Multi-channel amplitude phase correction system and method based on radio frequency chip
By using PL pure logic to drive the AD9361 chip and the automated initialization mechanism controlled by the state machine, combined with parallel register configuration, clock multiplication, zero-crossing detection and digital logic operation, the amplitude and phase inconsistency problem of multi-channel RF systems is solved, achieving high-precision, low-latency amplitude and phase consistency correction, which is suitable for scenarios such as 5G communication and phased array radar.
Patent Information
- Application Number
- CN202511880802.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-13
- Publication Date
- 2026-03-10
AI Technical Summary
Amplitude-phase inconsistency exists in multi-channel radio frequency systems. Existing technologies struggle to achieve high-precision, low-latency, and fully closed-loop amplitude-phase consistency correction, hindering the development of high-performance applications such as 5G communication and millimeter-wave radar.
The AD9361 chip is driven by PL pure logic. Through the automated initialization mechanism controlled by state machine, combined with parallel register configuration, clock frequency multiplication, zero-crossing detection and digital logic operation, multiple measurement averaging algorithm and complex plane rotation technology, multi-channel amplitude and phase consistency correction is achieved.
It achieves high-precision phase measurement (0.1 degree accuracy) for multi-channel RF systems, shortens phase correction response time, and uses a 2048-point sliding window and adaptive gain mapping table to achieve microsecond-level dynamic amplitude tracking. It is suitable for scenarios such as 5G communication and phased array radar, and improves system debugging efficiency and maintainability.
Smart Images

Figure CN121643941A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of radio frequency communication technology, and more specifically, to a multi-channel amplitude and phase correction system and method based on radio frequency chips. Background Technology
[0002] Currently, in multi-channel RF systems, due to the inherent differences in the hardware characteristics of each channel, including inconsistent filter group delays, amplifier gain deviations, and mixer phase shifts, there are significant amplitude and phase inconsistencies between channels. These inconsistencies can severely degrade system performance. For example, in MIMO systems, they can lead to a decrease in spatial multiplexing gain, and in beamforming applications, they can cause pointing deviations and sidelobe rises. Existing technologies often employ software-based digital predistortion (DPD) technology or application-specific integrated circuit (ASIC)-based correction schemes. Software solutions typically rely on a processor (such as an ARM core) to run the correction algorithm and configure the RF chip registers through a serial interface. While this method offers high flexibility, it suffers from inherent drawbacks such as poor real-time performance and large processing delays, making it difficult to meet the real-time correction requirements of high-speed communication scenarios.
[0003] Another common approach is to use analog circuits, such as using variable gain amplifiers (VGA) and phase shifters for correction. However, this method suffers from severe temperature drift, poor consistency and repeatability, and the hardware complexity increases exponentially with the number of channels, which is not conducive to system integration and cost control.
[0004] The existing technology has the following prominent problems: Firstly, at the system architecture level, the traditional PS software control mode introduces a large communication delay due to the need for data interaction through the operating system layer and the driver layer, making it difficult to achieve microsecond-level real-time response. Especially in multi-channel collaborative work, this delay will cause mismatch in the correction timing between channels, which will seriously affect the correction accuracy. Secondly, in terms of phase detection, conventional zero-crossing detection methods are susceptible to noise interference and false triggering, and lack a high-precision timing mechanism, resulting in large phase measurement errors. While digital phase detection methods based on correlation functions have better noise resistance, they have high computational complexity and require a lot of logic resources, making them unsuitable for resource-constrained embedded scenarios. Third, in the phase correction stage, most schemes use direct phase addition and subtraction operations without considering the mathematical nature of complex plane rotation, resulting in limited computational accuracy; and lack an efficient rotation factor generation mechanism, often implemented through lookup tables (LUTs), but there is an inherent contradiction between LUT capacity and accuracy, making it difficult to balance resource efficiency and correction quality. Fourth, in terms of amplitude control, traditional automatic gain control (AGC) circuits have slow response speed, limited dynamic range, and cannot achieve coordinated optimization with phase correction; digital AGC schemes mostly adopt fixed step size iterative adjustment strategies, which have slow convergence speed and are prone to overshoot or oscillation when the signal amplitude changes rapidly.
[0005] These technologies severely restrict the further development of multi-channel radio frequency systems in high-performance applications such as 5G communication and millimeter-wave radar, and there is an urgent need for a high-precision, low-latency, and fully closed-loop amplitude and phase consistency correction solution. Summary of the Invention
[0006] The purpose of this invention is to provide a multi-channel amplitude and phase correction system and method based on radio frequency chips, which solves the long-standing problem of amplitude and phase inconsistency in multi-channel radio frequency systems. By driving the AD9361 chip with pure logic using PL and adopting an automated initialization mechanism controlled by a state machine, the precise setting of 1024 registers can be completed in a single configuration, aiming to solve the problems in the prior art.
[0007] This invention is implemented as follows: a multi-channel amplitude and phase correction method based on an RF chip, applied to a phase correction device, specifically includes the following steps: S101: Receives the RF chip initialization command from the terminal, generates an executable initialization sequence based on the RF chip configuration parameters, and completes the full initialization of the RF chip's working mode, clock network, and interface parameters through a parallelized register configuration mechanism, establishing a stable basic environment for RF signal transmission and reception. S102: Based on the standard test signal stored in the embedded memory, it realizes the synchronous transmission and reception of multi-channel signals through a high-speed interface, uses the built-in signal chain of the RF chip to complete signal modulation, transmission and demodulation, and feeds back the multi-channel raw data containing amplitude and phase deviation to the phase correction device through a loopback architecture. S103: The phase correction device improves the system's time resolution by clock frequency multiplication, and uses a combination of zero-crossing detection and digital logic operation to convert the analog phase difference into a time-capable digital signal. Combined with a multi-measurement averaging algorithm, it effectively suppresses random errors and accurately obtains the phase difference parameters between multiple channels. S104: The phase difference parameter is converted into a complex plane rotation factor using a digital signal processing kernel. The phase rotation correction of the input signal is achieved through orthogonal signal synthesis and complex multiplication, thus completing the digital compensation of the phase deviation. S105: After outputting the digital compensation signal, the sliding window power detection mechanism is used to calculate the signal amplitude characteristics of the RF chip in real time. Combined with the pre-stored gain mapping table, adaptive correction coefficients are generated. Amplitude equalization is achieved through precise timing alignment. The amplitude and phase consistency of the RF chip under different scenarios is verified through multi-standard test signals, and verification data is output.
