An optical chip-based voltage source matrix control method

Abstract

The application discloses a voltage source matrix control method based on an optical chip, and comprises the following steps: S1, using an ESP32-S3 single-chip microcomputer as a core controller, responsible for data communication with eight DAC8568 chips through an SPI interface, the DAC8568 chip comprising eight independent 16-bit precision voltage output channels; S2, realizing channel selection through a 74HC4051 data selector, the ESP32-S3 single-chip microcomputer controlling the selection signal of the 74HC4051 through a GPIO, dynamically selecting a specific DAC8568 chip and a voltage output channel; S3, adjusting the voltage output value in real time through the ESP32-S3 single-chip microcomputer, sending a 16-bit digital signal to the selected DAC8568 voltage output channel through the SPI interface, and converting the digital signal into a corresponding analog voltage by the DAC8568; S4, dynamically adjusting the voltage output value of each voltage output channel based on a multi-agent deep deterministic policy gradient algorithm; and S5, integrating all control and communication circuits on a PCB board. The application realizes high-precision voltage matrix control of the optical chip by using the ESP32-S3 and the multi-agent algorithm.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of high-precision voltage source matrix, and particularly relates to a voltage source matrix control method based on an optical chip. BACKGROUND

[0002] In modern optoelectronic applications, the performance of optical chips is highly dependent on the precision and stability of the supply voltage. With the advancement of optical communication and optical measurement technology, higher requirements are placed on the voltage source of optical chips. Optical chips require high-precision, stable voltage sources to ensure that they can accurately perform the required functions during operation. For example, optical sensors and optical emitters often require voltage sources to provide very small voltage fluctuations to ensure the accuracy of signals and the reliability of systems.

[0003] Existing voltage source systems are mostly used for general power supply applications, and their precision and power output range cannot meet the needs of high-precision optical chips. Traditional voltage source systems have certain limitations in output precision, power driving capability, and system integration, especially in applications that require independent control of multiple channels.

[0004] Currently, there are some technical solutions that attempt to address the demand for high-precision voltage sources. For example, using DACs (Digital-to-Analog Converters) to provide precise voltage output is a common method. Among them, the DAC8568 chip, as a 16-bit precision DAC, can provide high-precision voltage output. Traditional solutions usually use a single DAC to drive one or more output channels, but when higher power or multiple channel control is required, complex circuit design and additional components are often needed to achieve the combination or switching of multiple DACs. For example, some existing solutions use multiple DAC8568 chips and use manual switching or simple selection circuits to achieve control of multiple channels. However, these solutions may have problems such as complex control, large system size, and insufficient power output in actual applications, which are particularly evident in high-density integration and high-power applications.

[0005] The implementation method of the prior art includes the use of the DAC8568 chip. The DAC8568 is a 16-bit high-precision DAC that can provide precise voltage output. It is usually used in single-channel or a small number of channel voltage control. For applications requiring multiple channel control, multiple DAC8568 chips and complex selection circuits are usually required. Use SPI interface to control DAC: SPI (Serial Peripheral Interface) is a commonly used communication protocol for controlling DACs. Through the SPI interface, precise control of the DAC can be achieved, but in the case of multiple DACs, additional selection circuits are usually required to switch different DACs.

[0006] The high-precision voltage source system in the prior art can provide a certain degree of voltage output precision, but still has significant defects in many aspects:

[0007] The first is the power output limitation. Many traditional voltage source systems have difficulty in providing sufficient output capacity in high-power applications. This may cause unstable output voltage and affect the performance of the optical chip when driving high-power optical chips.

[0008] The second is the system complexity and volume. In the existing solutions, using multiple DAC chips for multi-channel control often requires complex selection circuits, which not only increases the complexity of the system, but also makes the overall design volume larger. For compact integrated application scenarios, such design is not suitable.

[0009] The third is the lack of control flexibility. The channel selection and control mechanism in traditional voltage source systems is relatively simple, and usually cannot achieve flexible channel switching and independent control. This limits the applicability of the system in multi-channel, high-precision applications.

