power is expressed as: ; In the formula, Indicates the first Target photon flux at each frequency node under candidate common equalization power; Indicates the candidate common equalization power; Denotes Planck's constant; Indicates the first The target optical frequency of each frequency node; Indicates the first Output transmission efficiency of each channel.

[0077] No. The photon flux that can be supplied to the source side of each frequency node is expressed as: ; In the formula, Indicates the first Each frequency node can supply photon flux at the source side under the candidate common equalization power; Indicates in and Take the larger value; Indicates the first The existing photon flux of each frequency node; Indicates the first Target photon flux at each frequency node; This represents zero photon flux.

[0078] No. The target-side photon flux required for each frequency node is expressed as: ; In the formula, Indicates the first The target-side photon flux required by each frequency node under the candidate common equalization power; the remaining symbols have the same meaning as in the aforementioned formula.

[0079] Equalizer 10 will Frequency nodes with positive values ​​are set as supply nodes. Frequency nodes with positive values ​​are designated as recipient nodes, and successive minimum-cost augmentation is performed in the residual network of the frequency migration graph. Each augmentation starts from a supplier node that still has a supply of photon flux, selects a path to the recipient node based on the comprehensive cost of reaching photon flux per unit of migration edge, and takes the minimum value among path capacity, supply margin, recipient demand, and pump margin as the source-side photon flux for this augmentation.

[0080] No. The frequency node to the first The comprehensive cost of migration edges for each frequency node is expressed as: ; In the formula, Indicates the first The frequency node to the The dimensionless migration edge comprehensive cost of each frequency node; Indicates the weight of photon transmission loss; Indicates the total photon transmission efficiency at the migration edge; Indicates pump power weight; This represents the pump power coefficient corresponding to a unit source-side photon flux during migration, in watt-seconds. Indicates the added noise weight; This represents the normalized additional noise parameter of the migration edge; Indicates the weight of changes in migration amount; This represents the source-side photon flux to be allocated in the current control cycle; This represents the source-side photon flux that has been executed and confirmed at the corresponding migration edge in the previous control cycle; absolute value symbol. This indicates that the value within the parentheses is the non-negative absolute value. The photon transmission loss weight and the additional noise weight are dimensionless, the pump power weight is in the reciprocal of the watt-second, and the migration change weight is in the second. Therefore, the dimensions of each summation term in the overall cost are consistent; each weight is a non-negative value.

[0081] When the required photon flux of all receiving nodes is met and the total pump power does not exceed the pump power limit, the candidate common equalization power is... It was deemed feasible, so When there are unmet demands or pump power exceeding limits, the candidate common equalization power... Deemed infeasible, .when When the resolution is not higher than the preset solution resolution, take As a common equalization power .

[0082] The feasibility assessment must simultaneously satisfy the following constraints: ; In the formula, Indicates the first The frequency node to the first Source-side photon flux of each frequency node migration; This represents the maximum allowed source-side photon flux for the corresponding migration edge.

[0083] ; In the formula, Indicates the first The sum of the source-side photon fluxes that migrate from each frequency node to all target frequency nodes; Indicates the first The existing photon flux of each frequency node; Indicates the first Each frequency node must retain the lowest photon flux of the original channel main optical path 14.

[0084] ; In the formula, Indicates the first The frequency node to the first The equivalent pump power coefficient corresponding to a unit source-side photon flux at a frequency node migration, in watt-seconds; This represents the source-side photon flux of the corresponding migration edge; This represents the upper limit of the total pump power allowed to be output by the pump generation unit; the double summation represents the accumulation of the migration edges corresponding to all source nodes and target nodes. The product of this coefficient and the source-side photon flux is the pump optical power corresponding to that migration edge, in watts; here, the pump power is optical power, not the electrical power consumption of the pump driver.

[0085] The algorithm output includes common equalization power. Photon flux at each migrating edge source side The target extraction ratio of each power slice extraction unit 4, the target frequency combination of each frequency conversion unit 6, the power of the two coherent pump light, and the initial phase correction amount of each phase correction unit 7.

[0086] No. The total source-side photon flux that needs to be extracted from each power excess channel is expressed as: ; In the formula, Indicates the first The total source-side photon flux that needs to be extracted from each power excess channel; Indicates the first The frequency node to the first Source-side photon flux of each frequency node migration; This indicates the total number of frequency nodes.

[0087] No. The target extraction ratio of each power slice extraction unit 4 is expressed as follows: ; In the formula, Indicates the first The target extraction ratio of each power slice extraction unit 4; Indicates the first The total source-side photon flux that needs to be extracted from each power excess channel; Indicates the first The existing photon flux at each frequency node.

[0088] No. The photon flux retained in the main optical path 14 of each channel is expressed as follows: ; In the formula, Indicates the first After the power slice extraction is completed in each channel, the photon flux retained in the main optical path 14 of the channel is as follows: This represents the proportion of photon flux that was not extracted; the other symbols have the same meaning as in the aforementioned formula.