[0008] Furthermore, in S101, an executable initialization sequence based on the RF chip configuration parameters is generated. This sequence completes the comprehensive initialization of the RF chip's operating mode, clock network, and interface parameters through a parallelized register configuration mechanism, including: Register configuration is completed by controlling the SPI communication interface through a state machine. The state machine includes an idle state, a command assembly state, a data transmission state, and a state verification state. In the command assembly state, the configuration parameters are combined into a 24-bit command word, which includes a 3-bit opcode, a 10-bit address field, and an 8-bit data field. In the data transmission state, the command word is output to the RF chip bit by bit through the clock synchronization mechanism, and the response flag bit is read to verify the configuration result in the status verification state. For the 1024 configuration registers of the RF chip, a paging addressing mechanism is used to achieve complete coverage. A retry mechanism is set for the configuration of key parameters. When the verification fails, the configuration command is automatically resent until it succeeds.
[0009] Furthermore, in S102, based on a standard test signal preset and stored in the embedded memory, synchronous transmission and reception of multi-channel signals are achieved through a high-speed interface, including: The data interface adopts DDR double data rate mode, and data frame synchronization signal pulses are sampled on both the rising and falling edges of the clock. The I / Q data of the first channel is transmitted during the rising edge phase, and the I / Q data of the second channel is transmitted during the falling edge phase. The data format is a 12-bit signed number, and the I and Q data are transmitted on the same physical link using time-division multiplexing. The receiving end extracts the clock signal from the LVDS differential pair through the data recovery unit, and converts the serial data stream into parallel data through the serial-to-parallel conversion module for use by subsequent units.
[0010] Furthermore, in S103, the phase correction device improves the system's time resolution by clock multiplication and uses a combination of zero-crossing detection and digital logic operations to convert the analog phase difference into a timeable digital signal, including: The clock frequency is multiplied from 30.72MHz to 409.6MHz via a phase-locked loop; The zero-crossing detection circuit includes a signal shaping filter and a hysteresis comparator. The filter is a fourth-order Butterworth low-pass filter with a cutoff frequency set to 1.5 times the signal frequency. The hysteresis comparator is set with a 5mV hysteresis voltage to prevent false triggering caused by noise. Digital logic operations include XOR gates and precision counters. The XOR gates perform logic operations on two square wave signals to output a pulse-width modulated (PWM) signal. The precision counter, driven by a frequency-multiplied clock, counts the pulse width. The counting result is expressed by the formula Δφ=(N×T)×360° / (1 / f). c Calculate the phase difference, where N is the count value, T is the frequency multiplication clock period, and f c This is the signal carrier frequency.
[0011] Furthermore, by combining multiple measurement averaging algorithms to effectively suppress random errors, the phase difference parameters between multiple channels are accurately obtained, including: A continuous sampling mechanism is established. Within the set measurement time window, the phase difference signal of the two signals is sampled at least 1024 times continuously with a high-frequency clock of 409.6MHz as the reference. Each sampling process includes a complete zero-crossing detection, XOR processing and clock counting process. The obtained raw phase difference data is temporarily stored in a first-in-first-out buffer queue with a depth of 1024, which ensures that the phase change characteristics of the signal in different periods can be covered. Data filtering and averaging were performed. The 1024 phase difference measurements in the cache queue were sorted and analyzed. The median absolute deviation algorithm was used to identify and remove abnormal measurement points caused by transient interference. The remaining valid data were divided into 8 groups according to time order, and the arithmetic mean of each group was calculated. The arithmetic mean of each group is weighted and averaged, with the group with the most recent time period given a higher weight. The output phase difference parameter also includes a confidence index, which reflects the reliability of the measurement results.
[0012] Furthermore, in S104, the phase difference parameter is converted into a complex plane rotation factor using a digital signal processing kernel, including: The phase difference parameter is normalized and mapped to a standard phase range of 0-360 degrees. Then, the normalized phase value is quantized into a 14-bit precision phase control word through phase control word conversion, where the high 14 bits are used as the effective phase input parameter of the DDS core. Driven by a 40MHz operating clock, the DDS core generates corresponding quadrature signals in real time based on the phase accumulator principle. By querying the pre-stored sine / cosine waveform table, it outputs 16-bit precision cosθ and sinθ digital signals to form the real and imaginary parts of the complex plane rotation factor. The input IQ data and the rotation factor are fed into the complex multiplication unit, which consists of four parallel real multipliers that calculate the four product terms I1×cosθ, Q1×sinθ, I1×sinθ, and Q1×cosθ, respectively. Two carry-lookahead adders then perform the operations of merging the real parts I1×cosθ-Q1×sinθ and merging the imaginary parts I1×sinθ+Q1×cosθ, completing the complex multiplication in a single clock cycle. The output data undergoes dynamic range detection and saturation processing, and is automatically limited when the result exceeds the 16-bit representation range to maintain the accurate amplitude and phase characteristics of the output signal.
