Device for testing electrical interface performance of high-speed optical module

Abstract

The invention provides a device for testing the electrical interface performance of a high-speed optical module, and belongs to the technical field of optical module testing, and the device comprises an upper computer which is used for issuing a control command; the high-speed switch chip is in communication connection with the upper computer and is used for generating a random code signal according to the control command; the input end of each electric link module is connected with the signal output end of the high-speed switching chip, the output end of each electric link module is connected with the optical module to be detected, the insertion loss of each electric link module is consistent, and the reflection intensity is in gradient increase; the to-be-tested optical module is used for comparing the random code signal carrying the reflection interference with a standard random code based on an internal oDSP chip, generating an error rate and reporting the error rate to the upper computer for visualization; and the MCB calibration module is detachably connected with each electric link module, so that the problems that the existing test scene is single, the parameter control is inaccurate, and the result is separated from the actual application are effectively solved.

Description

Technical Field

[0001] This invention relates to the field of optical module testing technology, and in particular to a device for testing the performance of high-speed optical module electrical interfaces. Background Technology

[0002] In the research and development and production of high-speed optical modules, the SVT (System Verification Test) stage requires adaptation testing on different customers' service boards. However, existing testing methods have the following key technical challenges: The root cause of performance differences is unclear: When different customer boards are adapted to the same optical module, the test performance often shows significant differences. Existing technology cannot quickly distinguish whether the difference is due to electrical link design defects of the customer board or quality problems of the optical module itself, resulting in long problem localization cycles and high costs.

[0003] Troubleshooting performance issues: In the development of optical module products, the test performance of different versions may fluctuate. However, existing testing methods lack the ability to accurately control the key parameters of the electrical interface link, making it difficult to locate specific fault points that cause performance differences, such as electrical link reflection, connector performance, or chip characteristic failures.

[0004] The selection of oDSP chips lacks a basis: When the same type of optical module uses oDSP chips from different suppliers, the existing technology cannot separate the influence of optical module design and chip performance in a unified testing environment, making it difficult to objectively compare the advantages and disadvantages of different oDSP chips. This leads to high risk in the selection of key materials and affects the overall design quality of the optical module.

[0005] Furthermore, PAM4 signals, as a commonly used modulation signal in high-speed optical modules, are extremely sensitive to electrical link reflections and noise interference. Existing testing equipment lacks differentiated electrical link scenarios designed to address this characteristic, failing to simulate varying degrees of harshness in real-world electrical transmission environments. This leads to test results that are disconnected from actual application scenarios, making it difficult to comprehensively evaluate the electrical interface link design capabilities of optical modules. Especially in the ultra-high-speed, low-latency transmission requirements of 6G mobile communication, and in scenarios with multi-device concurrent access and high-bandwidth data interaction in Wi-Fi 6, the challenges of reflection interference and performance adaptation faced by optical module electrical interfaces are further exacerbated, making it even more difficult for existing testing equipment to meet the accurate testing needs of these emerging scenarios.

[0006] Therefore, this invention proposes a device for testing the performance of high-speed optical module electrical interfaces. Summary of the Invention

[0007] This invention provides a device for testing the performance of a high-speed optical module electrical interface, in order to solve the aforementioned technical problems.

[0008] This invention provides an apparatus for testing the performance of a high-speed optical module electrical interface, comprising: The host computer is used to issue control commands. A high-speed switching chip, which is connected to the host computer, is used to generate random code signals according to the control commands. Several differentiated electrical link modules are provided. The input of each electrical link module is connected to the signal output of the high-speed switching chip. It is used to receive random code signals and apply reflection interference to realize serial testing of environmental simulation. The output of each electrical link module is connected to the optical module under test. It is used to transmit random code signals carrying reflection interference to the optical module under test. The insertion loss of each electrical link module is consistent, the reflection intensity increases in a gradient, and each electrical link module corresponds to an application scenario with different reflection intensities. The optical module under test is used to compare the random code signal carrying reflection interference with the standard random code based on the internal oDSP chip, generate the bit error rate, and report it to the host computer. The host computer is used to analyze the bit error rate, determine the comparison results and judgment results, and visualize them. The MCB calibration module is detachably connected to each electrical link module and is used to pre-calibrate the insertion loss and reflection intensity of each electrical link module.

[0009] Preferably, the high-speed switching chip is a Broadcom Strata DNX series chip or a Cisco Silicon One series chip.

[0010] Preferably, the MCB calibration module includes a vector network analyzer for collecting the S-parameters of each electrical link module and adjusting the attenuator parameters of the electrical link modules to achieve insertion loss and reflection intensity calibration.

[0011] Preferably, the optical module under test is an optical module that supports PAM4 signal transmission.

[0012] Preferably, the high-speed switching chip includes: The parameter extraction unit is used to extract the random code generation parameters of the control command and store the verification instructions; The determination unit is used to determine whether the random code generation parameters are consistent with the random code generation configuration content pre-stored inside the high-speed switching chip according to the storage verification instruction. If they are inconsistent, the difference information between the random code generation parameters and the pre-stored random code generation configuration content will be used. A determining unit is configured to determine the number of different types of differences in the random code generation parameters and the difference symbol for each difference type based on the difference information. The recoding unit is used to extract the important and unimportant bits occupied by the difference symbols under each difference type, determine the temporal distribution parameters of the corresponding difference symbols in the difference information and the reflection intensity level factor of the current electrical link module to be accessed, and recode the corresponding difference symbols according to the type attribute of each difference type to generate a random code signal carrying interference features corresponding to the difference information. If they match, a random code signal is directly generated based on the pre-stored random code generation configuration.

[0013] Preferably, the host computer includes: The combined effect of the reflected interference of the electrical link module and the interference characteristics of the random code signal is determined, and the initial degree of influence is obtained by matching according to the influence-degree lookup table; If the random code signal is directly generated, the initial degree of influence shall be regarded as the current degree of influence; If the random code signal is generated by recoding, determine the degree of inconsistency of the difference information and adjust the initial degree of influence to obtain the current degree of influence; Based on the current level of influence, the visualization results of scene deviation interference are optimized.