[0089] The equalization controller 10, based on the drive voltage of the adjustable Mach-Zehnder coupler and the extraction ratio calibration table, will... The voltage is converted into a driving voltage. The driving voltage changes stepwise according to a preset slope. After each change, the system waits for at least five phase hold loop cycles before reading the power of the main optical path and the power of the migration branch. When the deviation between the actual extraction ratio and the target extraction ratio exceeds the allowable error of the extraction ratio, the equalization controller 10 corrects the driving voltage.

[0090] When the maximum permissible source-side photon flux of a single frequency conversion path is lower than At that time, the equalization controller 10 will Split into Individual photon flux: ; In the formula, Indicates the first The frequency node to the first Total source-side photon flux of each frequency node migration; Indicates the number of parallel frequency conversion paths; Indicates the first The source-side photon flux carried by the parallel frequency conversion paths; Indicates the parallel frequency conversion path number.

[0091] like Figure 4 As shown, the frequency is Power slice, frequency is The first coherent pump light and frequency are The second coherent pump light is input to the Bragg scattering four-wave mixing waveguide 61. The pump frequency satisfies: ; In the formula, This indicates the frequency of the first coherent pump light; This indicates the frequency of the second coherent pump light; Indicates the target optical frequency of the channel with excess power; This indicates the target optical frequency of the channel with insufficient power.

[0092] Within the calibration operating range where phase matching conditions are met and pump depletion is not significant, the initial estimate of the frequency conversion efficiency is expressed as: ; In the formula, Indicates the first The frequency node to the first The initial estimate of the frequency conversion efficiency under the approximate condition during the migration of a frequency node is a dimensionless quantity. Represents the sine function; This represents the equivalent nonlinear coupling coefficient of the corresponding nonlinear waveguide and frequency combination, expressed in units of watt per meter. This represents the effective nonlinear action length after considering propagation loss, in meters. This represents the optical power of the first coherent pump light entering the nonlinear region, expressed in watts. This represents the optical power of the second coherent pump light entering the nonlinear action region, measured in watts; square root symbol. This indicates taking the square root of the non-negative value within the parentheses. This initial estimate is only used to select the pump start-up power; the frequency conversion efficiency in the total photon transmission efficiency is taken as the calibrated value at the corresponding operating point, with the capped symbol indicating the measured value and the asterisk symbol indicating the target value, and each of the three uses a different complete symbol.

[0093] The initial estimate is only used to determine the starting point of the pump power. During the factory calibration phase, the frequency conversion unit 6 generates a multi-dimensional calibration table relating the input power, the first pump power, the second pump power, the device temperature, and the frequency conversion efficiency. During normal operation, the equalization controller 10 performs multilinear interpolation based on the current power slice optical power, device temperature, and target frequency conversion efficiency in the calibration table to obtain the two pump power setpoints.

[0094] The actual frequency conversion efficiency is expressed as: ; In the formula, Indicates the first The frequency node to the Measured frequency conversion efficiency of each frequency node migration path; This represents the converted photon flux measured at the target frequency output terminal of the wavelength-selective filter structure 63. This represents the measured input photon flux at the entrance of the Bragg scattering four-wave mixing waveguide 61. Both photon fluxes are measured in seconds and are obtained by dividing the optical power at the corresponding measurement point by the single-photon energy at the corresponding frequency. This input flux includes extraction and propagation losses before the entrance, and the outflow flux at the source node entrance reference plane. Defined separately; when the input flux is zero, this ratio is not calculated, and the efficiency measurement is marked as invalid.

[0095] Frequency conversion efficiency error is expressed as: ; In the formula, Indicates the first The frequency node to the Frequency conversion efficiency error of each frequency node migration path; This indicates the target frequency conversion efficiency of the corresponding migration path; This indicates the measured frequency conversion efficiency.

[0096] Equalization controller 10 according to The power of the first and second coherent pump lights is adjusted in limited steps. Each pump power increment does not exceed 1% of the corresponding rated pump power, and the pump power is stopped when the temperature of the frequency conversion unit exceeds the calibrated temperature range.

[0097] like Figure 7 As shown, before executing the target migration photon flux, the equalization controller 10 sets the power slice extraction ratio to 5% of the target extraction ratio, forming a low-power conversion carrier. The low-power conversion carrier is input to the coherent injection unit 8 via the phase correction unit 7 and synthesized with the original continuous wave carrier of the underpowered channel.

[0098] No. The frequency node is switched to the first The converted optical field of each frequency node is represented as: ; In the formula, Indicates the first The frequency node is converted to the first Complex optical field envelope at each frequency node; Indicates the amplitude of the converted optical field; Indicates the phase of the converted optical field, in radians; The imaginary unit is represented. This envelope is defined at the converted carrier input of the coherent injection unit 8, with the coherent reference light of the target channel as the common phase reference; the optical field amplitude adopts the same power normalization method as the aforementioned complex optical field.