[0013] Furthermore, in S105, a sliding window power detection mechanism is used to calculate the signal amplitude characteristics of the RF chip in real time, and adaptive correction coefficients are generated by combining them with a pre-stored gain mapping table, including: A sliding window is constructed using a 2048-point shift register array. In each clock cycle, the absolute value of the newly sampled signal is shifted into the beginning of the register, while the earliest stored absolute value is shifted out from the end. The power calculation maintains the dynamic cumulative sum in real time. The power sum is updated by adding the newly shifted-in absolute value to the current cumulative value and subtracting the shifted-out absolute value. The power is shifted right by 11 bits to obtain the average power value of 2048 points, which is used as the index address for looking up the gain mapping table. The gain mapping table adopts a piecewise linear interpolation structure, dividing the 20-bit gain value into 512 intervals according to the power range. When the average power value falls between two storage points, the accurate gain coefficient is calculated by linear interpolation. Timing alignment ensures precise synchronization between the input signal and the gain coefficient obtained from the lookup table through a 2149-stage pipeline register. Based on changes in the actual working environment, the gain mapping relationship is continuously optimized through background learning. When a significant change in channel conditions is detected, the gain table update process is automatically started. The gain mapping curve is recalculated based on the latest acquired signal characteristics. The update adopts a smooth transition to avoid signal distortion caused by sudden gain changes. Historical mapping data is retained for performance comparison and analysis. The timing alignment adopts a multi-stage pipeline structure, which includes precise compensation allocation for signal absolute value calculation delay, sliding window processing delay, table lookup delay, and multiplier delay. The allocation is as follows: signal preprocessing occupies 5 clock cycles, sliding window power calculation occupies 2048 clock cycles, gain table lookup occupies 64 clock cycles, and multiplication operation occupies 32 clock cycles. This ensures that each data sample can be accurately matched with the corresponding gain coefficient, achieving a jitter-free amplitude correction effect.
[0014] Furthermore, the amplitude and phase consistency of the RF chip under different scenarios is verified using multi-standard test signals, and verification data is output, including: A hierarchical verification system was established, and basic verification was performed using single-tone signals. The basic functions of the system were confirmed by observing the time-domain waveform and spectral characteristics. Subsequently, a 10.6MHz narrowband QPSK signal was used to verify the low-frequency performance, with a focus on detecting the convergence and phase noise characteristics of the constellation diagram. High-frequency stability was verified using 18MHz and 26MHz wideband QPSK signals, and in-band flatness and group delay consistency were analyzed. For each test signal, the -3dB bandwidth and in-band ripple were measured using a spectrum analyzer, and the error vector amplitude, IQ imbalance and constellation diagram symmetry were obtained using a vector signal analyzer. Test data is automatically recorded and a comprehensive performance report is generated, including amplitude consistency error, phase consistency error, and system stability coefficient. When any indicator exceeds the threshold, a recalibration process is automatically triggered.
[0015] Compared with the prior art, the multi-channel amplitude and phase correction system and method based on radio frequency chips provided by the present invention have the following advantages: 1. The long-standing amplitude-phase inconsistency problem in multi-channel RF systems has been solved. By driving the AD9361 chip with pure logic using PL and adopting an automated initialization mechanism controlled by a state machine, the precise setting of 1024 registers can be completed in a single configuration. The digital phase detection scheme multiplies the system clock to 409.6MHz. Combined with zero-crossing detection and multiple measurement averaging algorithms, the phase measurement accuracy reaches 0.1 degrees, effectively overcoming measurement errors caused by noise interference. The complex plane phase shifting technology generates a rotation factor through DDS verification, avoiding the capacity and accuracy contradiction of the traditional lookup table method. The phase correction response time is shortened to a single clock cycle. The digital AGC adopts a 2048-point sliding window and an adaptive gain mapping table to achieve microsecond-level dynamic amplitude tracking. 2. A dual-stage verification mechanism combining single-tone signal calibration and multi-bandwidth QPSK verification covers the entire operating bandwidth from 200kHz to 56MHz, demonstrating excellent adaptability in scenarios such as 5G communication and phased array radar. The integration of an ILA online logic analyzer and a VIO virtual debugging interface improves system debugging efficiency by over 60%. Furthermore, remote monitoring enables intelligent operation and maintenance of the equipment. SPI serial-to-parallel conversion and multi-master arbitration mechanisms ensure optimized allocation of system resources, supporting parallel operation of multiple processing units. The timing alignment unit employs a 2149-level precise timing pipeline, completely eliminating signal jitter during the calibration process. This solution maintains high precision while possessing excellent engineering practicality and maintainability, providing reliable amplitude and phase consistency guarantees for wireless communication equipment.
[0016] A multi-channel amplitude and phase correction system based on an RF chip is used to perform the above-described multi-channel amplitude and phase correction method. The system includes: The configuration initialization module is used to receive terminal instructions and generate an executable initialization sequence for the RF chip. It completes register configuration through the SPI interface controlled by the state machine and establishes a stable basic environment for RF signal transmission and reception. The signal transceiver processing module is used to read standard test signals from embedded memory, realize synchronous transmission and reception of multi-channel signals through LVDS interface, complete signal modulation and demodulation using the built-in signal chain of RF chip, and feed back the raw data containing amplitude and phase deviation to the processing system through loopback architecture. The phase detection module is used to improve the system's time resolution by multiplying the frequency through a phase-locked loop. It uses a combination of zero-crossing detection and digital logic operations to convert the analog phase difference into a digital signal, and combines multiple measurement averaging algorithms to accurately obtain the phase difference parameters between multiple channels. The phase correction module is used to convert the phase difference parameter into a complex plane rotation factor through the DDS digital signal processing core, and to perform phase rotation correction of the input signal by using complex multiplication operations, thereby completing the digital compensation of phase deviation. The amplitude correction module is used to calculate the signal amplitude characteristics in real time using a sliding window power detection mechanism, generate adaptive correction coefficients by combining a pre-stored gain mapping table, and achieve time-aligned amplitude equalization processing through a multi-stage pipeline. The system verification module is used to verify the amplitude and phase consistency of the system under different scenarios through multi-standard test signals, and automatically generates a comprehensive performance report including amplitude consistency error, phase consistency error and system stability coefficient.