[0014] Preferably, the parameter extraction unit includes: The syntax detection subunit is used to perform syntax parsing on the control command issued by the host computer before extracting the random code generation parameters of the control command. When a syntax error is detected in the control command, the command configuration metafunction and object mapping matrix are generated based on the design architecture of the high-speed switching chip. An error-locking subunit is used to lock the location of a syntax error and capture the context information of the syntax error location; The command correction subunit is used to generate initial completion parameters based on the completion rules in the command configuration metafunction according to the syntax error type, and adjust the initial completion parameters based on the current hardware operating status parameters of the high-speed switching chip. The adjusted completion parameters are then verified for cross-functional consistency based on the object mapping matrix. After successful verification, context information is embedded to obtain a corrected command that meets the requirements of syntax, semantics, and hardware execution feasibility. The command configuration metafunction generates completion parameters that match the hardware execution logic for the missing or erroneous fields in the context information according to the parameter interface specification of the high-speed switching chip. The object mapping matrix is ​​used to verify the cross-functional matching of the corrected parameters. Extract sub-units, used to extract the random code generation parameters of the corrected command; When no syntax errors are detected in the control command, the random code generation parameters of the control command are extracted.

[0015] Preferably, the optical module under test includes: The feature extraction unit is used to denote the random code signal carrying reflection interference as the Rx signal, and extract multi-dimensional feature parameters associated with the reflection interference from the Rx signal. The multi-dimensional feature parameters include: the timing offset corresponding to each PAM4 level, the amplitude attenuation coefficient corresponding to each PAM4 level, the symbol edge jitter value corresponding to each PAM4 level, and the reflection level of the corresponding electrical link module. The dynamic calibration unit is used to acquire the original PAM4 standard random code generated by the high-speed switching chip, denoted as the Ref signal, and to perform dynamic hierarchical calibration on the Ref signal based on multi-dimensional feature parameters. For each PAM4 level, the timing compensation sub-algorithm is used to offset the symbol trigger time of the corresponding level of the Ref signal according to the timing offset. The amplitude of the symbol at the corresponding level of the Ref signal is scaled according to the amplitude attenuation coefficient using the amplitude correction sub-algorithm; The edge smoothing sub-algorithm reduces the symbol edge jitter of the corresponding level of the Ref signal to the corresponding adaptation range of the symbol edge jitter value; Based on the reflection level, the overall amplitude and phase of the Ref signal after each level calibration are corrected twice to generate the Ref0 signal that dynamically matches the interference characteristics of the Rx signal and the electrical link reflection level. The Ref0 signal is a standard random code signal after dynamic calibration. The bit error rate calculation unit is used to compare the Rx signal and the Ref0 signal bit by bit and level by level. Based on the bit error sensitivity weight of each PAM4 level, it calculates the number of inconsistencies N1 after weighting and combines it with the total number of bits N2 to calculate the bit error rate.

[0016] Compared with the prior art, the beneficial effects of this application are as follows: Precise command control is achieved through a host computer, and compliant random code signals are generated using a dedicated high-speed switching chip. Differentiated electrical link modules with consistent insertion loss and increasing reflection intensity gradients are used to realistically simulate reflection interference in various practical application scenarios. An MCB calibration module ensures the accuracy of electrical link parameters. The optical module under test (DUT) accurately performs signal comparison and bit error rate calculation through its internal oDSP chip. Finally, the host computer performs result analysis and visualization. This effectively solves the problems of existing test scenarios being too limited, parameter control being inaccurate, and results being disconnected from actual applications. It provides reliable testing support for performance verification in the R&D stage and quality inspection in the production stage of high-speed optical modules, significantly improving the accuracy and scenario coverage of optical module electrical interface performance testing. Simultaneously, this device is adaptable to the technical requirements of 6G and Wi-Fi 6 scenarios, capable of simulating complex reflection interference environments under 6GTbps transmission and multiple concurrent Wi-Fi 6 devices. It provides a highly reliable optical module performance verification solution for emerging application scenarios such as 6G base station construction and Wi-Fi 6 network upgrades, facilitating the implementation and promotion of related high-speed communication technologies.

[0017] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0018] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0019] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a structural diagram of a device for testing the performance of a high-speed optical module electrical interface according to an embodiment of the present invention. Detailed Implementation

[0020] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0021] This invention provides a device for testing the performance of a high-speed optical module electrical interface, such as... Figure 1 As shown, it includes: The host computer is used to issue control commands. A high-speed switching chip, which is connected to the host computer, is used to generate random code signals according to the control commands. Several differentiated electrical link modules are provided. The input of each electrical link module is connected to the signal output of the high-speed switching chip. It is used to receive random code signals and apply reflection interference to realize serial testing of environmental simulation. The output of each electrical link module is connected to the optical module under test. It is used to transmit random code signals carrying reflection interference to the optical module under test. The insertion loss of each electrical link module is consistent, the reflection intensity increases in a gradient, and each electrical link module corresponds to an application scenario with different reflection intensities. The optical module under test is used to compare the random code signal carrying reflection interference with the standard random code based on the internal oDSP chip, generate the bit error rate, and report it to the host computer. The host computer is used to analyze the bit error rate, determine the comparison results and judgment results, and visualize them. The MCB calibration module is detachably connected to each electrical link module and is used to pre-calibrate the insertion loss and reflection intensity of each electrical link module.

[0022] Preferably, the high-speed switching chip is a Broadcom Strata DNX series chip or a Cisco Silicon One series chip.

[0023] Preferably, the MCB calibration module includes a vector network analyzer for collecting the S-parameters of each electrical link module and adjusting the attenuator parameters of the electrical link modules to achieve insertion loss and reflection intensity calibration.

[0024] Preferably, the optical module under test is an optical module that supports PAM4 signal transmission.

[0025] In this embodiment, the host computer is a control terminal with data processing and command issuance capabilities. It is typically an industrial computer or high-performance PC equipped with dedicated testing software, such as an industrial control computer with a Core i7 processor and 16GB of memory. The control commands it issues contain parameters required for the test, specifically including the random code pattern (e.g., PRBS31, PRBS15), transmission rate (e.g., 100Gbps, 200Gbps), and test duration (e.g., 5 minutes, 10 minutes). Specifically, test control software developed based on LabVIEW or Python is installed on the host computer. A communication connection is established with the high-speed switching chip via Ethernet or PCIe interface. The software interface allows for parameter input boxes; after the user inputs the parameters and clicks the "issue command" button, the control command is transmitted to the high-speed switching chip.

[0026] The high-speed switching chip is the core hardware responsible for receiving control commands and generating high-speed random code signals. It has high-speed signal processing and generation capabilities. Specifically, the high-speed switching chip is soldered onto a customized PCB motherboard. The motherboard needs to be equipped with a stable power supply circuit, a high-precision clock circuit, and a communication interface circuit to ensure that the chip can stably receive control commands from the host computer and generate corresponding random code signals according to the code pattern and rate parameters in the command. For example, when receiving the PRBS31, 100Gbps command, it generates a binary random code stream with a rate of 100Gbps that conforms to the PRBS31 encoding rules.