[0099] No. The complex optical field of the original continuous wave carrier in each channel is represented as: ; In the formula, Indicates the first The complex optical field envelope of the original continuous wave carrier of each channel; This represents the optical field amplitude of the original continuous wave carrier. This represents the optical field phase of the original continuous wave carrier, in radians. This envelope is defined at the original carrier input of the coherent injection unit 8 at the same frequency, and uses the same phase reference and amplitude normalization method as the converted optical field envelope.

[0100] Phase error is expressed as: ; In the formula, Indicates the transformation of the light field and the first The phase error between the original optical fields of each channel is expressed in radians. This indicates that the phase value within the parentheses is mapped to... Phase reversal operation within an interval; Indicates the phase of the converted optical field; Indicates the first The original optical field phase of each channel; This represents the target phase difference when the coherent injection unit 8 at the same frequency obtains the maximum synthesized optical power.

[0101] When phase acquisition is initiated, the equalization controller 10 scans the drive voltage of the thermo-optical coarse phase tuner, in a... The synthesized optical power at the output of the coherent injection unit 8 is recorded within the phase period, and the driving voltage corresponding to the point of maximum synthesized optical power is recorded as the initial phase-construction driving quantity. Subsequently, the proportional-integral closed loop of the fast phaser is started.

[0102] The phase corrector drive quantity is expressed as: ; In the formula, Indicates the first Phase corrector drive voltage per phase control cycle, in volts; This represents the initial phase-length drive voltage recorded at the end of phase capture, in volts. This represents the phase ratio control coefficient, with units of volts per radian. Indicates the first Phase error per phase control cycle, in radians; This represents the discrete phase integral control coefficient that has been factored into the fixed sampling period, expressed in volts per radian. Indicates the phase control cycle number. This indicates the cumulative error index within the control cycle range; both are non-negative integers. Indicates from the 0th to the 1st The cumulative phase error over each phase control cycle. The feedback direction of this positional proportional-integral relationship is determined based on the calibration results of the phase corrector's drive voltage and phase change.

[0103] When the absolute value of the phase error is no higher than 0.1 rad for ten consecutive phase control cycles, the equalization controller 10 marks the corresponding phase correction branch as locked and gradually increases the power slice extraction ratio in steps not exceeding 5% of the target extraction ratio per level. When the absolute value of the phase error is higher than 0.3 rad for three consecutive phase control cycles, the equalization controller 10 reduces the power slice extraction ratio to 20% of the current value and re-executes phase acquisition.

[0104] The update frequency of the phase-holding loop is set to 10kHz to 1MHz. The input data is the phase error between the converted carrier and the original continuous wave carrier of the underpowered channel, and the output data is the fast phaser drive quantity. The phase-holding loop does not change the power slice extraction ratio or the pump light power.

[0105] The conversion efficiency loop's update frequency is set from 1kHz to 10kHz. The input data is the difference between the target frequency conversion efficiency and the measured frequency conversion efficiency, and the output data is the power settings for the first and second coherent pump lights. The conversion efficiency loop only increases the pump power when the phase correction branch is locked.

[0106] The update frequency of the power distribution loop is set to 100Hz to 2kHz. The input data are the effective output optical power, frequency conversion efficiency, output transmission efficiency, migration edge capacity and pump power margin of each channel. The output data are the common equalization power, the source-side photon flux of each migration edge and the extraction ratio of each power slice.

[0107] The update cycle of the parameter update loop is set from 1s to 600s. The input data are long-term average output optical power, path loss, frequency conversion efficiency, pump power margin and injection lock margin. The output data are output transmission efficiency correction value, migration edge capacity correction value and calibration table correction value.

[0108] The data used in this invention are divided into factory calibration data, startup status data, real-time operation data, and long-term status data. The factory calibration data is stored according to the device serial number and includes the target light frequency, seed light injection power range, injection lock-in temperature range, power slice extraction calibration table, frequency conversion efficiency calibration table, phase correction calibration table, output transmission efficiency, migration edge capacity, pump power boundary, and monitoring light power calibration coefficient.

[0109] The startup status data is collected each time the array is powered on, including the reference comb frequency and power, the beat frequency residual of each high-power DFB laser channel 31, the phase error variance, the input optical power, the optical power at the data modulator input, the temperature of each frequency conversion unit 6, and the initial drive quantity of each phase correction unit 7. The startup status data is used to confirm whether the current operation is within the coverage range of the factory calibration data.

[0110] Real-time operational data is acquired by a field-programmable gate array (FPGA) according to a unified sampling clock, including... , The data includes the power slice extraction ratio, input optical power of the conversion branch, output optical power at the target frequency, pump power, pump phase, frequency conversion unit temperature, phase corrector drive quantity, and injection lock status. Each real-time running data entry is accompanied by the sampling time, channel number, frequency node number, and migration edge number.