[0017] Specifically, the phase detection module includes: The clock multiplier unit is used to multiply the reference clock from 30.72MHz to 409.6MHz via a phase-locked loop; The zero-crossing detection unit, which includes a signal shaping filter and a hysteresis comparator, is used to convert the input signal into a square wave signal; The digital logic unit, which includes an XOR gate and a precision counter, is used to perform logic operations on two square wave signals and output pulse width modulation signals. The data processing unit performs continuous sampling and data filtering, suppresses random errors through a multi-measurement averaging algorithm, and outputs accurate phase difference parameters with confidence indices. Attached Figure Description
[0018] Figure 1 This is a flowchart illustrating the multi-channel amplitude and phase correction method based on an RF chip proposed in this invention. Figure 2 This is a flowchart illustrating the process of fully initializing the operating mode, clock network, and interface parameters of the RF chip in the multi-channel amplitude and phase correction method based on the RF chip proposed in this invention. Figure 3 This is a flowchart of the serial-to-parallel conversion procedure in the multi-channel amplitude and phase correction method based on radio frequency chips proposed in this invention; Figure 4 This is a schematic diagram of the structure of the multi-channel amplitude and phase correction system based on an RF chip proposed in this invention; Figure 5 This is a schematic diagram of the phase detection module in the multi-channel amplitude and phase correction system based on an RF chip proposed in this invention. Detailed Implementation
[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
[0020] The implementation of the present invention will be described in detail below with reference to specific embodiments.
[0021] In the accompanying drawings of this embodiment, the same or similar reference numerals correspond to the same or similar components. In the description of this invention, it should be understood that if terms such as "upper," "lower," "left," and "right" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting this invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.
[0022] Reference Figure 1-3 As shown, a multi-channel amplitude and phase correction method based on an RF chip, applied to a phase correction device, specifically includes the following steps: S101: Receives the RF chip initialization command from the terminal, generates an executable initialization sequence based on the RF chip configuration parameters, and completes the full initialization of the RF chip's working mode, clock network, and interface parameters through a parallelized register configuration mechanism, establishing a stable basic environment for RF signal transmission and reception. This includes generating an executable initialization sequence based on the RF chip configuration parameters, and completing the comprehensive initialization of the RF chip's operating mode, clock network, and interface parameters through a parallelized register configuration mechanism, including: Register configuration is completed by controlling the SPI communication interface through a state machine. The state machine includes an idle state, a command assembly state, a data transmission state, and a state verification state. In the command assembly state, the configuration parameters are combined into a 24-bit command word, which includes a 3-bit opcode, a 10-bit address field, and an 8-bit data field. In the data transmission state, the command word is output to the RF chip bit by bit through the clock synchronization mechanism, and the response flag bit is read to verify the configuration result in the status verification state. For the 1024 configuration registers of the RF chip, a paging addressing mechanism is used to achieve complete coverage, and a retry mechanism is set for the configuration of key parameters. When the verification fails, the configuration command is automatically resent until it succeeds. S102: Based on the standard test signal stored in the embedded memory, it realizes the synchronous transmission and reception of multi-channel signals through a high-speed interface, uses the built-in signal chain of the RF chip to complete signal modulation, transmission and demodulation, and feeds back the multi-channel raw data containing amplitude and phase deviation to the phase correction device through a loopback architecture. Among them, based on standard test signals preset and stored in embedded memory, synchronous transmission and reception of multi-channel signals are achieved through a high-speed interface, including: The data interface adopts DDR double data rate mode, and data frame synchronization signal pulses are sampled on both the rising and falling edges of the clock. The I / Q data of the first channel is transmitted during the rising edge phase, and the I / Q data of the second channel is transmitted during the falling edge phase. The data format is a 12-bit signed number, and the I and Q data are transmitted on the same physical link using time-division multiplexing. The receiving end extracts the clock signal from the LVDS differential pair through the data recovery unit, and converts the serial data stream into parallel data through the serial-to-parallel conversion module for use by subsequent units. Reference Figure 3 The serial-to-parallel conversion module operates as follows: When the FPGA reads from the AD9361, it outputs the address to be read to the AD9361 via the command register through spi_sdo. Then, the AD9361 sends the data at this address to the FPGA's SPI interface via spi_sdi through SPI_DO, and stores it in an 8-bit read_data through a shift register. When waitrequest is low, these eight bits of data are passed to other modules in parallel. When the FPGA writes to the AD9361, it passes the address and writedata to the AD9361.