[0010] The fourth is the low integration. The existing technology usually requires multiple discrete components when implementing high-precision, multi-channel output, which results in low system integration, making it difficult to meet the needs of modern optical chips for compact design and high integration.

[0011] Therefore, how to provide a voltage source matrix control method based on an optical chip is a problem that those skilled in the art need to solve. SUMMARY

[0012] An object of the present application is to provide a voltage source matrix control method based on an optical chip. The present application uses an ESP32-S3 single-chip microcomputer as the core controller, combines multiple DAC8568 chips and a 74HC4051 data selector, and realizes accurate dynamic regulation and control of the voltage output of the optical chip through a multi-agent deep deterministic policy gradient algorithm. The system communicates at high speed through the SPI interface and realizes real-time voltage regulation based on external feedback, ensuring the voltage stability and precision of the optical chip in complex working environments. The entire system is integrated on a PCB board, has the advantages of high integration, flexible regulation and control, and stable operation, and is particularly suitable for high-power, multi-channel optical chip application scenarios.

[0013] According to the voltage source matrix control method based on an optical chip according to an embodiment of the present application, the following steps are included:

[0014] S1, an ESP32-S3 single-chip microcomputer is used as a core controller, responsible for data communication with eight DAC8568 chips through an SPI interface, the DAC8568 chip includes eight independent 16-bit precision voltage output channels, a total of 64 voltage output channels;

[0015] S2, channel selection is realized by a 74HC4051 data selector, and the ESP32-S3 single-chip microcomputer controls the selection signal of the 74HC4051 data selector through a GPIO to dynamically select a specific DAC8568 chip and a voltage output channel;

[0016] S3, the voltage output value of each voltage output channel is adjusted in real time by the ESP32-S3 single-chip microcomputer, 16-bit digital signals are sent to the selected DAC8568 voltage output channel through an SPI interface, the DAC8568 converts the digital signals into corresponding analog voltages, and outputs through the corresponding voltage output channel;

[0017] S4, based on a multi-agent deep deterministic policy gradient algorithm, real-time analysis is performed according to external environment feedback and the running state of the optical chip, and the voltage output value of each voltage output channel is dynamically adjusted;

[0018] S5, all control and communication circuits are integrated on a PCB board.

[0019] Optionally, the ESP32-S3 single-chip microcomputer communicates data with eight DAC8568 chips through a full-duplex SPI interface, and 16-bit digital signals are transmitted each time.

[0020] Optionally, the ESP32-S3 single-chip microcomputer controls the 3-bit selection signal of the 74HC4051 data selector through a GPIO to realize dynamic selection and switching of the eight DAC8568 chips and each voltage output channel.

[0021] Optionally, the ESP32-S3 single-chip microcomputer sends 16-bit digital signals to the voltage output channel of the selected DAC8568 chip through an SPI interface, and the output voltage range is set through an external reference voltage.

[0022] Optionally, the S4 specifically includes:

[0023] S41, an agent is defined for each voltage output channel, and the agent constructs a state space s t according to real-time feedback signals of voltage output, the state space s t includes a voltage output value, external environment feedback and a running state of the optical chip;

[0024] S42, an action space a t is designed for each agent, the action space a t determines an adjustment amount of voltage output, and each action corresponds to an increase or decrease operation of the voltage output channel;

[0025] S43, a reward function R based on multiple dimensions is designed for each agentt , the reward function R t measures the accuracy and stability of the voltage output:

[0026]

[0027] where V target represents the target voltage value, V output represents the actual voltage output value, ∈ and δ represent smoothing factors, exp represents the exponential function, E stability represents the stability measure of the voltage output, γ1, γ2 and γ3 represent weight parameters, represents the rate of change of the voltage output over time, ΔV output represents the voltage change rate, n represents the number of agents participating in voltage control, p, q and r represent adjustment exponents;

[0028] S44, based on the policy network μ(s t |θ) and the Q network , the decision of the agent is made, the policy network is used to generate the voltage adjustment strategy, and the Q network is used to evaluate the value of each action;

[0029] S45, sharing the experience pool among multiple agents through the deep deterministic policy gradient algorithm, and updating the policy:

[0030]