[0111] Long-term state data is statistically analyzed by the embedded processor according to a fixed time window. It includes average effective output optical power, rate of change of output transmission efficiency, rate of change of frequency conversion efficiency, rate of change of pump power required to achieve the same conversion efficiency, and rate of change of average drive quantity of the phase corrector. The long-term state data only corrects the migration side capacity and calibration parameters, and does not directly enter the fast phase closure loop.

[0112] The equalization controller 10 performs validity marking on the acquired data. When the photodetector is saturated, the DFB laser channel is unlocked, the pump light is unstable, the frequency conversion unit temperature exceeds the calibration range, or the analog-to-digital converter experiences sampling loss, the corresponding data is marked as invalid. Invalid data is not included in the common equalization power solution. Channels lacking valid data for three consecutive power allocation cycles are isolated from the current frequency shift graph.

[0113] like Figure 8 As shown, after the array is powered on, the equalization controller 10 first sets all power slice extraction units 4 to the pass-through state, sets all frequency conversion pumps to the off state, and sets all coherent injection branches to the off state. After the coherent reference generation unit 1 reaches a stable temperature state, the equalization controller 10 checks the reference light frequency, the number of reference comb teeth, the reference comb tooth spacing, and the power of each reference comb tooth.

[0114] After the reference light check is completed, the equalization controller 10 performs seed light injection and DFB chip temperature scanning channel by channel. After each high-power DFB laser channel 31 enters the injection lock state, it maintains the corresponding drive current and temperature settings. After all the high-power DFB laser channels 31 participating in equalization have completed injection lock, the equalization controller 10 collects the entrance optical power, data modulator input optical power, and optical field phase in the non-migration state.

[0115] During power slice extraction and calibration, the equalization controller 10, with the frequency conversion pump off, gradually increases the adjustable Mach-Zehnder coupler drive voltage from the minimum value to the maximum calibration value. At each voltage point, the power of the main optical path and the power of the migration branch are collected simultaneously to form a three-dimensional calibration table between the drive voltage, device temperature and actual extraction ratio.

[0116] During frequency conversion calibration, the equalization controller 10 inputs a continuous wave carrier with known optical power to the frequency conversion unit 6, adjusts the two coherent pump optical powers according to the preset pump power grid, and records the target frequency output optical power, unconverted carrier optical power, and pump leakage optical power at multiple frequency conversion unit temperature points to form a five-dimensional calibration table between input optical power, first pump power, second pump power, device temperature, and frequency conversion efficiency.

[0117] During phase correction calibration, the equalization controller 10 generates a low-power conversion carrier at 5% of the target migrated photon flux, scans the driving voltage of the thermo-optical coarse phase tuner, records the synthesized optical power at the output of the coherent injection unit 8 at the same frequency, records the position of the maximum synthesized optical power within one phase period as the initial phase-contraction driving amount, and records the position of the minimum synthesized optical power as the phase-destruction verification position.

[0118] After the above calibration is completed, the equalization controller 10 performs test migration with a source-side photon flux not exceeding 2% of the existing photon flux of a single channel. After the test migration satisfies the calibration relationship between the reduction in photon flux of the main optical path of the supply channel, the photon flux of the target frequency conversion carrier, and the increase in the effective output optical power of the yield channel, the system enters the normal equalization state.

[0119] Under normal operating conditions, the equalization controller 10 collects the effective output optical power and optical field phase at the input of each data modulator according to the power distribution cycle. Based on the output transmission efficiency and the migration amount executed in the previous control cycle, it converts the effective output optical power back to the existing photon flux before node migration and updates the frequency conversion efficiency, coherent injection efficiency, and pump power margin of each migration edge. The conversion relationship is as follows: ; Among them, the migration amount with the superscript "pre" is the migration amount of the previous cycle that has been executed and confirmed at the sampling time. The efficiency used corresponds to the operating point of the migration amount. The meanings of the other symbols are the same as those in the previous text.

[0120] The equalization controller 10 performs a maximum feasible common power search and minimum cost photon flux allocation to obtain the common equalization power, the photon flux on the source side of each migration edge, and the target values ​​for each actuator. For a newly enabled migration edge, the equalization controller 10 first sets the frequency and initial power of the two coherent pump beams, and then performs phase capture with a power slice of 5% of the target source-side photon flux. After the phase correction branch enters the locked state, the power slice extraction ratio and pump power are gradually increased.

[0121] After each level of power slice extraction ratio is adjusted, the equalization controller 10 checks the decrease in effective output optical power of the supply channel, the target frequency conversion carrier optical power, and the increase in effective output optical power of the receiving channel. When the deviation of these three from the predicted relationship of the total photon transmission efficiency of the migration side exceeds 5%, the increase in the power slice extraction ratio is stopped, and the extraction ratio calibration, pump frequency difference, frequency conversion unit temperature, filter center frequency, and phase error are checked in sequence.