[0023] S103: The phase correction device improves the system's time resolution by clock frequency multiplication, and uses a combination of zero-crossing detection and digital logic operation to convert the analog phase difference into a time-capable digital signal. Combined with a multi-measurement averaging algorithm, it effectively suppresses random errors and accurately obtains the phase difference parameters between multiple channels. The phase correction device improves the system's time resolution by multiplying the clock frequency and uses a combination of zero-crossing detection and digital logic operations to convert the analog phase difference into a timeable digital signal, including: The clock frequency is multiplied from 30.72MHz to 409.6MHz via a phase-locked loop; The zero-crossing detection circuit includes a signal shaping filter and a hysteresis comparator. The filter is a fourth-order Butterworth low-pass filter with a cutoff frequency set to 1.5 times the signal frequency. The hysteresis comparator is set with a 5mV hysteresis voltage to prevent false triggering caused by noise. Digital logic operations include XOR gates and precision counters. The XOR gates perform logic operations on two square wave signals to output a pulse-width modulated (PWM) signal. The precision counter, driven by a frequency-multiplied clock, counts the pulse width. The counting result is expressed by the formula Δφ=(N×T)×360° / (1 / f). c Calculate the phase difference, where N is the count value, T is the frequency multiplication clock period, and f c The signal carrier frequency; S104: The phase difference parameter is converted into a complex plane rotation factor using a digital signal processing kernel. The phase rotation correction of the input signal is achieved through orthogonal signal synthesis and complex multiplication, thus completing the digital compensation of the phase deviation. The process of converting the phase difference parameter into a complex plane rotation factor using a digital signal processing kernel includes: The phase difference parameter is normalized and mapped to a standard phase range of 0-360 degrees. Then, the normalized phase value is quantized into a 14-bit precision phase control word through phase control word conversion, where the high 14 bits are used as the effective phase input parameter of the DDS core. Driven by a 40MHz operating clock, the DDS core generates corresponding quadrature signals in real time based on the phase accumulator principle. By querying the pre-stored sine / cosine waveform table, it outputs 16-bit precision cosθ and sinθ digital signals to form the real and imaginary parts of the complex plane rotation factor. The input IQ data and the rotation factor are fed into the complex multiplication unit, which consists of four parallel real multipliers that calculate the four product terms I1×cosθ, Q1×sinθ, I1×sinθ, and Q1×cosθ respectively. Two carry-lookahead adders then perform the operations of merging the real parts I1×cosθ-Q1×sinθ and merging the imaginary parts I1×sinθ+Q1×cosθ. The complex multiplication is completed in a single clock cycle. The output data undergoes dynamic range detection and saturation processing. When the result exceeds the 16-bit representation range, it is automatically limited to maintain the accurate amplitude and phase characteristics of the output signal. S105: After outputting the digital compensation signal, the sliding window power detection mechanism is used to calculate the signal amplitude characteristics of the RF chip in real time. Combined with the pre-stored gain mapping table, adaptive correction coefficients are generated. Amplitude equalization is achieved through precise timing alignment. The amplitude and phase consistency of the RF chip under different scenarios is verified through multi-standard test signals, and verification data is output.
[0024] In S103 of this embodiment, the multiple measurement averaging algorithm effectively suppresses random errors and accurately obtains the phase difference parameters between multiple channels, including: A continuous sampling mechanism is established. Within the set measurement time window, the phase difference signal of the two signals is sampled at least 1024 times continuously with a high-frequency clock of 409.6MHz as the reference. Each sampling process includes a complete zero-crossing detection, XOR processing and clock counting process. The obtained raw phase difference data is temporarily stored in a first-in-first-out buffer queue with a depth of 1024, which ensures that the phase change characteristics of the signal in different periods can be covered. Data filtering and averaging were performed. The 1024 phase difference measurements in the cache queue were sorted and analyzed. The median absolute deviation algorithm was used to identify and remove abnormal measurement points caused by transient interference. The remaining valid data were divided into 8 groups according to time order, and the arithmetic mean of each group was calculated. The arithmetic mean of each group is weighted and averaged, with the group with the most recent time period given a higher weight. The output phase difference parameter also includes a confidence index, which reflects the reliability of the measurement results.
[0025] In S105 of this embodiment, a sliding window power detection mechanism is used to calculate the signal amplitude characteristics of the RF chip in real time, and adaptive correction coefficients are generated by combining the pre-stored gain mapping table, including: A sliding window is constructed using a 2048-point shift register array. In each clock cycle, the absolute value of the newly sampled signal is shifted into the beginning of the register, while the earliest stored absolute value is shifted out from the end. The power calculation maintains the dynamic cumulative sum in real time. The power sum is updated by adding the newly shifted-in absolute value to the current cumulative value and subtracting the shifted-out absolute value. The power is shifted right by 11 bits to obtain the average power value of 2048 points, which is used as the index address for looking up the gain mapping table. The gain mapping table adopts a piecewise linear interpolation structure, dividing the 20-bit gain value into 512 intervals according to the power range. When the average power value falls between two storage points, the accurate gain coefficient is calculated by linear interpolation. Timing alignment ensures precise synchronization between the input signal and the gain coefficient obtained from the lookup table through a 2149-stage pipeline register. Based on changes in the actual working environment, the gain mapping relationship is continuously optimized through background learning. When a significant change in channel conditions is detected, the gain table update process is automatically started. The gain mapping curve is recalculated based on the latest acquired signal characteristics. The update adopts a smooth transition to avoid signal distortion caused by sudden gain changes. Historical mapping data is retained for performance comparison and analysis. Timing alignment employs a multi-stage pipeline structure, with precise compensation allocations for signal absolute value calculation delay, sliding window processing delay, lookup table delay, and multiplier delay: signal preprocessing occupies 5 clock cycles, sliding window power calculation occupies 2048 clock cycles, gain lookup table occupies 64 clock cycles, and multiplication operation occupies 32 clock cycles, ensuring that each data sample can accurately match the corresponding gain coefficient, achieving a jitter-free amplitude correction effect.
[0026] In this embodiment, the amplitude and phase consistency of the RF chip under different scenarios is verified using multi-standard test signals, and verification data is output, including: A hierarchical verification system was established, and basic verification was performed using single-tone signals. The basic functions of the system were confirmed by observing the time-domain waveform and spectral characteristics. Subsequently, a 10.6MHz narrowband QPSK signal was used to verify the low-frequency performance, with a focus on detecting the convergence and phase noise characteristics of the constellation diagram. High-frequency stability was verified using 18MHz and 26MHz wideband QPSK signals, and in-band flatness and group delay consistency were analyzed. For each test signal, the -3dB bandwidth and in-band ripple were measured using a spectrum analyzer, and the error vector amplitude, IQ imbalance and constellation diagram symmetry were obtained using a vector signal analyzer. Test data is automatically recorded and a comprehensive performance report is generated, including amplitude consistency error, phase consistency error, and system stability coefficient. When any indicator exceeds the threshold, a recalibration process is automatically triggered.