[0027] Several differentiated electrical link modules are hardware units used to simulate reflection interference in different real-world application scenarios. Several are usually set to 3-5 modules to cover different scenarios. The core of differentiation is reflected in the different reflection intensities of each module. Consistent insertion loss means that the attenuation of random code signals by each module is the same (e.g., all controlled at 2dB±0.1dB). The reflection intensity increases in a gradient, that is, the reflection coefficient of each module gradually increases. For example, the reflection coefficient of the first module is -30dB (corresponding to the low reflection scenario of short-distance data centers), the second is -25dB (corresponding to the medium-distance transmission scenario), and the third is -20dB (corresponding to the high reflection scenario of long-distance complex links). Each reflection intensity corresponds to a real-world application scenario. Specifically, each module uses shielded coaxial cable or high-frequency PCB transmission line to build a signal path. An adjustable attenuator (such as Mini-Circuits' ATT-0200 attenuator) is connected in series in the path to calibrate insertion loss, and different terminating load resistors (such as 50Ω and 75Ω resistors) are connected in parallel to adjust the reflection intensity. The module input is connected to the signal output of the high-speed switching chip through an SFP+ interface, and the output is connected to the electrical interface of the optical module under test through a QSFP28 interface. A serial test mode is adopted, that is, the individual modules are connected to the optical module under test one by one through relay switching to avoid signal interference between modules.

[0028] The optical module under test is the object of this test, namely the high-speed optical transmission module whose electrical interface performance needs to be verified. It must support PAM4 signal transmission (PAM4 is a 4-level pulse amplitude modulation technology, which is the mainstream modulation method for 100G and above high-speed optical modules). Specifically, a 100GBASE-SR4 optical module or a 200GBASE-FR4 optical module can be selected. Its internal oDSP chip (optical digital signal processing chip) is the core of signal comparison and bit error rate calculation. Broadcom's BCM84750oDSP chip is used. This chip has high-speed signal sampling, symbol comparison and bit error rate statistics functions. Specifically, the optical module under test is designed according to the SFF-8636 industry protocol. The electrical interface adopts the QSFP28 specification to receive signals transmitted by the electrical link module. The internal oDSP chip implements its functions through firmware: first, it receives a random code signal carrying reflection interference (denoted as Rx signal), and at the same time, it retrieves the standard random code signal stored in the chip's internal Flash (denoted as Ref signal, which is from the same source as the random code generated by the high-speed switching chip). The hardware logic circuit compares the Rx signal and the Ref signal symbol by symbol, counts the number of inconsistent symbols, calculates the bit error rate by combining the total number of symbols, and finally reports the bit error rate data to the host computer through the I2C communication interface.

[0029] The MCB calibration module is an auxiliary unit used for pre-calibrating the parameters of the electrical link module. MCB is short for Module Calibration Board. Its core component is a vector network analyzer used to measure the transmission and reflection parameters of the electrical link (specifically, a Keysight N5247A vector network analyzer can be used). S-parameters are parameters that describe the signal transmission characteristics of the electrical link. Among them, the S11 parameter reflects the reflection characteristics at the input end of the link, and the S21 parameter reflects the transmission attenuation characteristics of the link. The attenuator parameter refers to the attenuation amount of the adjustable attenuator in the electrical link module (e.g., adjustable from 0-10dB). Specifically: Connect the MCB calibration module to each electrical link module via a high-frequency test cable, start the vector network analyzer, set the test frequency range (e.g., 100MHz-50GHz), and collect the S11 and S21 parameters of each electrical link module; if the S21 parameter of a certain module is -2.5dB (i.e., insertion loss of 2.5dB, exceeding the requirement of 2dB±0.1dB), then adjust the attenuation of the adjustable attenuator in the module to reduce it by 0.5dB, so that the S21 parameter reaches -2.0dB; if the S11 parameter of a certain module is -28dB (reflection intensity does not meet the design requirement of -30dB), then replace the parallel terminating load resistor with a resistor whose resistance is closer to 50Ω, so that the S11 parameter is optimized to -30dB, ensuring that the insertion loss of all electrical link modules is consistent and the reflection intensity gradient meets the design.

[0030] Suppose we are testing the electrical interface performance of a 100GPAM4 high-speed optical module, and let it pass through three different electrical link modules (corresponding to low, medium, and high reflection intensity scenarios). The bit error rate collected by the host computer is as follows: Low reflection scenario - Electrical link module 1: BER1; Medium reflection scenario - Electrical link module 2: BER2; High reflection scenario - Electrical link module 3: BER3.

[0031] It should be noted that the comparison results, i.e., the horizontal comparison, are: BER3 / BER2, BER2 / BER1; trend analysis: the change of bit error rate with transmission intensity.

[0032] The judgment result needs to be combined with the preset qualified threshold to determine whether the performance of different reflection scenarios meets the standard. If the bit error rate is less than the corresponding threshold, it is considered qualified; otherwise, it is considered unqualified. The preset qualified threshold is set by experts in advance and can be used directly.

[0033] In this embodiment, the visualization is achieved using Python's Matplotlib or LabVIEW to present the results intuitively. Bar chart: The horizontal axis represents the reflection scene (low / medium / high), and the vertical axis represents the bit error rate. The three bars correspond to BER1, BER2, and BER3, respectively, clearly showing the differences in bit error rate in different scenes.

[0034] Trend line chart: The horizontal axis represents the reflection intensity level (1 / 2 / 3), and the vertical axis represents the bit error rate. The line connects the three data points, visually showing the trend that the stronger the reflection, the higher the bit error rate.

[0035] Threshold Comparison Chart: Adding dashed lines to the chart represents the acceptable threshold. The relative position of the bit error rate of each scenario with the dashed line can quickly determine whether the standard is met.

[0036] The beneficial effects of the above technical solution are as follows: precise command control is achieved through the host computer; compliant random code signals are generated by a dedicated high-speed switching chip; differentiated electrical link modules with consistent insertion loss and increasing reflection intensity gradients are used to realistically simulate reflection interference in different practical application scenarios; the accuracy of electrical link parameters is ensured by the MCB calibration module; the optical module under test accurately completes signal comparison and bit error rate calculation through its internal oDSP chip; and finally, the host computer performs result analysis and visualization. This effectively solves the problems of existing single test scenarios, inaccurate parameter control, and disconnect between results and actual applications. It can provide reliable test support for performance verification in the R&D stage and quality inspection in the production stage of high-speed optical modules, and significantly improve the accuracy and scenario coverage of optical module electrical interface performance testing.

