A low-cost driving method of SOA bias current for PON link

Abstract

The application discloses a low-cost PON link SOA bias current driving method and belongs to the technical field of optical communication. By introducing a current distribution network composed of an adjustable current mirror, a local energy storage capacitor pre-charging mechanism and a central cooperative controller, simultaneous, independent and stable biasing of a laser and a semiconductor optical amplifier during PON uplink burst transmission is realized. The central cooperative controller configures the current mirror ratio in advance and pre-charges the energy storage capacitor according to the uplink burst enable signal before the burst starts, maintains simultaneous power supply during the burst, and cuts off the power supply after the burst to reduce power consumption. Compared with the prior art, the application completely avoids the use of an external independent current source chip, significantly reduces the bill of materials cost and system complexity, while ensuring the burst response speed, biasing accuracy and nonlinear distortion suppression effect, and can be widely applied to a passive optical network system with high link budget.

Description

Technical Field

[0001] This invention belongs to the field of optical communication technology, specifically a low-cost method for driving SOA bias current in a PON link. Background Technology

[0002] Passive optical networks (PONs), as the mainstream technology for fiber optic access, are facing increasing link budget requirements. Introducing semiconductor optical amplifiers (SOAs) can effectively compensate for optical power attenuation caused by increased splitting ratios or extended transmission distances, but the stable operation of SOAs depends on precise bias current driving. In existing PON products, electro-absorption modulated laser (EML) driver chips typically integrate only one constant current source for laser bias, unable to provide independent bias currents for both the laser and the SOA simultaneously. Traditional solutions involve adding an independent current source chip (such as an IDAC) or using a multi-current source integrated chip, but this significantly increases bill of materials costs, board area, and control complexity, contradicting the pursuit of extreme cost-effectiveness in PON systems.

[0003] At a deeper level, the core contradiction of a single constant current source driving dual loads lies in the different bias current requirements of the laser and the SOA (the laser needs a stable threshold current, while the SOA needs to dynamically adjust according to link losses), and both need to operate simultaneously in PON uplink burst mode. If simply connected in parallel, load coupling will lead to problems such as power supply voltage fluctuations and gain crosstalk; if time-sharing power is used, it cannot meet the requirement of simultaneous operation during bursts.

[0004] Therefore, there is an urgent need for a low-cost driving method based on a single constant current source that can simultaneously provide independent and controllable bias current for the laser and SOA. Summary of the Invention

[0005] The purpose of this invention is to provide a low-cost method for driving the SOA bias current of a PON link, in order to solve the problem in the prior art that the laser and semiconductor optical amplifier cannot be driven independently and dynamically in a coordinated manner under low-cost constraints, resulting in a difficulty in balancing functional implementation and system cost.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0007] A low-cost method for driving the SOA bias current of a PON link, applied to a drive system containing a single constant current source, includes the following steps:

[0008] A current distribution network is provided, the input of which is connected to the output of a single constant current source. The current distribution network has at least two outputs, which are respectively used to connect a laser as a first load and a semiconductor optical amplifier as a second load.

[0009] The central coordination controller acquires status signals that characterize uplink burst transmissions of the PON link in real time.

[0010] According to the status signal, the central coordinating controller dynamically controls the current distribution network during the burst transmission period, and simultaneously distributes the total current output of the single constant current source to the first load and the second load. The current distributed to the first load meets the bias requirements of the laser, and the current distributed to the second load meets the bias requirements of the semiconductor optical amplifier.

[0011] During non-burst transmission periods, the central coordinating controller controls the current distribution network to cut off or reduce power supply to the first and second loads.

[0012] According to the above scheme, the current distribution network is specifically composed of a first adjustable current mirror and a second adjustable current mirror connected in parallel. The input terminals of the first adjustable current mirror and the second adjustable current mirror are connected to the output terminal of a single constant current source, and their output terminals are connected to the first load and the second load, respectively. The central coordinating controller realizes the proportional distribution of the total current between the two loads by adjusting the mirror ratio of the first adjustable current mirror and the second adjustable current mirror.