[0027] The serial-to-parallel conversion module logic specifically includes: State 0: Detects if there is a read or write signal. If so, it passes the read / write signal, address, and data to command. Here, address and writedata are input signals, which are input from the Avalon_Mux module. State 1: Output the received instructions bit by bit to AD9361 via spi_sdo; State 2: If spi_sdo is not fully transmitted, return to state 1 to continue transmission; if transmission is complete, enter state 3. State 3: After the transmission is complete, set waitrequest (output) to zero, indicating that the read data (readdata) can be sent out. State 4: set waitrequest to 1, stop transmission, end.
[0028] This technical solution solves the long-standing problem of amplitude and phase inconsistency in multi-channel RF systems. By driving the AD9361 chip with pure logic using PL and employing an automated initialization mechanism controlled by a state machine, the precise setting of 1024 registers can be completed in a single configuration. The digital phase detection scheme multiplies the system clock to 409.6MHz and, combined with zero-crossing detection and multiple measurement averaging algorithms, achieves a phase measurement accuracy of 0.1 degrees, effectively overcoming measurement errors caused by noise interference. The complex plane phase shifting technology generates a rotation factor through DDS verification, avoiding the capacity and accuracy contradiction of the traditional lookup table method. The phase correction response time is shortened to a single clock cycle. The digital AGC uses a 2048-point sliding window and an adaptive gain mapping table to achieve microsecond-level dynamic amplitude tracking. Reference Figure 4-5 As shown, a multi-channel amplitude and phase correction system based on an RF chip is used to execute the aforementioned multi-channel amplitude and phase correction method. The system includes: a configuration initialization module, used to receive terminal instructions and generate an initialization sequence executable by the RF chip, complete register configuration through a state machine-controlled SPI interface, and establish a stable RF signal transmission and reception environment; a signal transmission and reception processing module, used to read standard test signals from embedded memory, realize synchronous transmission and reception of multi-channel signals through an LVDS interface, complete signal modulation and demodulation using the built-in signal chain of the RF chip, and feed back the raw data containing amplitude and phase deviations to the processing system through a loopback architecture; a phase detection module, used to improve the system's time resolution through phase-locked loop frequency multiplication, convert analog phase difference into digital signal using a combination of zero-crossing detection and digital logic operation, and accurately obtain the phase difference parameters between multiple channels using a multiple measurement averaging algorithm; and a phase correction module, used to correct the phase difference parameters between multiple channels through a DDS digital signal... The first processing core converts the phase difference parameter into a complex plane rotation factor, and uses complex multiplication to perform phase rotation correction on the input signal, thus completing digital compensation for phase deviation. The amplitude correction module uses a sliding window power detection mechanism to calculate the signal amplitude characteristics in real time, and generates adaptive correction coefficients by combining a pre-stored gain mapping table. It then performs timing-aligned amplitude equalization processing through a multi-stage pipeline. The system verification module verifies the amplitude and phase consistency of the system under different scenarios using multi-standard test signals, and automatically generates a comprehensive performance report including amplitude consistency error, phase consistency error, and system stability coefficient. Through a two-stage verification mechanism of single-tone signal calibration + multi-bandwidth QPSK verification, it covers the complete working bandwidth of 200kHz~56MHz, and shows excellent adaptability in scenarios such as 5G communication and phased array radar. It integrates an ILA online logic analyzer and a VIO virtual debugging interface, which improves the system debugging efficiency by more than 60%. In this embodiment, the phase detection module specifically includes: a clock multiplication unit, used to multiply the reference clock from 30.72MHz to 409.6MHz through a phase-locked loop; a zero-crossing detection unit, including a signal shaping filter and a hysteresis comparator, used to convert the input signal into a square wave signal; and a digital logic operation unit, including an XOR gate and a precision counter, used to perform logical operations on the two square wave signals and output a pulse width modulation signal. The data processing unit performs continuous sampling and data filtering, suppresses random errors through a multi-measurement averaging algorithm, and outputs accurate phase difference parameters with confidence indicators. Remote monitoring enables intelligent operation and maintenance of the equipment. SPI serial-to-parallel conversion and a multi-master arbitration mechanism ensure optimal allocation of system resources and support parallel operation of multiple processing units. The timing alignment unit employs a 2149-level precise timing pipeline, completely eliminating signal jitter during the calibration process. This solution maintains high precision while possessing good engineering practicality and maintainability, providing reliable amplitude and phase consistency guarantees for wireless communication equipment.
[0029] This technical solution achieves intelligent operation and maintenance of equipment through remote monitoring. The SPI serial-to-parallel conversion and multi-master arbitration mechanism ensure the optimal allocation of system resources and support the parallel operation of multiple processing units. The timing alignment unit adopts a 2149-level precise timing pipeline, which completely eliminates the signal jitter phenomenon during the calibration process. This solution maintains high precision while also having good engineering practicality and maintainability, providing reliable amplitude and phase consistency guarantee for wireless communication equipment.
[0030] In this embodiment, the entire operation process can be controlled by a computer to provide signal feedback and implement the steps sequentially. These are all conventional knowledge in current automation control, and will not be elaborated on in this embodiment.