[0031] where θ represents the parameters of the policy network, η represents the learning rate, T represents the total number of steps in the policy network training, represents the gradient of the policy network, b(s t ) represents the baseline function, |a t -μ9s t |θ represents the error between the action a t and the action predicted by the policy network μ(s t |θ), ∈1 and v2 represent smoothing factors, σ t represents the uncertainty of the action, σ target represents the target uncertainty, R i (s t ,a t ) represents the reward function gradient of the agent i, represents the parameters of the agent i, represents the loss function, β i represents the adjustment term, represents the gradient of the parameter ;

[0032] S46, the intelligent agent dynamically adjusts the action strategy according to the external environment feedback and global information, and each intelligent agent adjusts the voltage output value according to the global information of the experience pool.

[0033] Optionally, the control and communication circuit integrated on the PCB specifically includes an ESP32-S3 single-chip microcomputer, a DAC8568 chip, a 74HC4051 data selector, and a power management module.

[0034] The beneficial effects of the present application are:

[0035] First of all, by using the ESP32-S3 single-chip microcomputer as the core controller, the present application communicates with multiple DAC8568 chips to achieve high-precision voltage output of 64 independent channels. Compared with traditional voltage source systems, the present application has higher precision and flexibility, and can meet the demand for precise voltage control of modern optical chips. Especially through full-duplex communication via the SPI interface, 16-bit digital signals are transmitted each time, ensuring that the system can adjust the voltage output in real time and efficiently, and ensuring the response speed and stability of the system.

[0036] In addition, through the dynamic selection mechanism of the 74HC4051 data selector, the present application can flexibly control the output voltage of each channel, realizing independent control and real-time adjustment of multiple voltage channels. Combined with GPIO control selection signals, the system can dynamically adjust the output channels as needed in complex application scenarios, improving the adaptability and expansion capability of the system. The system also fine-tunes the output voltage range through an external reference voltage, making it flexible to adapt to different application requirements.

[0037] By introducing the multi-agent deep deterministic policy gradient algorithm, the present application can analyze and adjust the voltage value of the voltage output channel in real time based on external environment feedback and the running state of the optical chip. This algorithm can utilize the collaborative learning mechanism of multiple agents to optimize the voltage output strategy, and share global information through the experience pool to ensure the stability and accuracy of the voltage in the global range. The algorithm can also adaptively adjust according to a complex reward function to improve the control accuracy of the system in dynamic environments, especially suitable for optical chip application scenarios that require high precision and low fluctuation.

[0038] Finally, by integrating the control and communication circuit on a PCB, the present application realizes compact integrated design, greatly reducing the size of the system and improving the integration and reliability of the system. This integrated design is not only suitable for high-density optical chip application scenarios, but also ensures the stability and continuous power supply capability of the system under high-power conditions through the power management module, thereby ensuring the long-term operation and reliability of the system. BRIEF DESCRIPTION OF DRAWINGS

[0039] The accompanying drawings are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and are meant to explain the present application but are not intended to limit the application. In the drawings:

[0040] Figure 1 A flow chart of a voltage source matrix control method based on an optical chip according to the present application;

[0041] Figure 2 A voltage output value dynamic adjustment flow chart of a voltage source matrix control method based on an optical chip according to the present application. DETAILED DESCRIPTION

[0042] The present application will now be described in further detail with reference to the drawings. These drawings are simplified schematic diagrams which show the basic structure of the present application in a schematic manner only, and thus only show the components relevant to the present application.

[0043] Reference Figure 1 and Figure 2 A voltage source matrix control method based on an optical chip, comprising the following steps:

[0044] S1, using an ESP32-S3 single-chip microcomputer as a core controller, responsible for data communication with eight DAC8568 chips through an SPI interface, the DAC8568 chip including 8 independent 16-bit precision voltage output channels, a total of 64 voltage output channels;

[0045] S2, channel selection is achieved through a 74HC4051 data selector, the ESP32-S3 single-chip microcomputer controls the selection signal of the 74HC4051 data selector through GPIO, dynamically selecting a specific DAC8568 chip and voltage output channel;

[0046] S3, the voltage output value of each voltage output channel is adjusted in real time by the ESP32-S3 single-chip microcomputer, a 16-bit digital signal is sent to the selected DAC8568 voltage output channel through the SPI interface, the DAC8568 converts the digital signal into a corresponding analog voltage and outputs through the corresponding voltage output channel;

[0047] S4, based on the multi-agent deep deterministic policy gradient algorithm, real-time analysis is performed according to external environment feedback and the running state of the optical chip, and the voltage output value of each voltage output channel is dynamically adjusted;

[0048] S5, all control and communication circuits are integrated on a PCB board.