[0013] According to the above scheme, the central coordinating controller is based on the target bias current of the laser. and the target bias current of the semiconductor optical amplifier Set the mirror ratio of the first adjustable current mirror. The mirror ratio of the second adjustable current mirror The specific relationships are as follows:

[0014]

[0015] in, The total current output by a single constant current source, and satisfying the following conditions: At the same time ensure This ensures that the total current is fully distributed between the two loads.

[0016] According to the above scheme, the current distribution network is also equipped with a first local energy storage capacitor and a second local energy storage capacitor. The first local energy storage capacitor is connected in parallel across the first load, and the second local energy storage capacitor is connected in parallel across the second load. The local energy storage capacitor is used to precharge the load by a single constant current source during non-burst transmission periods, and to supply power to the corresponding load together with the single constant current source at the moment of burst transmission, thereby compensating for the transient current demand caused by load switching and suppressing voltage drop.

[0017] According to the above scheme, the central coordination controller will adjust the timing based on a preset lead time. Before the effective edge of the burst transmission enable signal arrives, the control current distribution network switches the single constant current source to the ready-to-power state and pre-charges the local energy storage capacitor; the lead time must simultaneously meet the following conditions:

[0018]

[0019] in, The time constant for the load current is determined by both the carrier lifetime of the semiconductor optical amplifier and the circuit time constant. The minimum protection interval specified by the PON link standard ensures that current switching does not occupy effective transmission time slots.

[0020] According to the above scheme, it also includes a cross-interference suppression step: real-time monitoring of voltage fluctuation at the output of a single constant current source; when the detected voltage fluctuation exceeds the preset threshold Vth, the central coordinating controller adjusts the control timing or current distribution ratio of the current distribution network accordingly in subsequent burst transmission cycles to smooth the current change process and suppress power supply voltage oscillation caused by load switching.

[0021] According to the above scheme, when the drive system needs to drive multiple loads simultaneously, and the number of loads is N and N≥2, the central coordination controller maintains a load priority list and determines the load priority based on the service priority label. and link loss Calculate the urgency factor for each load. The calculation formula is as follows:

[0022]

[0023] in, and The weighting coefficients are preset; the central coordination controller is set according to... The current is allocated in descending order of value, if the total current demand of all loads is... Exceeding the maximum output current of a single constant current source If the current demand is not met, the load with the highest urgency factor will be prioritized until the current resources are exhausted. The unmet load will be suspended during the current burst cycle, and its current demand will be recorded and prioritized for compensation in the next cycle.

[0024] According to the above scheme, the central coordinating controller contains a load state analysis engine. This engine maintains a dynamic model of the bias current of the semiconductor optical amplifier, which is used to analyze the total link loss. and target output optical power The required target gain is calculated, and then the corresponding target bias current is determined. Specifically, the target gain is first calculated based on the total link loss and the target output optical power. :

[0025]

[0026] in, The output optical power of the laser is obtained through the laser bias-optical power characteristic curve; then, the corresponding target bias current is obtained through an offline calibrated three-dimensional lookup table. ,in For ambient temperature, this lookup table is stored in the controller's non-volatile memory.

[0027] According to the above scheme, it also includes a machine learning optimization step: first, construct a feature vector. It includes the following real-time monitoring data:

[0028]

[0029] in, For ambient temperature, This is the laser bias current. To output optical power for semiconductor optical amplifiers, For link loss, The bias current of the semiconductor optical amplifier in the previous burst cycle is used; an online learning model is employed to predict the bias current in the current cycle. :

[0030]

[0031] in, The model parameters are updated online using stochastic gradient descent; the predicted values ​​are compared with the baseline values ​​calculated by the load state analysis engine. Weighted fusion is performed to obtain the final bias current:

[0032]

[0033] Among them, weight The system adaptively adjusts based on the prediction error of the model on the most recent samples to ensure that the system can smoothly transition to the optimal operating point when the environment changes.