[0031] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A multi-channel amplitude and phase correction method based on a radio frequency chip, characterized in that, The application is applied to a phase correction device, and specifically comprises the following steps: S101: receiving a radio frequency chip initialization instruction of a terminal, generating an executable initialization sequence based on radio frequency chip configuration parameters, completing comprehensive initialization of radio frequency chip working modes, clock networks and interface parameters through a parallelized register configuration mechanism, and establishing a stable radio frequency signal transmission and reception basic environment; S102: based on a standard test signal preset and stored in an embedded memory, realizing synchronous transmission and reception of multi-channel signals through a high-speed interface, completing signal modulation, transmission and demodulation by using a signal chain built in the radio frequency chip, and feeding back multi-channel original data containing amplitude and phase deviations to the phase correction device through a loopback architecture; S103: the phase correction device improves system time resolution by clock frequency multiplication, converts analog phase differences into countable digital signals by combining zero-crossing detection and digital logic operation, effectively suppresses random errors by combining multiple measurement average algorithms, and accurately obtains phase difference parameters between the multi-channels; S104: converting the phase difference parameters into complex plane rotation factors by using a digital signal processing core, realizing phase rotation correction of input signals by using orthogonal signal synthesis and complex multiplication operation, and completing digital compensation of phase deviations; S105: after outputting the digital compensation signal, a sliding window power detection mechanism is used to calculate the signal amplitude characteristics of the radio frequency chip in real time, an adaptive correction coefficient is generated by combining a pre-stored gain mapping table, amplitude equalization processing is realized by accurate time sequence alignment, the amplitude and phase consistency of the radio frequency chip in different scenarios is verified by using multi-standard test signals, and verification data is output.
2. The method of claim 1, wherein the plurality of channels are corrected by the RFIC based on a plurality of amplitude and phase correction values. In S101, an executable initialization sequence based on radio frequency chip configuration parameters is generated, and comprehensive initialization of radio frequency chip working modes, clock networks and interface parameters is completed through a parallelized register configuration mechanism, including: A register configuration is completed by using a state machine to control an SPI communication interface, the state machine includes an idle state, a command assembly state, a data transmission state and a state verification state, and the configuration parameters are combined into a 24-bit command word in the command assembly state, including 3-bit operation codes, 10-bit address fields and 8-bit data fields; In the data transmission state, the command word is output to the radio frequency chip bit by bit through a clock synchronization mechanism, and the configuration result is verified by reading a response flag bit in the state verification state; For 1024 configuration registers of the radio frequency chip, a paging addressing mechanism is used to realize complete coverage, and a retry mechanism is set for key parameter configuration, and when verification fails, the configuration command is automatically resent until success.
3. The method of claim 2, wherein the plurality of channels are determined by a plurality of RF paths. In S102, based on a standard test signal preset and stored in an embedded memory, synchronous transmission and reception of multi-channel signals are realized through a high-speed interface, including: A data interface uses a DDR (double data rate) mode, and data frame synchronization signal pulses are sampled at both clock rising and falling edges, I / Q data of a first channel is transmitted at the rising edge stage, and I / Q data of a second channel is transmitted at the falling edge stage; The data format is a 12-bit signed number, and I and Q data are transmitted on the same physical link by using time division multiplexing. The receiving end extracts the clock signal from the LVDS differential pair through a data recovery unit, and converts the serial data stream into parallel data through a serial-to-parallel conversion module for use by subsequent units.
4. The method of claim 3, wherein the plurality of channels are corrected by the RFIC based on a plurality of amplitude and phase correction values. In S103, the phase correction device improves the system time resolution by clock frequency multiplication, converts the analog phase difference into a countable digital signal by combining zero-crossing detection with digital logic operation, including: The clock frequency multiplication multiplies the reference clock from 30.72MHz to 409.6MHz through a phase-locked loop; The zero-crossing detection circuit includes a signal shaping filter and a hysteresis comparator, wherein the filter uses a 4th order Butterworth low-pass filter with a cutoff frequency set to 1.5 times the signal frequency, and the hysteresis comparator sets a hysteresis voltage of 5mV to prevent false triggering caused by noise; The digital logic operation includes an exclusive OR gate and a precision counter. The exclusive OR gate performs logic operation on two square wave signals to output a pulse width modulation signal. The precision counter counts the pulse width under the drive of a frequency multiplication clock. The phase difference is calculated by the formula Δφ=(N×T)×360° / (1 / f c ), where N is the counting value, T is the period of the frequency multiplication clock, and f c is the signal carrier frequency.
5. The method of claim 4, wherein the plurality of channels are corrected by the RFIC based on a plurality of amplitude and phase correction values. Combining the multiple measurement average algorithm effectively suppresses random errors and accurately obtains the phase difference parameters between multiple channels, including: A continuous sampling mechanism is established, and within the set measurement time window, the phase difference signal of the two signals is sampled at least 1024 times continuously based on the high-frequency clock of 409.6MHz. Each sampling process includes complete zero-crossing detection, XOR processing and clock counting process. The obtained raw phase difference data is temporarily stored in a first-in-first-out cache queue with a depth of 1024, ensuring that the phase change characteristics of the signal in different periods can be covered; Perform data screening and average processing, sort and analyze the 1024 phase difference measurement values in the cache queue, identify and remove abnormal measurement points caused by transient interference using the median absolute deviation algorithm, and divide the remaining valid data into 8 groups in chronological order. Calculate the arithmetic mean of each group respectively; Weighted average processing is performed on each group of arithmetic mean values, with higher weights given to the most recent time period. The output phase difference parameter also includes a confidence index reflecting the reliability of the measurement result.