[0049] In this embodiment, the ESP32-S3 single-chip microcomputer communicates data with eight DAC8568 chips through a full-duplex SPI interface, and transmits a 16-bit digital signal each time.

[0050] In this embodiment, the ESP32-S3 microcontroller controls the 3-bit selection signal of the 74HC4051 data selector through GPIO to achieve dynamic selection and switching of eight DAC8568 chips and each voltage output channel.

[0051] In this embodiment, the ESP32-S3 microcontroller sends a 16-bit digital signal to the voltage output channel of the selected DAC8568 chip via the SPI interface, and the output voltage range is set by an external reference voltage.

[0052] In this embodiment, the S4 specifically includes:

[0053] S41. Define an agent for each voltage output channel. The agent constructs a state space s according to the real-time feedback signal of the voltage output. t , the state space s t Including voltage output value, external environment feedback and the operating status of the optical chip;

[0054] S42. Design an action space a for each agent t , the action space a t Determines the adjustment amount of voltage output, and each action corresponds to the increase or decrease operation of the voltage output channel;

[0055] S43. Design a reward function R based on multi-dimensional factors for each agent t , the reward function R t To measure the accuracy and stability of voltage output:

[0056]

[0057] Among them, V target Indicates the target voltage value, V output represents the actual voltage output value, ∈ and δ represent smoothing factors, exp represents the exponential function, E stability represents the stability measure of the voltage output, γ1, γ2 and γ3 represent weight parameters, Indicates the rate of change of voltage output over time, ΔV output represents the voltage change rate, n represents the number of agents involved in voltage control, and p, q, and r represent the regulation index;

[0058] S44, based on the policy network μ(s t |θ) and Q network The policy network is used to generate a voltage regulation policy, and the Q network is used to evaluate the value of each action;

[0059] S45, sharing the experience pool among multiple agents through the deep deterministic policy gradient algorithm, and updating the policy:

[0060]

[0061] wherein, θ represents the parameters of the policy network, η represents the learning rate, T represents the total number of steps in the policy network training, represents the gradient of the policy network, b(s t ) represents the baseline function, |a t -μ(s t |θ) represents the error between the action a t and the action μ(s t |θ) predicted by the policy network, ∈1 and ∈2 represent the smoothing factors, σ t represents the uncertainty of the action, σ target represents the target uncertainty, R i (s t ,a t ) represents the reward function gradient of the agent i, represents the parameters of the agent i, represents the loss function, β i represents the adjustment term, represents the gradient of the parameters .

[0062] S46, the agent dynamically adjusts the action policy according to the external environment feedback and the global information, and each agent adjusts the voltage output value according to the global information of the experience pool.

[0063] In the embodiment, the control and communication circuit integrated on the PCB specifically includes an ESP32-S3 single-chip microcomputer, a DAC8568 chip, a 74HC4051 data selector, and a power management module.

[0064] Example 1

[0065] In order to verify the feasibility of the application in implementation, the application is applied to the production line of a large optical communication equipment factory. There are multiple high-power optical chips in the production line that need accurate voltage control. The traditional voltage source system uses single-chip DAC control, and the number of voltage output channels of the system is limited, and the voltage regulation response time is slow, which cannot adapt to complex production requirements. In the actual production application of the factory, due to the extremely high voltage precision requirement of the optical chip, the traditional system is difficult to provide sufficient voltage stability, resulting in a high error rate of up to 5% in the operation of the optical chip. Such errors are unacceptable in high-precision optical sensing and optical communication equipment, directly affecting the performance of the optical chip and the out-of-factory qualification rate of the product.