[0034] According to the above scheme, a digital predistortion compensation step is also included: first, a nonlinear model of the semiconductor optical amplifier is established to describe its input optical power. With output optical power The relationship between them:

[0035]

[0036] in, This is the bias current for the semiconductor optical amplifier. This is the small-signal gain coefficient. This is the reciprocal of the gain saturation coefficient; then, based on the desired linear gain... Construct a predistortion function to preprocess the modulated signal, and then measure the amplitude of the predistorted signal. It is given by the following formula:

[0037]

[0038] During burst transmission, the baseband signal to be modulated is first processed by a predistortion function to generate a predistorted amplitude sequence, which then drives the laser to generate an optical signal. After being amplified by a semiconductor optical amplifier, the signal is output, thereby canceling the nonlinear distortion of the semiconductor optical amplifier itself and improving the signal transmission quality.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] This invention, under the hardware constraint of a single constant current source, achieves simultaneous, independent, and stable biasing of the laser and SOA during PON uplink bursts by introducing mechanisms such as an adjustable current mirror distribution network, local energy storage capacitor pre-charging, cross-interference suppression, and a load dynamic model. This method completely avoids the use of external independent current source chips, significantly reducing system cost and complexity. Simultaneously, through precise timing control and closed-loop feedback, it ensures burst setup speed and long-term stability, providing a feasible solution for low-cost integration of SOA in high link-budget PON systems. Attached Figure Description

[0041] Figure 1 This is a schematic diagram of the overall system architecture of a low-cost PON link SOA bias current driving method of the present invention.

[0042] Figure 2 This is a flowchart illustrating the driving method for driving the SOA bias current of a low-cost PON link according to the present invention. Detailed Implementation

[0043] 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 skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] Example 1

[0045] This embodiment provides a method for co-driving a laser and SOA based on a single constant current source, and its system architecture is as follows: Figure 1 As shown, the system specifically includes a single constant current source (maximum output current is...) integrated within the EML driver chip, featuring low noise and fast startup characteristics. A current distribution network consisting of a first adjustable current mirror and a second adjustable current mirror connected in parallel (the input terminals of the two current mirrors are connected to the output terminal of a single constant current source, and the output terminals are respectively connected to the laser and the SOA to achieve precise current distribution), a central collaborative controller with an embedded load state analysis engine, timing collaborative control logic, cross-interference suppression module, and standardized communication interface, is connected in parallel across the laser. With parallel connection at both ends of SOA The system consists of a local energy storage capacitor array (used to compensate for transient current demands), an optical power detection unit containing a high-response-speed photodiode and a low-noise transimpedance amplifier, and a high-precision temperature sensor (used to acquire link status data in real time).

[0046] like Figure 2 As shown, the central coordinating controller acquires in real time the uplink burst enable signal TX_EN and burst length information issued by the PONMAC layer through the communication interface, as well as the SOA output optical power data fed back by the optical power detection unit and the ambient temperature data collected by the temperature sensor. The load status analysis engine is based on the link loss estimate. and target output optical power Calculate the required bias current for the laser and SOA separately: where the laser bias current is... The SOA bias current is precisely determined based on its own threshold current characteristics and temperature compensation coefficient to ensure stable laser lasing; This is obtained through a two-stage calculation model. First, the required gain is calculated based on the ratio of the target output optical power to the laser output optical power, as shown in the following formula:

[0047]

[0048] in, The value was obtained by looking up the laser bias-optical power characteristic curve in the table.

[0049] Then, based on the SOA gain-current-temperature three-dimensional lookup table, which is offline calibrated and optimized by piecewise linear interpolation (interpolation error less than 0.1dB), the current temperature is used to... and desired gain Find the corresponding value The lookup table is pre-stored in the controller's non-volatile memory and supports online updates.