[0037] This invention provides an apparatus for testing the performance of a high-speed optical module electrical interface, wherein the high-speed switching chip includes: The parameter extraction unit is used to extract the random code generation parameters of the control command and store the verification instructions; The determination unit is used to determine whether the random code generation parameters are consistent with the random code generation configuration content pre-stored inside the high-speed switching chip according to the storage verification instruction. If they are inconsistent, the difference information between the random code generation parameters and the pre-stored random code generation configuration content will be used. A determining unit is configured to determine the number of different types of differences in the random code generation parameters and the difference symbol for each difference type based on the difference information. The recoding unit is used to extract the important and unimportant bits occupied by the difference symbols under each difference type, determine the temporal distribution parameters of the corresponding difference symbols in the difference information and the reflection intensity level factor of the current electrical link module to be accessed, and recode the corresponding difference symbols according to the type attribute of each difference type to generate a random code signal carrying interference features corresponding to the difference information. If they match, a random code signal is directly generated based on the pre-stored random code generation configuration.

[0038] In this embodiment, a difference sign bit Si,i=1,2,3,...,N is set for each difference type, where N is the number of types, and the difference sign bit Si occupies a significant number of bits. Unimportant digits The difference sign bit Si corresponds to the type attribute Ti of the difference type; Determine the temporal distribution parameters of the difference symbol Si in the difference information. Simultaneously, obtain the reflection intensity level factor Rj of the current electrical link module to be connected; Constructing recoding formulas The difference symbol Si is re-encoded, where f is a multi-dimensional weight allocation function based on type attribute, temporal distribution and reflection intensity. By adjusting the temporal weight of important bits, the basic weight of non-important bits and the coupling weight associated with reflection intensity, the generated interference feature carries a random code signal of interference feature corresponding to the difference information and electrical link reflection feature.

[0039] Specifically, the number of significant digits is determined based on the type attribute of the difference type. The time-series weighting benchmark Wi(Ti) is used as a reference, and the time-series weighting benchmark is modified by combining the time-series distribution parameters to obtain the real-time time-series weights of the important bits. Where g is the time series distribution correction function; In this embodiment, The value ranges from 0 to 1, according to The difference between the corresponding occurrence time and the occurrence time of the first difference symbol in the corresponding difference information is used to match the corresponding value according to the difference-value lookup table. For example, if the difference is 100ms, the corresponding... The value is 0.7.

[0040] Based on non-significant digits The ratio to the preset benchmark value determines the basic weight of the unimportant digits. Where Mref is the preset baseline value for unimportant digits; h is the ratio mapping function; In this embodiment, .

[0041] Based on the reflection intensity level Rj of the electrical link module, calculate the coupling weight WR=k(Rj) associated with the reflection intensity, where k is the reflection intensity-coupling weight mapping function; In this embodiment, k(Rj)=a0 Rj, where a0 is a linear coefficient, typically taking the value of 1.

[0042] Encode according to the following formula: ,in, For encoding important bits, subfunction This is a subfunction for encoding non-significant bits; This is a reflection-coupled coding subfunction.

[0043] In this embodiment, the random code generation parameters may include code type, transmission rate, etc. The code type can be a random binary sequence PRBS31, PRBS15, etc.; the transmission rate can be 100Gbps, 200Gbps, etc.; the storage verification instruction is used to trigger a consistency determination between subsequent parameters and the chip's pre-stored configuration. For example, the instruction code 0x01 represents starting verification. Specifically, the chip integrates an instruction parsing circuit written in a hardware description language. When the host computer sends a data packet containing parameters and instructions via the PCIe bus, such as code type: PRBS15; rate: 200Gbps; verification instruction: 0x01, this circuit parses the data packet fields and extracts the corresponding random code generation parameters and storage verification instructions.

[0044] In this embodiment, the pre-stored random code generation configuration is a standard configuration either manufactured at the chip factory or pre-written, such as a default configuration of code pattern PRBS31 and a rate of 100Gbps. Specifically, the chip is designed with a comparison logic circuit that compares the parameters obtained by the parameter extraction unit, such as code pattern PRBS15, with the pre-stored configuration, such as code pattern PRBS31, field by field, and outputs a result indicating whether they match or do not match.

[0045] In this embodiment, the logic analysis circuit inside the chip will break down the difference information, count the number of difference types in different dimensions, and assign a preset difference symbol to each type. For example, the specific content of the parameter inconsistency is: the code pattern changes from PRBS31 to PRBS15 and the rate changes from 100Gbps to 200Gbps. At this time, the number of difference types is code pattern and rate, a total of 2 types. At this time, 01 represents code pattern difference and 10 represents rate difference.

[0046] In this embodiment, the important bits are the bits in the difference symbol that have a significant impact on the interference characteristics of the random code. For example, the first bit in the difference symbol 01 (binary) is an important bit, while the non-important bits are the bits with a smaller impact, such as the second bit in the difference symbol 01. In this embodiment, the temporal distribution parameter is the time when the difference symbol appears in the difference information; In this embodiment, the reflection intensity level factor is a quantized value corresponding to the reflection intensity of the electrical link module. The low reflection module can be set to 0.2 and the high reflection module can be set to 0.8. It is obtained through the I2C communication interface between the chip and the electrical link module.

[0047] In this embodiment, the type attribute of code pattern difference is encoding rule difference, the type attribute of rate difference is transmission rate difference, etc. It should be noted that the type attribute of each difference type is pre-defined and can be used directly. Moreover, the timing weight benchmarks determined by different type attributes are all obtained by matching from the attribute-weight lookup table. For example, the timing weight benchmark corresponding to the type attribute of code pattern difference is 0.6.

[0048] The important bit encoding subfunction, the unimportant bit encoding subfunction, and the reflection coupling encoding subfunction perform specific encoding operations on the corresponding bits of the difference symbol, such as adjusting the bit level and inserting interference pulses, to ultimately generate a random code carrying interference characteristics.

[0049] In this embodiment, if the random code generation parameters are consistent with the pre-stored configuration, the chip directly calls the pre-stored random code generation logic, such as generating a random code signal without additional interference features according to the PRBS31, 100Gbps standard rules.

[0050] In this embodiment, it is assumed that there is a difference type of code configuration deviation, the difference symbol Si is binary 1010, where the 1st and 3rd bits are important bits (corresponding to binary bits 1 and 1), the 2nd and 4th bits are non-important bits (corresponding to binary bits 0 and 0), and the reflection intensity level factor Rj of the current electrical link module to be connected is 0.6 (representing medium intensity reflection).