[0122] Once the effective output optical power of each channel enters the allowable power range corresponding to the common equalization power, the equalization controller 10 maintains the current migration path and compensates for phase drift and conversion efficiency changes with a phase preservation loop and a conversion efficiency loop. The power distribution loop only recalculates the migration photon flux when the effective output optical power of any channel deviates from the allowable power range, the migration edge efficiency changes by a preset amplitude, or the array operating mode is switched.

[0123] Exception handling and path reconstruction: When any high-power DFB laser channel 31 loses injection lock, the equalization controller 10 reduces the power slice extraction ratio connected to that channel to zero within one power allocation cycle, shuts down the coherent injection branch that merges the converted carrier into that channel, and isolates that frequency node from the current frequency migration graph. The underlying DFB controller re-executes injection lock, and after the lock is restored, it is added back to the frequency migration graph according to the test migration procedure.

[0124] When the measured frequency conversion efficiency of any frequency conversion unit 6 is lower than the corresponding lower limit of conversion efficiency for five consecutive conversion efficiency cycles, the equalization controller 10 checks the pump frequency, pump power, nonlinear waveguide temperature, and input polarization. If the state does not recover, the corresponding migration edge is marked as unavailable, and an alternative path is searched in the residual frequency migration graph.

[0125] When the phase error of any phase correction branch exceeds 0.3 rad for three consecutive phase hold loop periods, the equalization controller 10 reduces the corresponding power slice extraction ratio to 20% of the current value, rescans the coarse phase tuner, and restores the fast phase closure loop. If phase acquisition fails three times consecutively, the corresponding migration edge is closed.

[0126] When any high-power DFB laser channel 31 stops outputting, the equalization controller 10 sets the existing photon flux of that frequency node to zero and retains that frequency node as the receiving node. The equalization controller 10 recalculates the common equalization power based on the surplus photon flux of the remaining nodes, the migration edge capacity, and the pump power margin. If the target photon flux of all nodes cannot be satisfied, the common equalization power is reduced until a migration scheme that satisfies all constraints is obtained.

[0127] When the total pump power of the pump generation unit 62 approaches the upper limit of the pump power, the equalization controller 10 does not continue to increase the pump power, but instead reduces the candidate common equalization power and re-executes the photon flow allocation. When the temperature of the frequency conversion unit is higher than the corresponding upper limit of the temperature, the two pump lights of the frequency conversion unit are immediately turned off. Specific Implementation Example 2: This embodiment employs eight high-power DFB laser channels 31, with target light frequencies arranged at 100 GHz intervals and each channel having a rated output optical power of 250 mW. The eight high-power DFB laser channels 31 are respectively injected and locked to the eight reference combs output by the coherent reference generation unit 1. An equalization structure is positioned before the eight data modulators, and each channel loads independent service data after photon migration equalization.

[0129] The initial effective output optical powers of the eight channels at the data modulator input are 205mW, 192mW, 214mW, 181mW, 207mW, 198mW, 188mW, and 211mW, respectively. The output transmission efficiencies of each channel are 0.91, 0.90, 0.92, 0.88, 0.91, 0.90, 0.89, and 0.92, respectively. The total photon transmission efficiency calibrated for each adjacent frequency shift edge is 0.83 to 0.88, and the total photon transmission efficiency calibrated for a shift edge spanning two frequency slots is 0.78 to 0.82.

[0130] The equalization controller 10 performs a common equalization power search based on the effective output optical power, target optical frequency, and output transmission efficiency of each channel. Using a total photon transmission efficiency of 0.85 on the main migration edge, the calculated common equalization power is 198.7 mW. Channels 3, 5, 8, and 1 are designated as supply channels, while channels 4, 7, 2, and 6 are designated as receiving channels.

[0131] The equalization controller 10 assigns migration tasks to channels 3 to 4, 8 to 7, 5 to 2, and 1 and 5 to 6. Each migration task first performs phase capture with a power slice of 5% of the target migration amount. Once the absolute value of the phase error is within 0.1 rad, the extraction ratio is gradually increased. After the effective output optical power of each channel enters the allowable power range corresponding to 198.7 mW, the migration path is maintained.

[0132] After equalization, the driving current of the eight high-power DFB laser channels 31 remains within the injection-locked current range, and the channel power deviation at the data modulator input does not exceed 2% of the common equalization power. The specific optical power values ​​in this embodiment are used to illustrate the control steps and parameter relationships and do not limit the scope of protection of this invention. Specific Implementation Example 3: In this embodiment, a channel with excess power needs to migrate a 40mW source-side continuous wave carrier to a channel with insufficient power. The rated input slice optical power of a single frequency conversion unit 6 is 15mW. The equalization controller 10 distributes the source-side photon flux to three parallel frequency conversion paths, with input optical powers of 14mW, 13mW, and 13mW for the three paths, respectively.