[0050] During normal operation of PONONU, the system is divided into two core operating states: standby and burst transmission. When the uplink burst enable signal TX_EN is invalid, the system is in standby mode. The central coordinating controller configures both adjustable current mirrors to be off, and simultaneously controls the single constant current source to output a microamp-level standby current or directly shuts it off to reduce power consumption. At the same time, an auxiliary pre-charging circuit consisting of a high-resistance current-limiting resistor and a unidirectional conduction diode charges the local energy storage capacitor. and A slow pre-charge is performed to raise the capacitor voltage to near the rated operating voltage of the corresponding load (deviation less than 10%), and the pre-charge current is strictly controlled within 5μA to avoid affecting standby power consumption; when the uplink burst enable signal TX_EN is detected to be about to become effective, the system initiates the burst transmission state preparation procedure, and the central coordinating controller determines the timing based on the preset lead time. The lead time for initiating current distribution preparation must meet certain conditions. and The time constant for the load current is determined by both the SOA carrier lifetime and the circuit's RC time constant. This is the minimum protection interval specified by the PON standard. The measured value in this embodiment is... It is approximately 0.8 μs, therefore it is set to... This ensures stable current establishment without occupying effective transmission time slots.

[0051] Lead time At that moment, the controller starts the single-channel constant current source and sets the total output current. Simultaneously, the first adjustable current mirror image ratio is configured. Second adjustable current mirror mirror ratio Strictly guarantee To meet the current conservation constraint, this adjustable current mirror is implemented using a binary weighted transistor array based on CMOS technology. Each current mirror is independently adjusted by an 8-bit control word, achieving a current distribution step accuracy of up to [percentage missing]. Since the local energy storage capacitor has been pre-charged to near the operating voltage during the standby phase, the capacitor and the constant current source form a cooperative power supply mechanism during the current switching instant, effectively compensating for the transient current demand of the load, suppressing the voltage drop at the load end to within 35mV, and the current settling time to less than 1μs. Therefore, when the effective edge of TX_EN arrives, the laser and SOA are already in a stable bias state, and can immediately respond to the modulation signal to realize instantaneous transmission and amplification of the optical signal.

[0052] Throughout the burst transmission (e.g., a 512-byte burst frame), the central coordinating controller maintains the current current distribution ratio to ensure a stable load operating point. If the burst frame length is long or the link loss changes dynamically, the controller will output optical power based on the real-time feedback from the optical power detection unit. Fine-tuning through a closed-loop feedback mechanism ,when At this time, the proportional-integral (PI) algorithm is activated for current correction, and the correction formula is as follows:

[0053]

[0054] The error signal , , The adjustment coefficients are optimized through experiments to ensure a fast correction process without overshoot. Once the burst transmission is complete, the TX_EN signal is removed, the controller immediately shuts down the two adjustable current mirrors, switches the single constant current source to standby mode, and restarts the local energy storage capacitor pre-charging process to prepare for the next burst transmission.

[0055] In addition, the cross-interference suppression module built into the central coordinating controller monitors the voltage fluctuation at the constant current source output in real time. When the detected voltage fluctuation exceeds a preset threshold, the module will detect the voltage fluctuation. When this happens, the module automatically records the disturbance event and adopts an adaptive suppression strategy in the next burst cycle by extending or shortening the lead time. (Step length) Alternatively, a multi-step gradual adjustment method can be used to adjust the rate of change of the current mirror ratio, smoothing the current switching process to suppress circuit oscillation. According to actual measurements, this mechanism can reduce the peak-to-peak voltage fluctuation from 180mV in the traditional solution to below 35mV, significantly improving system stability.

[0056] Example 2

[0057] Building upon the aforementioned Embodiment 1, this embodiment provides a dynamic priority scheduling method for multiple loads in multi-load application scenarios (such as multi-wavelength ONUs and multi-channel OLT optical modules) to achieve optimal allocation of limited current resources. In this embodiment, the current distribution network is expanded to N output terminals, each corresponding to one of N loads (laser-SOA components). Each load is configured with an independent adjustable current mirror, and its mirror ratio can be independently programmed and adjusted. The central coordinating controller maintains a dynamically updated load priority list based on real-time scheduling information from the PON system. This scheduling information includes service priority tags provided by the MAC layer according to ITU-T standards. (Values ​​range from 1 to 5, with higher values ​​indicating higher priority), and real-time loss values ​​for the links corresponding to each load. (Unit: dB) The controller calculates based on the above information using the following formula:

[0058]

[0059] Among them, the urgency factor of each load (where (These are dynamically configurable positive real-valued weighting coefficients), the priority list is sorted by urgency factor. The loads are sorted in descending order to ensure that high-priority loads with high link loss receive priority power supply.