[0051] Important bit encoding subfunction For the first and third bits (logic 1) of the important bits, the original standard +2V level of these two bits is reduced by 10% (to +1.8V), and the rising edge of these two bits is delayed by 50 picoseconds (ps), so that the 1 of the important bit changes from the standard strong level and quasi-timing to a weak level and delayed timing interference state.

[0052] Non-significant bit encoding subfunction For the 2nd and 4th bits of the non-significant bits (logic 0, level 0V), insert a narrow pulse with an amplitude of +0.5V and a width of 20 picoseconds at the midpoint of the duration of each non-significant bit, so that the originally pure 0 becomes a 0 with a spike pulse.

[0053] Reflection Coupling Subfunction Combined with the reflection intensity level factor Rj=0.6, within the entire symbol period corresponding to the difference symbol 1010, a positive echo of +0.3V is superimposed for the first 50 picoseconds, and a negative echo of -0.3V is superimposed for the last 50 picoseconds, so that the signal of the entire difference symbol carries the reflection echo characteristic of positive first and negative later.

[0054] Through the coordinated operation of these three sub-functions, the original difference symbol 1010 is finally encoded into random code bits carrying interference characteristics. It also carries level / timing interference caused by code pattern configuration deviation, spike noise of minor configuration difference, and echo interference of medium-intensity electrical link reflection, accurately reproducing the composite interference scenario of configuration deviation and link reflection in actual applications.

[0055] The beneficial effects of the above technical solution are: when the parameters are inconsistent with the pre-stored configuration, by combining the difference information between the control command and the chip configuration and the reflection intensity of the electrical link module, a random code carrying cooperative interference characteristics is generated in a targeted manner, so that the subsequent optical module test is closer to the actual complex scenario of configuration deviation and electrical link reflection, solving the problem that traditional test can only simulate a single interference and is out of touch with actual application, and improving the authenticity and comprehensiveness of optical module electrical interface performance test.

[0056] This invention provides a device for testing the performance of a high-speed optical module electrical interface, wherein the host computer includes: The combined effect of the reflected interference of the electrical link module and the interference characteristics of the random code signal is determined, and the initial degree of influence is obtained by matching according to the influence-degree lookup table; If the random code signal is directly generated, the initial degree of influence shall be regarded as the current degree of influence; If the random code signal is generated by recoding, determine the degree of inconsistency of the difference information and adjust the initial degree of influence to obtain the current degree of influence; Based on the current level of influence, the visualization results of scene deviation interference are optimized.

[0057] In this embodiment, the reflection interference of the electrical link module refers to the distortion caused to the signal after the electrical link module reflects the random code signal, such as signal amplitude attenuation and phase shift. The interference characteristics of the random code signal are interferences such as code pattern abnormalities and timing jitter caused when the random code is generated due to differences in control command parameters and chip pre-stored configurations. The superposition effect is the combined effect of these two types of interference on the signal. The influence-degree comparison table is a comparison table that determines the combination of reflection interference level and random code interference characteristics and the corresponding influence degree value through pre-testing. For example, when the reflection interference level is 3 and the random code interference characteristic is code pattern abnormality type A, the corresponding initial influence degree in the comparison table is 0.6.

[0058] The degree of inconsistency in the difference information is a quantification of the difference between the random code generation parameters in the control command and the chip's pre-stored configuration. This difference exists across multiple dimensions, such as code type, rate, and length. The greater the deviation in each dimension, the higher the degree of inconsistency. This is achieved through standardization and normalization of the differences in each dimension, followed by a weighted calculation, which is a well-known computational method. Specifically, the host computer uses the status feedback from the high-speed switching chip to determine whether the random code is directly generated or re-encoded. If it is re-encoded, the difference information is analyzed. For example, if the code type changes from PRBS31 to PRBS15, or the rate changes from 100Gbps to 200Gbps, this multi-dimensional difference is quantified as an inconsistency level of 0.8. Then, an adjustment algorithm is used to adjust the initial impact level to obtain the current impact level. For example, multiplying the inconsistency level by 0.2 and adding the initial impact level, we get 0.8 × 0.2 + 0.6 = 0.76. If it is directly generated, the current impact level equals the initial impact level.

[0059] The visualization results of scene deviation interference are presented in the form of waveform graphs, bit error rate trend graphs, and interference level heatmaps, showing the impact of interference on the optical module's electrical interface. Optimization involves adjusting the display parameters of these graphs based on the current level of impact, such as color intensity, curve thickness, and data annotation precision, to make the interference level more intuitive. Specifically, in the host computer's visualization software, when the impact level is high, the corresponding data points in the bit error rate trend graph are marked in striking red and the curve is thickened; when the impact level is low, they are marked in green and the curve is thinned. Additionally, legends indicating high-impact and low-impact areas are added next to the graphs to help testers quickly and intuitively understand the interference situation.

[0060] The beneficial effects of the above technical solution are: determining the superimposed impact of reflection interference and random code interference and matching the initial degree; adjusting the current impact degree according to the random code generation method and the degree of inconsistency of difference information; optimizing the visualization of scene deviation interference; solving the problems of non-intuitive interference impact display and weak correlation with actual deviation in traditional testing; and improving the interpretability and analysis efficiency of high-speed optical module electrical interface performance test results.

[0061] This invention provides a device for testing the performance of a high-speed optical module electrical interface, including a parameter extraction unit comprising: The syntax detection subunit is used to perform syntax parsing on the control command issued by the host computer before extracting the random code generation parameters of the control command. When a syntax error is detected in the control command, the command configuration metafunction and object mapping matrix are generated based on the design architecture of the high-speed switching chip. An error-locking subunit is used to lock the location of a syntax error and capture the context information of the syntax error location; The command correction subunit is used to generate initial completion parameters based on the completion rules in the command configuration metafunction according to the syntax error type, and adjust the initial completion parameters based on the current hardware operating status parameters of the high-speed switching chip. The adjusted completion parameters are then verified for cross-functional consistency based on the object mapping matrix. After successful verification, context information is embedded to obtain a corrected command that meets the requirements of syntax, semantics, and hardware execution feasibility. The command configuration metafunction generates completion parameters that match the hardware execution logic for the missing or erroneous fields in the context information according to the parameter interface specification of the high-speed switching chip. The object mapping matrix is ​​used to verify the cross-functional matching of the corrected parameters. Extract sub-units, used to extract the random code generation parameters of the corrected command; When no syntax errors are detected in the control command, the random code generation parameters of the control command are extracted.