[0134] The pump frequency differences of the three frequency conversion paths all satisfy the frequency difference relationship between the source channel and the target channel. The three converted carriers are phase-acquired by independent phase correction units 7, and then sequentially merged with the original continuous wave carrier of the target channel in two-stage coherent injection units 8. The first stage merges with the first converted carrier, the second stage merges with the second converted carrier, and the third stage merges with the third converted carrier.

[0135] The equalizer controller 10 detects the phase of the three converted carriers relative to the corresponding reference comb teeth of the target channel, ensuring that the three converted carriers are in a constructive in-phase state with the original continuous wave carrier of the target channel. When any parallel frequency conversion path exits operation, the equalizer controller 10 reduces the corresponding power slice extraction amount and redistributes the migration photon flux of the remaining paths. Specific Implementation Example 4: In this embodiment, the sixth high-power DFB laser channel 31 stops outputting. The equalization controller 10 sets the existing photon flux of the sixth frequency node to zero, retains the sixth frequency node as the transferee node, and calculates the surplus photon flux and migration edge capacity of the remaining seven channels.

[0137] The equalization controller 10 selects two migration edges, one from the fifth frequency node to the sixth frequency node and the other from the seventh frequency node to the sixth frequency node, extracts power slices from the fifth and seventh channels respectively, and converts them to the target optical frequency of the sixth channel. The two converted carriers undergo phase correction and are then combined into the input of the sixth data modulator. The sixth high-power DFB laser channel 31 remains closed, and the sixth data modulator receives the target frequency continuous wave carrier generated by other frequency nodes.

[0138] When the excess photon flux of the remaining channels is insufficient to maintain the original common equalization power, the equalization controller 10 reduces the common equalization power and redistributes all migration tasks, so that the inputs of the eight data modulators are in the new allowable power range. Specific Implementation Example 5: In this embodiment, the frequency lattice equalization unit is positioned before each continuous wave carrier enters the independent sweep frequency modulator. Each high-power DFB laser channel 31 outputs a continuous wave carrier with a fixed target light frequency. The equalization controller 10 performs photon migration equalization based on the effective output light power at the input of each sweep frequency modulator. After equalization is completed, each sweep frequency modulator applies a frequency modulation signal.

[0140] When the fixed optical path loss of a transmission channel increases, the effective output optical power at the input of the sweep modulator of that channel decreases. The equalization controller 10 increases the target-side photon flux migrating to that frequency node based on the updated output transmission efficiency and extracts a power slice from the power-excess frequency node. Photon migration occurs before sweep modulation and does not transmit the sweep information supplied to the channel. Specific Implementation Example Six: In this embodiment, the target light frequency of each high-power DFB laser channel 31 corresponds to a different gas absorption line, the frequency lattice equalization unit is set before each carrier enters the independent modulator, and the actual use end 13 is set at the sample cell inlet.

[0142] The equalization controller 10 measures the output transmission efficiency of each channel from the frequency lattice equalization unit inlet to the sample cell inlet under a calibration state with no sample absorption. When a channel experiences an increase in transmission loss due to optical window contamination or a change in fiber coupling position, the equalization controller 10 recalculates the migration photon flux based on the effective output optical power at the sample cell inlet and converts the surplus continuous wave carriers of other frequency nodes to the target optical frequency of that channel. In this embodiment, the effective output optical power at the sample cell inlet is denoted as... The unit is watts. The output transmission efficiency from the equalization unit inlet to the sample cell inlet is denoted separately as... , which are dimensionless quantities; when calculating the equalization power at this monitoring location, replace the aforementioned data modulator input power and corresponding output transmission efficiency with the two quantities mentioned above. The superscript cell only indicates the monitoring location at the sample cell inlet.

[0143] The four-channel verification platform employs a narrow-linewidth reference laser, an electro-optic frequency comb generation device, four high-power DFB lasers, four power slice extraction units, four adjacent frequency shift edges, four phase correction units, and four coherent injection units at the same frequency. After completing parallel verification with single and double shift edges, it is expanded to an eight-channel platform.

[0144] The field-programmable gate array (FPGA) software includes a sampling synchronization module, a digital down-conversion module, an in-phase quadrature calculation module, a phase expansion module, a phase proportional-integral control module, an actuator timing module, and a hardware limiting module. The embedded processor software includes a device registration module, a calibration parameter module, a node state module, a migration edge state module, a common equalization power solution module, a minimum cost photon flow module, a running state machine module, an anomaly diagnosis module, and a running log module.