[0060] Before the start of each burst cycle, the controller first uses the load state analysis engine to calculate the target bias current for each load. And sum them up to get the total required current. ,like (Maximum output current of the constant current source) is then directly allocated according to the target current of each load, and the ratio of each current mirror is set to... To ensure full power supply; if Then, it enters the priority allocation mode, using a greedy algorithm to allocate current sequentially according to the priority list: allocating all the required current to the first load. Residual current Then, the second load was allocated. The remaining current is updated, and this process continues until the current resources are exhausted. Loads that do not receive sufficient current will suspend operation during the current cycle, and their current demand will be recorded and prioritized for compensation allocation in the next burst cycle. The core objective of this allocation strategy is to maximize the system's weighted current allocation efficiency. Within a limited current capacity, the controller prioritizes meeting the operational needs of high-urgent loads, effectively ensuring the transmission quality of critical services such as management frames and real-time voice in resource-constrained scenarios. After allocation, the controller determines the appropriate current based on the allocated current for each load. Set the mirror ratio of the corresponding current mirror, and coordinate the power supply start-up time of all loads through the timing synchronization control module to ensure precise alignment with the sudden enable signal and eliminate time slot waste.

[0061] Example 3

[0062] To further improve the prediction accuracy of SOA bias current and adapt to device aging and environmental abrupt changes, this embodiment introduces a machine learning-based adaptive current allocation optimization mechanism based on Embodiment 1. This mechanism continuously optimizes the bias current configuration through online learning. The central coordinating controller embeds a lightweight online machine learning engine, which constructs an input feature vector from multi-dimensional real-time monitoring data, as shown below:

[0063]

[0064] in For real-time temperature, The current bias current of the laser, To output optical power for SOA, For real-time link loss value, (The SOA bias current of the previous burst cycle) is used as the output target, and the model parameters are continuously updated through online learning.

[0065] After each burst cycle ends, the controller records the current feature vector. And calculate the error between the actual output optical power and the target value. If the absolute value of the error Then the sample Stored in the training buffer (where To ensure the validity of the sample data, the ideal bias current is obtained by inverse solution through the SOA gain-current-temperature calibration model.

[0066] To balance prediction accuracy and computational complexity, the machine learning engine employs a linear regression model, as shown in the following equation:

[0067]

[0068] in The model parameters are defined, and the stochastic gradient descent (SGD) method is used to update the parameters online. The update formula is as follows:

[0069]

[0070] Learning rate The training process is executed in a low-priority thread in the background of the controller, and the training time for a single session is less than 100ns, which does not affect the real-time control process.

[0071] In the next burst cycle, the controller will use the current feature vector The SOA bias current prediction is obtained by inputting the trained model. A weighted fusion strategy is used to combine it with the baseline value calculated by the load status analysis engine. The final bias current obtained by fusion is calculated using the following formula:

[0072]

[0073] Among them, weight The model is adaptively adjusted based on its recent prediction accuracy, using the following formula:

[0074]

[0075] in, For the model in recent Root mean square prediction error on each sample (The root mean square error of the baseline model), when the model's prediction accuracy improves Approaching 1, when sudden environmental changes cause the model to fail. The system automatically reduces its output power, reverting to the baseline model to ensure operational stability. Simultaneously, the controller continuously monitors the root mean square value of the SOA output optical power error; if this error persists... The error per cycle is less than If the current model parameters are fixed, online learning will be paused to reduce power consumption. If the error continues to increase or the temperature change rate exceeds [a certain threshold], [further action will be taken]. If the learning process is reactivated, the model parameters will be updated quickly. Experimental results show that after adaptive optimization using machine learning in this embodiment, the SOA output optical power fluctuation is further reduced from ±0.5dB to within ±0.2dB, the gain attenuation caused by device aging can be compensated adaptively by the model, and the module calibration-free period is extended to more than 3 years.