[0062] In this embodiment, syntax parsing checks whether the command format and fields conform to the rules. For example, commands must be organized in the format of field name: value;. If there is a missing separator between fields, it is judged as a syntax error.

[0063] In this embodiment, the design architecture refers to the connection method of the internal functional units of the chip and the specification of the parameter interface of each unit. For example, the module that generates random codes requires the code pattern field to be an 8-bit binary identifier and the rate field to be a 16-bit value.

[0064] When a syntax error is detected, such as a control command lacking a code pattern field, based on the chip architecture, a command configuration metafunction is generated to generate a default value for the missing code pattern field. The function and object mapping matrix records the correspondence between the chip's code pattern generation unit, rate control unit, etc., and the command parameters, which is used for subsequent parameter matching and verification.

[0065] In this embodiment, for example, in the command code pattern PRBS31 with a rate of 100Gbps, there is a missing separator between PRBS31 and the rate, and the error location is at this point. At this time, the context information, namely the valid command segments before and after the error location, such as the preceding code pattern PRBS31 and the following 100Gbps, specifically involves: locating the index of the error character using a string matching algorithm implemented by a state machine, and then temporarily storing the command content of fixed lengths before and after the error location as a reference for subsequent correction.

[0066] In this embodiment, syntax error types include missing fields and parameter out of bounds. For example, if a command rate is set to 500Gbps, but the chip supports a maximum of 400Gbps, it is considered parameter out of bounds. In this case, when the parameter is out of bounds, it is padded to the maximum allowed value, that is, 500Gbps is changed to 400Gbps. In this embodiment, the hardware operating status parameters include: the load rate of the rate control unit, the cache utilization rate, etc. For example, the original default test duration of 10 minutes is changed to 8 minutes. The adjusted parameters are verified using an object mapping matrix to check whether the 400Gbps rate and the 8-minute duration simultaneously meet the working requirements of multiple functional units such as the rate control unit and the duration scheduling unit. After verification, the completed parameters are embedded into the previously captured context information to obtain a corrected command that meets the requirements of syntax, semantics, and hardware execution feasibility, such as code type: PRBS31; rate: 400Gbps; duration: 8min.

[0067] The beneficial effects of the above technical solution are: through the coordinated operation of syntax detection, error locking, precise correction and parameter extraction, the problem that the high-speed switching chip cannot accurately generate random codes when there are syntax errors in the host computer control commands is solved, and the robustness of the testing process is greatly improved.

[0068] This invention provides an apparatus for testing the electrical interface performance of a high-speed optical module, wherein the optical module under test includes: The feature extraction unit is used to denote the random code signal carrying reflection interference as the Rx signal, and extract multi-dimensional feature parameters associated with the reflection interference from the Rx signal. The multi-dimensional feature parameters include: the timing offset corresponding to each PAM4 level, the amplitude attenuation coefficient corresponding to each PAM4 level, the symbol edge jitter value corresponding to each PAM4 level, and the reflection level of the corresponding electrical link module. The dynamic calibration unit is used to acquire the original PAM4 standard random code generated by the high-speed switching chip, denoted as the Ref signal, and to perform dynamic hierarchical calibration on the Ref signal based on multi-dimensional feature parameters. For each PAM4 level, the timing compensation sub-algorithm is used to offset the symbol trigger time of the corresponding level of the Ref signal according to the timing offset. The amplitude of the symbol at the corresponding level of the Ref signal is scaled according to the amplitude attenuation coefficient using the amplitude correction sub-algorithm; The edge smoothing sub-algorithm reduces the symbol edge jitter of the corresponding level of the Ref signal to the corresponding adaptation range of the symbol edge jitter value; Based on the reflection level, the overall amplitude and phase of the Ref signal after each level calibration are corrected twice to generate the Ref0 signal that dynamically matches the interference characteristics of the Rx signal and the electrical link reflection level. The Ref0 signal is a standard random code signal after dynamic calibration. The bit error rate calculation unit is used to compare the Rx signal and the Ref0 signal bit by bit and level by level. Based on the bit error sensitivity weight of each PAM4 level, it calculates the number of inconsistencies N1 after weighting and combines it with the total number of bits N2 to calculate the bit error rate.

[0069] In this embodiment, the bit error rate is calculated as follows: ,in, γ1 is the error identifier for the u-th symbol and the l-th level, with the error being 1 or 0, and BER being the bit error rate; γ1 is the error sensitivity weight for the l-th PAM4 level. The number of inconsistent symbols after weighting is N1.

[0070] In this embodiment, the Rx signal refers to the random code signal generated by the high-speed switching chip, which, after being transmitted through the differentiated electrical link module, is distorted due to the module's reflection effect and is ultimately received by the optical module under test. For example, the PRBS31, 100Gbps PAM4 random code output by the high-speed switching chip, after passing through an electrical link module with a reflection intensity level of 2 (corresponding to a reflection coefficient of -25dB), experiences amplitude attenuation and timing shift, forming an Rx signal carrying reflection interference. Specifically, this signal is received through the QSFP28 electrical interface of the optical module under test. The interface integrates a signal buffer circuit to ensure that the Rx signal is stably input to the feature extraction unit.

[0071] In this embodiment, the multidimensional characteristic parameters are a set of key parameters that characterize the degree of reflection interference in the Rx signal. Specifically, they include: the timing offset corresponding to each PAM4 level, the amplitude attenuation coefficient corresponding to each PAM4 level, the symbol edge jitter value corresponding to each PAM4 level, and the reflection intensity level of the corresponding electrical link module. The PAM4 levels are the four signal levels in PAM4 modulation technology (usually denoted as 0, 1, 2, and 3 levels), where 0 corresponds to 0V, 1 corresponds to 1V, 2 corresponds to 2V, and 3 corresponds to 3V, used to carry high-speed data transmission. The timing offset corresponding to each PAM4 level refers to the deviation between the actual trigger time and the standard trigger time of a certain PAM4 level symbol in the Rx signal. For example, the standard trigger time of the 0 level symbol is 100ns, but due to reflection interference, it is actually delayed to 102ns. In this case, the timing offset of the 0 level is +2ns. Specifically, the timing offset value of each level symbol is calculated by acquiring the Rx signal waveform with an oscilloscope, comparing it with the standard timing signal output by the high-speed switching chip.