[0066] To solve the above problems, the factory decided to use the voltage source matrix control method based on optical chip. In actual application scenarios, ESP32-S3 single-chip microcomputer is used as the core controller, and communication is carried out with eight DAC8568 chips through the SPI interface, forming 64 independent voltage output channels. Each channel can accurately control 16-bit voltage output, covering the voltage demand range of the device. In this scenario, ESP32-S3 controls 74HC4051 data selector through GPIO interface to select the output channel of DAC8568 chip in real time, ensuring that multiple channels can flexibly and independently regulate voltage.

[0067] To verify the effectiveness of the method, the factory conducted a 3-month actual test, focusing on the voltage stability of the optical chip and the overall working efficiency of the device.

[0068] Table 1 Performance comparison table of traditional method and voltage source matrix control method based on optical chip

[0069] Test item Conventional method Method of the present application Voltage error rate 5.0％ 0.5％ Voltage adjustment response time 20 ms 3 ms Voltage fluctuation range 0.025V 0.005V Product qualification rate 92.0％ 98.5％ Equipment failure rate 4 times 2 times Installation and maintenance cost 150,000 yuan / year 100,000 yuan / year Production line space occupation 50 square meters 42.5 square meters Continuous working time 72 hours 120 hours

[0070] The above Table 1 shows the differences between the voltage source matrix control system based on optical chip of the present application and the traditional voltage source system in multiple key performance indicators. First, in terms of voltage error rate, the error rate of the traditional system is 5.0%, while the system of the present application significantly reduces the error rate to 0.5% through the dynamic regulation of the multi-agent deep deterministic policy gradient algorithm. This indicates that the system of the present application can provide higher stability and reliability in voltage precision.

[0071] In terms of voltage adjustment response time, the response speed of the traditional system is slower, and it takes an average of 20 milliseconds to adjust the voltage, while the system of the present application shortens the response time to 3 milliseconds through an efficient control and communication architecture, greatly improving the real-time performance of the system. This improvement is of great significance for quickly responding to external environmental changes and ensuring the stable operation of the optical chip.

[0072] Voltage fluctuation range is also an advantage of the system of the present application. The voltage fluctuation range of the traditional system is 0.025V, while the system of the present application controls the fluctuation within 0.005V, further ensuring the stability of voltage output. This is crucial for the high-precision work requirements of the optical chip, and can reduce the impact of voltage fluctuation on chip performance.

[0073] In terms of product pass rate, the improvement of the system of the present application is also very obvious. The pass rate of the traditional system is 92.0%, while the system of the present application improves the pass rate to 98.5% through more accurate voltage control, significantly reducing the number of unqualified products and significantly improving production efficiency and product quality.

[0074] The reduction in device failure rate reflects the reliability of the system in long-term operation. The traditional system has an average of 4 failures per month, while the system of the present application reduces the failure rate to 2 times, which means that the device runs more stably, reducing maintenance costs and downtime.

[0075] The system of the present application also has a reduction in installation and maintenance costs. The annual maintenance cost of the traditional system is about 150,000 yuan, while the cost of the system of the present application is 100,000 yuan, saving 33% of the cost. Through the highly integrated design, the maintenance of the device is more convenient, and at the same time, the potential failure points are reduced.

[0076] In addition, in terms of production line space occupation, the system of the present application also shows higher integration. The traditional system occupies a space of 50 square meters, while the system of the present application reduces the space occupation to 42.5 square meters by integrating the control and communication circuit on a PCB board, reducing the space requirement by about 15%. This provides more flexibility for the layout of the production line and the installation of other devices.

[0077] In terms of continuous working time, the system of the present application also shows obvious advantages. The traditional system has voltage fluctuations and failures after 72 hours of continuous work, while the system of the present application can work continuously for 120 hours under high load conditions, further verifying its stability in long-term operation.

[0078] From the above analysis, it can be seen that the voltage source matrix control system of the present application is superior to the traditional system in multiple key performance indicators, showing the advantages of higher voltage precision, faster response speed, more stable operation, lower maintenance cost, and higher integration, greatly improving the working efficiency of the optical chip and the overall reliability of the system.