[0076] Example 4

[0077] To suppress nonlinear distortion of the SOA in large-signal modulation scenarios and improve signal transmission quality, this embodiment, based on Embodiment 1, incorporates digital predistortion (DPD) technology to construct a driving scheme that coordinates the optimization of bias current and signal processing. The system adds a digital predistortion module to the modulation path. This module works in real-time with the central coordinating controller, adjusting the bias current based on the current SOA bias current. The modulation signal is preprocessed by nonlinear inverse transformation of the input signal amplitude to counteract the inherent gain saturation nonlinearity of SOA.

[0078] During the offline calibration phase, the SOA was experimentally measured under different bias currents. The input-output optical power relationship is derived based on the carrier rate equation and fitted using an exponential model, as shown in the following equation:

[0079]

[0080] in This is the small-signal gain coefficient. (The reciprocal of the gain saturation coefficient) was obtained by fitting multiple sets of experimental data. (Fitting error less than 0.2dB) and The polynomial expression (fitting error less than 5%) is pre-stored in the digital predistortion module. The desired linear gain is set. Then, the predistortion mapping relationship is derived in reverse based on the SOA nonlinear model:

[0081]

[0082] To reduce real-time computational complexity, this function uses piecewise linear interpolation to store data in a lookup table (interpolation error less than 0.1dB), and also supports calculations based on the current... Parameters are calculated in real time.

[0083] During each burst cycle, the central coordination controller determines the target gain based on link loss. The SOA bias current is calculated by the load state analysis engine. And transmit it to the digital predistortion module in real time, the module according to the current Update the predistortion function and The parameters are used to predistort the baseband signal to be modulated (such as PAM4 or NRZ format) to generate an amplitude sequence. This sequence is then converted from digital to analog to drive the laser to generate an optical signal. After being input into the SOA, its nonlinear distortion is canceled by the predistortion process, achieving linear amplification and output. The optical power detection unit monitors the SOA output signal in real time and extracts the third-order intermodulation distortion (IMD3) and error vector amplitude (EVM) as distortion indicators. If IMD3 exceeds -35dBc or EVM exceeds 15%, the feedback correction process is initiated to fine-tune the desired linear gain in the predistortion function. Alternatively, the SOA model parameters can be updated until the distortion index meets the requirements. Feedback correction can be triggered at a fixed period or by temperature changes. To avoid storing massive lookup tables, a Taylor series expansion is used to approximate the predistortion function (taking the first four terms), with coefficients varying accordingly. Dynamic updates save 70% of storage resources while ensuring distortion suppression. Actual test data shows that after adopting the digital predistortion collaborative control of this embodiment, the IMD3 of the SOA output signal is reduced by more than 15dB, the EVM is improved by about 30%, the system can stably support high-order modulation formats such as PAM4-512Gbps, and the transmission distance is extended by about 20% under the same bit error rate requirements.

[0084] In summary, the present invention, through the technical solutions described in Embodiments 1 to 4, constructs a complete, flexible, and low-cost SOA bias current driving system for PON links. Those skilled in the art can adapt the above embodiments to any combination according to actual application scenarios, or supplement additional optimization mechanisms without departing from the core concept of the present invention. All such modifications and extensions should fall within the protection scope of the present invention.

[0085] This invention addresses the technical pain points of traditional PON systems, such as high cost, high complexity, slow burst response, low bias accuracy, and severe signal distortion in laser and SOA driving schemes. By innovatively designing a single-channel constant current source + adjustable current mirror architecture, and combining core technologies such as local energy storage capacitor pre-charging, cross-interference suppression, dynamic priority scheduling, machine learning adaptive optimization, and digital predistortion collaborative control, it achieves synchronous, independent, and high-precision stable biasing of the laser and SOA during PON uplink burst transmission, achieving significant technical results.