[0072] The amplitude attenuation coefficient is the ratio (value 0-1) of the actual amplitude of a certain PAM4 level in the Rx signal to the standard amplitude. For example, the standard amplitude of the 3 level is 3V, and after electrical link reflection interference, the actual amplitude of the 3 level in the Rx signal is 2.7V. At this time, the amplitude attenuation coefficient of the 3 level is 0.9. Specifically, the amplitude attenuation coefficient of each PAM4 level is calculated by measuring the S21 parameter (transmission coefficient) of the electrical link module through a vector network analyzer and combining the correspondence between the S21 parameter and amplitude attenuation.

[0073] The symbol edge jitter value refers to the time fluctuation range of the rising or falling edge of a certain PAM4 level symbol in the Rx signal. For example, the rising edge of a 1 level symbol is sometimes triggered at 5.0ns and sometimes at 5.2ns. In this case, the symbol edge jitter value of the 1 level symbol is 0.2ns. Specifically, the edge signal of each PAM4 level symbol is collected by a jitter analyzer, and the maximum fluctuation value at the edge trigger time is counted.

[0074] The reflection level is a classification indicator of the degree of reflection interference from the electrical link module. It is usually divided into three levels: level 1 is low reflection, level 2 is medium reflection, and level 3 is high reflection. For example, a reflection coefficient of -30dB corresponds to level 1, -25dB corresponds to level 2, and -20dB corresponds to level 3. Specifically, the S11 parameter (reflection coefficient) of the electrical link module is measured by the vector network analyzer of the MCB calibration module. According to the preset correspondence between the S11 parameter and the reflection intensity level, the reflection intensity level of the currently accessed electrical link module is determined. For example, the reflection coefficient is between -30dB and -20dB, with a gradient interval of 5dB.

[0075] The Ref signal refers to the standard PAM4 random code signal of the high-speed switching chip that has not been subjected to any electrical link interference and conforms to the preset configuration. This signal is obtained from the dedicated test output interface of the high-speed switching chip.

[0076] In this embodiment, the timing compensation sub-algorithm is used to adjust the triggering time of the corresponding PAM4 level symbol in the Ref signal. The core is to align the timing of the Ref signal with the timing of the Rx signal. For example, if the timing offset of the 2 level in the Rx signal is -1ns (triggering 1ns earlier), then the algorithm advances the triggering time of the 2 level symbol in the Ref signal by 1ns. The implementation method is to integrate a delay line circuit inside the dynamic calibration unit and control the delay time of the delay line according to the timing offset to achieve timing compensation of the Ref signal.

[0077] In this embodiment, the amplitude correction sub-algorithm is used to adjust the amplitude of the corresponding PAM4 level symbol in the Ref signal. The core is to match the amplitude of the Ref signal with the amplitude of the Rx signal. For example, if the amplitude attenuation coefficient of the 1 level in the Rx signal is 0.8 (actual amplitude 0.8V), then the algorithm will scale the amplitude of the 1 level in the Ref signal from 1V to 0.8V. The implementation method is to integrate a variable gain amplifier inside the dynamic calibration unit and adjust the gain of the amplifier according to the amplitude attenuation coefficient to achieve amplitude correction of the Ref signal.

[0078] In this embodiment, the edge smoothing sub-algorithm is used to reduce the edge jitter of the corresponding PAM4 level symbols in the Ref signal. The core is to control the edge jitter of the Ref signal within the same range as the Rx signal. For example, if the edge jitter of the 0 level symbol in the Rx signal is 0.15ns, then the algorithm can reduce the edge jitter of the 0 level symbol in the Ref signal to less than 0.15ns. The implementation method is to integrate a level edge shaping circuit inside the dynamic calibration unit to shape the symbol edges of the Ref signal and reduce jitter.

[0079] In this embodiment, the secondary correction of overall amplitude and phase refers to the overall amplitude fine-tuning and phase adjustment of the Ref signal after completing the timing, amplitude, and jitter calibration of each PAM4 level, combined with the reflection intensity level of the electrical link module. For example, when the reflection intensity level is level 3 (high reflection), the overall amplitude after each level calibration needs to be increased by an additional 5% to match the overall attenuation trend of the Rx signal, and the phase needs to be fine-tuned by 1° to compensate for the phase shift caused by reflection. The implementation method is to use the phase adjuster and gain controller inside the dynamic calibration unit to query the preset correction parameter table according to the reflection intensity level and perform secondary correction.

[0080] In this embodiment, the Ref0 signal refers to a standard random code signal whose timing, amplitude, jitter, and phase all match the interference characteristics of the Rx signal after dynamic hierarchical calibration and secondary correction. For example, the timing offset, amplitude attenuation, and edge jitter of each PAM4 level in the Ref0 signal are consistent with those of the Rx signal. The implementation method is to store the corrected signal in the high-speed cache of the dynamic calibration unit. The cache is linked with the bit error rate calculation unit through a synchronous clock to ensure that the Ref0 signal is stably output for comparison.

[0081] In this embodiment, the symbol-by-symbol and level-by-level comparison refers to comparing the Rx signal and the Ref0 signal one by one in symbol order, and judging whether the PAM4 level value of each symbol is consistent. For example, when comparing the 100th symbol, it is necessary to judge whether the 0 level value of the 100th symbol of the Rx signal is consistent with the corresponding level value of the Ref0 signal and whether the 1 level value is consistent (if the symbol is actually at level 1, then the consistency of level 1 is the focus). The bit error rate calculation unit integrates a hardware comparator, with one comparator corresponding to each PAM4 level, and outputs the level consistency judgment result one by one, that is, 0 for consistency and 1 for inconsistency.

[0082] The error sensitivity weights for each PAM4 level are determined based on the anti-interference capability of each PAM4 signal level. The error impact weights are set for each level. The middle level usually has a higher weight than the edge level because the level interval is small and the anti-interference capability is weak. For example, the preset weights are 0.8 for level 0, 1.2 for level 1, 1.2 for level 2, and 0.8 for level 3. This means that when a bit error occurs at level 1 or level 2, the impact on the overall bit error rate is greater. Specifically, the weight values ​​are pre-stored in the Flash memory of the oDSP chip of the optical module under test and directly called when calculating the bit error rate.

[0083] The beneficial effects of the above technical solution are as follows: It captures the multi-dimensional reflection interference characteristics of the Rx signal; the dynamic calibration unit performs layered calibration and secondary correction for each level of PAM4, ensuring a high degree of matching between the interference characteristics of the Ref0 signal and the Rx signal; and the bit error rate calculation unit combines the level error sensitivity weights to statistically calculate the weighted number of bit errors, ultimately achieving accurate bit error rate calculation. This significantly improves the accuracy of determining the performance of the optical module's electrical interface under different reflection scenarios, providing a reliable quantitative basis for optimizing the anti-interference capability of the optical module.