[0145] The device registration module binds the power slice extraction calibration table, frequency conversion efficiency calibration table, and phase correction calibration table to the device serial number. The operation log records the input node status, output migration scheme, actuator target value, phase locking result, and anomaly handling result for each common equalization power solution. The operation log does not participate in rapid control; it is only used to reproduce the operating state and correct long-term calibration parameters.

[0146] The project acceptance process begins with all migration branches closed, recording the drive current, chip temperature, target light frequency, linewidth, and effective output light power of each high-power DFB laser channel 31. Subsequently, photon migration equalization is initiated, and the same parameters are recorded. Before and after equalization, the DFB drive current should remain within the corresponding injection lock-in current range, and the changes in target light frequency and linewidth should be within the product specifications.

[0147] Channel power equalization acceptance is performed with the data modulator input as the measurement location, continuously collecting effective output optical power for no less than one thousand power distribution cycles, and calculating the maximum deviation and root mean square deviation of each channel relative to the common equalization power. The reduction in source-side photon flux in channels with excess power and the increase in target-side photon flux in channels with insufficient power should meet the calibration value of the total photon transmission efficiency on the migration side.

[0148] Phase stability acceptance involves continuously recording the phase error of each phase correction branch and statistically analyzing the mean, standard deviation, and number of out-of-limit occurrences of the phase error. Frequency conversion acceptance measures the source carrier suppression, target frequency conversion efficiency, pump isolation, and conversion-added noise.

[0149] The abnormal reconstruction acceptance test sequentially simulates one migration edge shutdown, one frequency conversion unit overheating, one phase correction branch lockout, and one high-power DFB laser channel 31 stopping output. The equalization controller 10 should isolate the abnormal object, reconstruct the frequency migration map, and output a new common equalization power and migration scheme within the preset fault handling cycle.

[0150] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising a reference structure" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0151] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for equalizing a high-power DFB array light source, characterized in that, The process includes the following steps: generating coherent reference beams corresponding to multiple target light frequencies, injecting each coherent reference beam into a corresponding high-power DFB laser channel, and placing each high-power DFB laser channel in an injection-locked state corresponding to the target light frequency. The unmodulated continuous wave carriers output from each of the high-power DFB laser channels are input into their respective main optical paths. The effective output optical power and optical field phase of each main optical path at the corresponding data modulator input are collected. Based on the effective output optical power, the common equalization power, the power excess channel, the power deficiency channel, and the source-side photon flux migrating from the power excess channel to the power deficiency channel are determined. A power slice corresponding to the source-side photon flux is extracted from the main optical path of the power excess channel. Based on the frequency difference between the target optical frequency of the power excess channel and the target optical frequency of the power deficiency channel, the power slice is converted into a converted carrier with the same frequency as the target optical frequency of the power deficiency channel. Based on the phase error between the optical field phase of the converted carrier and the optical field phase of the original continuous wave carrier of the power deficiency channel, the optical field phase of the converted carrier is adjusted, and the adjusted converted carrier is incorporated into the main optical path of the power deficiency channel. During the extraction, conversion, and incorporation of the power slice, the driving current of each high-power DFB laser channel is maintained within the corresponding injection lock-in current range.

2. The high-power DFB array light source equalization method according to claim 1, characterized in that, The coherent reference light is generated by a narrow-linewidth reference laser and an optical frequency comb generation structure, and the target light frequency of the i-th coherent reference light satisfies the following... In the formula, This represents the target light frequency of the i-th high-power DFB laser channel. This indicates the reference light frequency corresponding to the narrow linewidth reference laser. This represents the integer frequency slot number corresponding to the i-th high-power DFB laser channel. The basic frequency interval formed by the optical frequency comb generation structure is represented by i, which represents the high-power DFB laser channel number. Each coherent reference light is injected into the corresponding high-power DFB laser channel after wavelength demultiplexing and seed light power adjustment. The injection locking state is determined based on the beat frequency residual and phase error variance between the corresponding high-power DFB laser channel and the coherent reference light.

3. The high-power DFB array light source equalization method according to claim 1, characterized in that, The acquisition of the effective output optical power and optical field phase of each of the above includes: extracting monitoring light from the input terminal of each of the data modulators, inputting the monitoring light of the i-th channel and the corresponding coherent reference light into a 90-degree optical mixer, and obtaining the in-phase detection component through balanced photodetector. and orthogonal detection components According to the in-phase detection component and the orthogonal detection components Construct a complex optical field estimate for the i-th channel, determine the effective output optical power of the i-th channel based on the square of the amplitude of the complex optical field estimate, and determine the optical field phase of the i-th channel based on the complex argument of the complex optical field estimate.