[0086] Firstly, in terms of cost and complexity optimization, compared to the traditional dual-path independent constant current source drive scheme, this invention significantly simplifies the hardware circuit architecture through single-path constant current source multiplexing and integrated design, reducing the bill of materials (BOM) cost by approximately 30%, shrinking the chip area by more than 25%, and simultaneously reducing the number of peripheral components, thereby improving system integration and reliability. Secondly, regarding burst response speed, through local energy storage capacitor pre-charging and advance timing control mechanisms, the SOA gain settling time is shortened to less than 1μs, meeting the PON standard's requirement for minimum protection interval, achieving instantaneous start-up and stable transmission of burst signals, and avoiding time slot waste caused by current settling delay in traditional schemes. Thirdly, regarding bias accuracy and stability, by integrating offline calibration lookup tables, PI closed-loop feedback, and machine learning adaptive optimization technology, the SOA output optical power fluctuation is strictly controlled within ±0.2d. Within the range of B, it can adapt to changes in ambient temperature (-40℃~85℃) and device aging effects, extending the module's calibration-free cycle to more than 3 years and reducing maintenance costs. Fourth, in terms of signal transmission quality, the cross-interference suppression module suppresses voltage fluctuations caused by current switching to below 35mV. Combined with digital predistortion collaborative control, SOA third-order intermodulation distortion (IMD3) is reduced by more than 15dB, and error vector amplitude (EVM) is improved by 30%, effectively supporting high-order modulation formats. Under the same bit error rate, the transmission distance is extended by 20%, significantly improving the link budget and transmission capacity of the PON system. Fifth, in terms of scenario adaptability, through a multi-load dynamic priority scheduling mechanism, it can flexibly adapt to complex application scenarios such as multi-wavelength ONUs and multi-channel OLT optical modules, prioritizing the transmission quality of critical services under limited current resources, and improving system robustness and scalability.

[0087] In summary, the technical solution of this invention fully meets the stringent requirements of high link budget PON systems for low cost, fast response, high precision, and low distortion, and has broad application prospects and industrialization value.

[0088] It should be noted that, in this document, relational terms such as "first" and "second" are used only 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 process, method, article, or apparatus.

[0089] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A low-cost method for driving the SOA bias current of a PON link, applied to a driving system, the system comprising a single constant current source, characterized in that: Includes the following steps: A current distribution network is provided, the input of which is connected to the output of a single constant current source. The current distribution network has at least two outputs, which are respectively used to connect a laser as a first load and a semiconductor optical amplifier as a second load. The central coordination controller acquires status signals that characterize uplink burst transmissions of the PON link in real time. According to the status signal, the central coordinating controller dynamically controls the current distribution network during the burst transmission period, and simultaneously distributes the total current output of the single constant current source to the first load and the second load. The current distributed to the first load meets the bias requirements of the laser, and the current distributed to the second load meets the bias requirements of the semiconductor optical amplifier. The central co-controller is based on the target bias current of the laser. and the target bias current of the semiconductor optical amplifier Set the mirror ratio of the first adjustable current mirror. The mirror ratio of the second adjustable current mirror The specific relationships are as follows: , in, The total current output by a single constant current source, and satisfying the following conditions: At the same time ensure To ensure that the total current is fully distributed between the two loads; The central coordination controller operates based on a preset lead time. Before the effective edge of the burst transmission enable signal arrives, the control current distribution network switches the single constant current source to the ready-to-power state and pre-charges the local energy storage capacitor; lead time The following conditions must be met simultaneously: , in, The time constant for the load current is determined by both the carrier lifetime of the semiconductor optical amplifier and the circuit time constant. The minimum protection interval specified by the PON link standard ensures that current switching does not occupy effective transmission time slots; During non-burst transmission periods, the central coordinating controller controls the current distribution network to cut off or reduce power supply to the first and second loads.

2. The method for driving the SOA bias current of a low-cost PON link according to claim 1, characterized in that: The current distribution network is specifically composed of a first adjustable current mirror and a second adjustable current mirror connected in parallel. The input terminals of the first adjustable current mirror and the second adjustable current mirror are connected to the output terminal of a single constant current source, and their output terminals are connected to the first load and the second load, respectively. The central coordinating controller adjusts the mirror ratio of the first adjustable current mirror and the second adjustable current mirror to achieve proportional distribution of the total current between the two loads.