[0079] The above is only the preferred specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any skilled person in the art can make equivalent replacements or changes to the technical solutions and inventive concepts of the present application within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A voltage source matrix control method based on an optical chip, characterized in that: The steps include: S1, using the ESP32-S3 microcontroller as the core controller, responsible for data communication with eight DAC8568 chips through the SPI interface. The DAC8568 chip includes eight independent 16-bit precision voltage output channels, totaling 64 voltage output channels; S2. Channel selection is implemented through the 74HC4051 data selector. The ESP32-S3 microcontroller controls the selection signal of the 74HC4051 data selector through GPIO to dynamically select a specific DAC8568 chip and voltage output channel. S3, using the ESP32-S3 microcontroller to adjust the voltage output value of each voltage output channel in real time, sending a 16-bit digital signal to the selected DAC8568 voltage output channel through the SPI interface. The DAC8568 converts the digital signal into a corresponding analog voltage and outputs it through the corresponding voltage output channel; S4, based on a multi-agent deep deterministic policy gradient algorithm, performs real-time analysis based on external environment feedback and the operating status of the optical chip, and dynamically adjusts the voltage output value of each voltage output channel; S5. Integrate all control and communication circuits on a single PCB board; The S4 specifically includes: S41. Define an agent for each voltage output channel. The agent constructs a state space s according to the real-time feedback signal of the voltage output. t , the state space s t Including voltage output value, external environment feedback and the operating status of the optical chip; S42. Design an action space a for each agent t , the action space a t Determines the adjustment amount of voltage output, each action corresponds to the increase or decrease operation of the voltage output channel; S43. Design a reward function R based on multi-dimensional factors for each agent t , the reward function R t To measure the accuracy and stability of voltage output: Among them, V target Indicates the target voltage value, V output represents the actual voltage output value, ∈ and δ represent smoothing factors, exp represents the exponential function, E stability represents the stability measure of the voltage output, γ1, γ2 and γ3 represent weight parameters, Indicates the rate of change of voltage output over time, ΔV output represents the voltage change rate, n represents the number of agents involved in voltage control, and p, q, and r represent the regulation index; S44, based on the policy network μ(s t |θ) and Q network The policy network is used to generate a voltage regulation policy, and the Q network is used to evaluate the value of each action; S45. Share the experience pool among multiple agents through the deep deterministic policy gradient algorithm to update the strategy: Among them, θ represents the parameters of the policy network, η represents the learning rate, and T represents the total number of time steps in the policy network training. represents the gradient of the policy network, b(s t ) represents the baseline function, |a t -μ(s t |θ)| represents action a t and the policy network predicts action μ(s t |θ), ∈1 and ∈2 represent the smoothing factor, σ t represents the uncertainty of the action, σ target represents the target uncertainty, R i (s t ,a t ) represents the gradient of the reward function of agent i, represents the parameters of agent i, represents the loss function, β i Indicates the adjustment item, Representation parameters gradient; S46. The intelligent agent dynamically adjusts its action strategy based on external environment feedback and global information. Each intelligent agent adjusts its voltage output value based on the global information in the experience pool.

2. The voltage source matrix control method based on an optical chip according to claim 1, characterized in that: The ESP32-S3 microcontroller communicates with eight DAC8568 chips via a full-duplex SPI interface, transmitting 16-bit digital signals each time.

3. The voltage source matrix control method based on an optical chip according to claim 1, characterized in that: The ESP32-S3 microcontroller controls the 3-bit selection signal of the 74HC4051 data selector through GPIO to achieve dynamic selection and switching of eight DAC8568 chips and each voltage output channel.

4. The voltage source matrix control method based on an optical chip according to claim 1, characterized in that: The ESP32-S3 microcontroller sends a 16-bit digital signal to the voltage output channel of the selected DAC8568 chip through the SPI interface, and the output voltage range is set by the external reference voltage.

5. The voltage source matrix control method based on an optical chip according to claim 1, characterized in that: The control and communication circuit integrated on the PCB board specifically includes an ESP32-S3 single-chip microcomputer, a DAC8568 chip, a 74HC4051 data selector and a power management module.

Images