4. The high-power DFB array light source equalization method according to claim 1, characterized in that, Determining the common equalization power and source-side photon flux includes: defining each target optical frequency as a frequency node, defining the source frequency node connected to the frequency conversion unit and the target frequency node as a migration edge, and constructing a frequency migration map from each frequency node and each migration edge; dividing the effective output optical power of each channel's main optical path by the product of the corresponding target optical frequency and Planck's constant to obtain the existing photon flux of each frequency node; and determining the common equalization power and the source-side photon flux corresponding to each migration edge under the constraints that the source-side photon flux of each migration edge does not exceed the corresponding maximum allowable source-side photon flux, the photon flux retained by each power excess channel is not lower than the corresponding minimum retained photon flux, the total frequency conversion pump power does not exceed the pump power upper limit, and the effective output optical power of each data modulator input is within the same allowable power range.

5. The high-power DFB array light source equalization method according to claim 1, characterized in that, Extracting the power slice from the main optical path of the power-excess channel includes: inputting the main optical path of the power-excess channel into an adjustable Mach-Zehnder coupler; determining a target extraction ratio based on the ratio of the sum of source-side photon fluxes migrating from the power-excess channel to each power-deficient channel to the existing photon flux of the power-excess channel; adjusting the adjustable Mach-Zehnder coupler according to a pre-calibrated correspondence between the driving voltage and the extraction ratio; inputting the power slice corresponding to the target extraction ratio into a photon migration bus; and outputting the remaining continuous wave carrier along the main optical path of the power-excess channel; when the maximum allowable source-side photon flux of a single frequency conversion path is less than the source-side photon flux migrating from the power-excess channel to the power-deficient channel, dividing the source-side photon flux into multiple sub-photon fluxes, and inputting the multiple sub-photon fluxes into parallel frequency conversion paths respectively.

6. The high-power DFB array light source equalization method according to claim 1, characterized in that, Converting the power slice into the converted carrier includes: inputting the power slice, the first coherent pump light, and the second coherent pump light into a Bragg scattering four-wave mixing waveguide, wherein the frequency of the first coherent pump light is... The frequency of the second coherent pump light is The target optical frequency of the power excess channel is The target light frequency of the underpowered channel is and satisfy The frequency generated by the Bragg scattering four-wave mixing waveguide is... The converted carrier is then guided to the corresponding output port through a wavelength selective filtering structure.

7. The high-power DFB array light source equalization method according to claim 1, characterized in that, Adjusting the optical field phase of the converted carrier includes: determining the phase error based on the phase of the converted carrier relative to the coherent reference light corresponding to the underpowered channel and the phase of the original continuous wave carrier of the underpowered channel relative to the coherent reference light, and correcting the phase corrector drive amount based on the phase error.

8. The high-power DFB array light source equalization method according to claim 4, characterized in that, Determining the source-side photon flux corresponding to each migration edge includes: using the output photon flux of each frequency node after photon migration satisfying the target photon flux corresponding to that frequency node as node constraints, and using the upper and lower limits of the source-side photon flux of each migration edge, the minimum retained photon flux of each power excess channel, the upper limit of the total pump power, and the locking state of each phase correction branch as boundary constraints, to obtain the source-side photon flux of each migration edge.

9. The high-power DFB array light source equalization method according to claim 1, characterized in that, Before executing the high-power DFB array light source equalization method, the following calibrations are performed sequentially: the output transmission efficiency of each high-power DFB laser channel from the entrance of the frequency lattice equalization unit to the input of the corresponding data modulator is collected; the driving voltage of each adjustable Mach-Zehnder coupler is adjusted step by step, and a calibration relationship between the driving voltage and the extraction ratio is formed based on the main optical path power of the corresponding channel and the input power of the photon migration bus. A continuous wave carrier with known optical power is input to each frequency conversion path, and the power and phase of the first coherent pump light and the second coherent pump light are adjusted to form a calibration relationship between the input optical power, pump power and frequency conversion efficiency. The low-power converted carrier is input to the tunable coherent multiplexer corresponding to the underpowered channel. The phase corrector drive is scanned and the synthesized optical power is recorded to form a calibration relationship between the phase corrector drive and the synthesized optical power.

10. The high-power DFB array light source equalization method according to claim 4, characterized in that, During the execution of the high-power DFB array light source equalization method, the injection lock status of each high-power DFB laser channel, the frequency conversion efficiency of each frequency conversion path, and the temperature of each frequency conversion unit are detected. When any high-power DFB laser channel loses its injection lock, the extraction of power slices from that high-power DFB laser channel is stopped, and the addition of conversion carriers to that high-power DFB laser channel is also stopped. When the frequency conversion efficiency of any frequency conversion path is lower than the corresponding lower limit of conversion efficiency or the temperature of the frequency conversion unit is higher than the corresponding upper limit of temperature, that frequency conversion path is deleted from the frequency migration map, and the source-side photon flux corresponding to each migration edge is re-determined. When any high-power DFB laser channel stops outputting, the existing photon flux of the frequency node corresponding to that high-power DFB laser channel is set to zero, and the common equalization power is re-determined based on the existing photon flux of the remaining high-power DFB laser channels.

Images