3. The method for driving the SOA bias current of a low-cost PON link according to claim 1, characterized in that: The current distribution network is also equipped with a first local energy storage capacitor and a second local energy storage capacitor. The first local energy storage capacitor is connected in parallel across the first load, and the second local energy storage capacitor is connected in parallel across the second load. The local energy storage capacitor is used to precharge the load by a single constant current source during non-burst transmission periods, and to supply power to the corresponding load together with the single constant current source at the start of burst transmission, thereby compensating for the transient current demand caused by load switching and suppressing voltage drop.

4. The method for driving the SOA bias current of a low-cost PON link according to claim 1, characterized in that: It also includes a cross-interference suppression step: real-time monitoring of voltage fluctuations at the output of a single constant current source; when the detected voltage fluctuation exceeds the preset threshold Vth, the central coordinating controller adjusts the control timing or current distribution ratio of the current distribution network accordingly in subsequent burst transmission cycles to smooth the current change process and suppress power supply voltage oscillations caused by load switching.

5. The method for driving the SOA bias current of a low-cost PON link according to claim 1, characterized in that: When the drive system needs to drive multiple loads simultaneously, and the number of loads is N and N≥2, the central coordination controller maintains a load priority list and determines the load priority based on the service priority label. and link loss Calculate the urgency factor for each load. The calculation formula is as follows: in, and The weighting coefficients are preset; the central coordination controller is set according to... The current is allocated in descending order of value, if the total current demand of all loads is... Exceeding the maximum output current of a single constant current source If the current demand is not met, the load with the highest urgency factor will be prioritized until the current resources are exhausted. The unmet load will be suspended during the current burst cycle, and its current demand will be recorded and prioritized for compensation in the next cycle.

6. The method for driving the SOA bias current of a low-cost PON link according to claim 1, characterized in that: The central coordinating controller contains a load state analysis engine that maintains a dynamic model of the bias current of the semiconductor optical amplifier. This model is used to analyze the total link loss and the target output optical power. The required target gain is calculated, and then the corresponding target bias current is determined. Specifically, the target gain is first calculated based on the total link loss and the target output optical power. : in, The output optical power of the laser is obtained through the laser bias-optical power characteristic curve; then, the corresponding target bias current is obtained through an offline calibrated three-dimensional lookup table. ,in For ambient temperature, this lookup table is stored in the controller's non-volatile memory.

7. The method for driving the SOA bias current of a low-cost PON link according to claim 6, characterized in that: It also includes machine learning optimization steps: first, constructing a feature vector. It includes the following real-time monitoring data: in, For ambient temperature, This is the laser bias current. To output optical power for semiconductor optical amplifiers, For link loss, The bias current of the semiconductor optical amplifier in the previous burst cycle is used; an online learning model is employed to predict the bias current in the current cycle. : in, The model parameters are updated online using stochastic gradient descent; the predicted values ​​are compared with the baseline values ​​calculated by the load state analysis engine. Weighted fusion is performed to obtain the final bias current: Among them, weight The system adaptively adjusts based on the prediction error of the model on the most recent samples to ensure that the system can smoothly transition to the optimal operating point when the environment changes.

8. The method for driving the SOA bias current of a low-cost PON link according to claim 1, characterized in that: It also includes a digital predistortion compensation step: first, a nonlinear model of the semiconductor optical amplifier is established to describe its input optical power. With output optical power The relationship between them: in, This is the bias current for the semiconductor optical amplifier. For small signal gain coefficient, This is the reciprocal of the gain saturation coefficient; then, based on the desired linear gain... Construct a predistortion function to preprocess the modulated signal, and then measure the amplitude of the predistorted signal. It is given by the following formula: During burst transmission, the baseband signal to be modulated is first processed by a predistortion function to generate a predistorted amplitude sequence, which then drives the laser to generate an optical signal. After being amplified by a semiconductor optical amplifier, the signal is output, thereby canceling the nonlinear distortion of the semiconductor optical amplifier itself and improving the signal transmission quality.

Images