[0084] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A device for testing the performance of a high-speed optical module electrical interface, characterized in that, include: The host computer is used to issue control commands. A high-speed switching chip, which is connected to the host computer, is used to generate random code signals according to the control commands. Several differentiated electrical link modules are provided. The input of each electrical link module is connected to the signal output of the high-speed switching chip. It is used to receive random code signals and apply reflection interference to realize serial testing of environmental simulation. The output of each electrical link module is connected to the optical module under test. It is used to transmit random code signals carrying reflection interference to the optical module under test. The insertion loss of each electrical link module is consistent, the reflection intensity increases in a gradient, and each electrical link module corresponds to an application scenario with different reflection intensities. The optical module under test is used to compare the random code signal carrying reflection interference with the standard random code based on the internal oDSP chip, generate the bit error rate, and report it to the host computer. The host computer is used to analyze the bit error rate, determine the comparison results and judgment results, and visualize them. The MCB calibration module is detachably connected to each electrical link module and is used to pre-calibrate the insertion loss and reflection intensity of each electrical link module.

2. The apparatus for testing the performance of a high-speed optical module electrical interface according to claim 1, characterized in that, The high-speed switching chip is either a Broadcom Strata DNX series chip or a Cisco Silicon One series chip.

3. The apparatus for testing the performance of a high-speed optical module electrical interface according to claim 1, characterized in that, The MCB calibration module includes a vector network analyzer, which is used to collect the S-parameters of each electrical link module and adjust the attenuator parameters of the electrical link modules to achieve insertion loss and reflection intensity calibration.

4. The apparatus for testing the performance of a high-speed optical module electrical interface according to claim 1, characterized in that, The optical module under test is an optical module that supports PAM4 signal transmission.

5. The apparatus for testing the performance of a high-speed optical module electrical interface according to claim 1, characterized in that, The high-speed switching chip includes: The parameter extraction unit is used to extract the random code generation parameters of the control command and store the verification instructions; The determination unit is used to determine whether the random code generation parameters are consistent with the random code generation configuration content pre-stored inside the high-speed switching chip according to the storage verification instruction. If they are inconsistent, the difference information between the random code generation parameters and the pre-stored random code generation configuration content will be used. A determining unit is configured to determine the number of different types of differences in the random code generation parameters and the difference symbol for each difference type based on the difference information. The recoding unit is used to extract the important and unimportant bits occupied by the difference symbols under each difference type, determine the temporal distribution parameters of the corresponding difference symbols in the difference information and the reflection intensity level factor of the current electrical link module to be accessed, and recode the corresponding difference symbols according to the type attribute of each difference type to generate a random code signal carrying interference features corresponding to the difference information. If they match, a random code signal is directly generated based on the pre-stored random code generation configuration.

6. The apparatus for testing the performance of a high-speed optical module electrical interface according to claim 5, characterized in that, The host computer includes: The combined effect of the reflected interference of the electrical link module and the interference characteristics of the random code signal is determined, and the initial degree of influence is obtained by matching according to the influence-degree lookup table; If the random code signal is directly generated, the initial degree of influence shall be regarded as the current degree of influence; If the random code signal is generated by recoding, determine the degree of inconsistency of the difference information and adjust the initial degree of influence to obtain the current degree of influence; Based on the current level of influence, the visualization results of scene deviation interference are optimized.

7. The apparatus for testing the performance of a high-speed optical module electrical interface according to claim 1, characterized in that, The parameter extraction unit includes: The syntax detection subunit is used to perform syntax parsing on the control command issued by the host computer before extracting the random code generation parameters of the control command. When a syntax error is detected in the control command, the command configuration metafunction and object mapping matrix are generated based on the design architecture of the high-speed switching chip. An error-locking subunit is used to lock the location of a syntax error and capture the context information of the syntax error location; The command correction subunit is used to generate initial completion parameters based on the completion rules in the command configuration metafunction according to the syntax error type, and adjust the initial completion parameters based on the current hardware operating status parameters of the high-speed switching chip. The adjusted completion parameters are then verified for cross-functional consistency based on the object mapping matrix. After successful verification, context information is embedded to obtain a corrected command that meets the requirements of syntax, semantics, and hardware execution feasibility. The command configuration metafunction generates completion parameters that match the hardware execution logic for the missing or erroneous fields in the context information according to the parameter interface specification of the high-speed switching chip. The object mapping matrix is ​​used to verify the cross-functional matching of the corrected parameters. Extract sub-units, used to extract the random code generation parameters of the corrected command; When no syntax errors are detected in the control command, the random code generation parameters of the control command are extracted.

8. The apparatus for testing the performance of a high-speed optical module electrical interface according to claim 1, characterized in that, The optical module under test includes: The feature extraction unit is used to denote the random code signal carrying reflection interference as the Rx signal, and extract multi-dimensional feature parameters associated with the reflection interference from the Rx signal. The multi-dimensional feature parameters include: the timing offset corresponding to each PAM4 level, the amplitude attenuation coefficient corresponding to each PAM4 level, the symbol edge jitter value corresponding to each PAM4 level, and the reflection level of the corresponding electrical link module. The dynamic calibration unit is used to acquire the original PAM4 standard random code generated by the high-speed switching chip, denoted as the Ref signal, and to perform dynamic hierarchical calibration on the Ref signal based on multi-dimensional feature parameters. For each PAM4 level, the timing compensation sub-algorithm is used to offset the symbol trigger time of the corresponding level of the Ref signal according to the timing offset. The amplitude of the symbol at the corresponding level of the Ref signal is scaled according to the amplitude attenuation coefficient using the amplitude correction sub-algorithm; The edge smoothing sub-algorithm reduces the symbol edge jitter of the corresponding level of the Ref signal to the corresponding adaptation range of the symbol edge jitter value; Based on the reflection level, the overall amplitude and phase of the Ref signal after each level calibration are corrected twice to generate the Ref0 signal that dynamically matches the interference characteristics of the Rx signal and the electrical link reflection level. The Ref0 signal is a standard random code signal after dynamic calibration. The bit error rate calculation unit is used to compare the Rx signal and the Ref0 signal bit by bit and level by level. Based on the bit error sensitivity weight of each PAM4 level, it calculates the number of inconsistencies N1 after weighting and combines it with the total number of bits N2 to calculate the bit error rate.

Images