6. The method of claim 5, wherein the plurality of channels are corrected by the RFIC based on a plurality of amplitude and phase correction values. In S104, the phase difference parameter is converted into a complex plane rotation factor using a digital signal processing core, including: The phase difference parameter is normalized to map it to a standard phase range of 0-360 degrees, and then quantized to a 14-bit precision phase control word through a phase control word conversion, with the high 14 bits serving as the effective phase input parameter of the DDS core. The DDS core generates corresponding quadrature signals in real time based on the phase accumulator principle under the drive of a 40MHz working clock, outputs 16-bit precision cosθ and sinθ digital signals by querying the pre-stored sine / cosine waveform table, and forms the real part and imaginary part of the complex plane rotation factor. The input IQ data and the rotation factor are sent to a complex multiplication unit, which is composed of four parallel real multipliers, which calculate four product terms I1*cosθ, Q1*sinθ, I1*sinθ and Q1*cosθ respectively, and two carry-advance adders subsequently perform real part combination I1*cosθ-Q1*sinθ and imaginary part combination I1*sinθ+Q1*cosθ operations, completing complex multiplication in a single clock cycle, and the output data is subjected to dynamic range detection and saturation processing, and when the result exceeds the 16-bit representation range, it is automatically limited to maintain the accurate amplitude and phase characteristics of the output signal.
7. The method of claim 6, wherein the plurality of channels are corrected by the RFIC based on a plurality of amplitude and phase correction values. In S105, the sliding window power detection mechanism is used to calculate the signal amplitude characteristics of the radio frequency chip in real time, and the pre-stored gain mapping table is used to generate adaptive correction coefficients, including: A 2048-point shift register array is used to construct a sliding window, and the absolute value of the new sampled signal is moved to the front end of the register every clock cycle, while the earliest stored absolute value is moved out from the end, and the power calculation is maintained in real time. The power and is updated by adding the new moved absolute value and subtracting the moved absolute value from the current cumulative value; The power sum is right-shifted by 11 bits to obtain the average power value of 2048 points, which is used as the index address for searching the gain mapping table. The gain mapping table uses a piecewise linear interpolation structure, and the 20-bit gain value is divided into 512 intervals according to the power range. When the average power value falls between two stored points, the accurate gain coefficient is calculated by linear interpolation; The timing alignment ensures that the input signal and the gain coefficient obtained by table lookup are accurately synchronized through a 2149-stage pipeline register. According to the changes in the actual working environment, the gain mapping relationship is continuously optimized through background learning. When significant changes in channel conditions are detected, the gain table update process is automatically started, and the gain mapping curve is recalculated based on the latest collected signal characteristics. The update uses a smooth transition to avoid signal distortion caused by sudden gain changes and preserves historical mapping data for performance comparison and analysis; The timing alignment uses a multi-stage pipeline structure, which includes accurate compensation distribution of signal absolute value calculation delay, sliding window processing delay, table lookup delay and multiplier delay: signal preprocessing occupies 5 clock cycles, sliding window power calculation occupies 2048 clock cycles, gain table lookup occupies 64 clock cycles, and multiplication operation occupies 32 clock cycles, ensuring that each data sample can be accurately matched with the corresponding gain coefficient, achieving amplitude correction effect without jitter.
8. The method of claim 7, wherein the plurality of channels are corrected by the RFIC based on a plurality of amplitude and phase correction values. The amplitude and phase consistency of the radio frequency chip in different scenarios is verified by multi-standard test signals, and the verification data is output, including: A hierarchical verification system is established, and a single-tone signal is used for basic verification. The basic functions of the system are confirmed by observing the time domain waveform and spectral characteristics, and then a 10.6MHz narrowband QPSK signal is used to verify the low frequency band performance, focusing on detecting the convergence of the constellation diagram and the phase noise characteristics; High frequency band stability is verified using 18MHz and 26MHz wideband QPSK signals, in-band flatness and group delay consistency are analyzed, -3dB bandwidth and in-band fluctuation of each test signal are measured by a spectrum analyzer, error vector magnitude, IQ imbalance and constellation symmetry indicators are obtained by a vector signal analyzer; Test data is automatically recorded and a comprehensive performance report containing amplitude consistency error, phase consistency error and system stability coefficient is generated, and a re-calibration process is automatically triggered when any indicator exceeds the threshold.
9. A multi-channel amplitude and phase correction system based on radio frequency chips, characterized in that, The system is used for executing the multi-channel amplitude and phase correction method of any one of claims 1-8, and the system comprises: A configuration initialization module is configured to receive terminal instructions and generate an initialization sequence executable by a radio frequency chip, complete register configuration through an SPI interface controlled by a state machine, and establish a stable radio frequency signal transmission and reception basic environment; A signal transmission and reception processing module is configured to read standard test signals from an embedded memory, realize synchronous transmission and reception of multi-channel signals through an LVDS interface, complete signal modulation and demodulation by using a built-in signal chain of the radio frequency chip, and feed back original data containing amplitude and phase deviation to a processing system through a loopback architecture; A phase detection module is configured to improve system time resolution by frequency multiplication of a phase-locked loop, convert analog phase difference into digital signals by combining zero-crossing detection and digital logic operation, and accurately obtain phase difference parameters between multi-channels by combining multiple measurement average algorithms; A phase correction module is configured to convert the phase difference parameters into complex plane rotation factors by a DDS digital signal processing core, realize phase rotation correction of input signals by using complex multiplication operation, and complete digital compensation of phase deviation; An amplitude correction module is configured to realize real-time calculation of signal amplitude characteristics by using a sliding window power detection mechanism, generate adaptive correction coefficients by combining a pre-stored gain mapping table, and realize time alignment and amplitude equalization processing by a multi-stage pipeline; A system verification module is configured to verify amplitude and phase consistency of the system in different scenarios by using multi-standard test signals, and automatically generate a comprehensive performance report containing amplitude consistency error, phase consistency error and system stability coefficient.
10. The radio frequency chip based multi-channel amplitude and phase correction system based on video world model as claimed in claim 9, wherein, The phase detection module specifically comprises: A clock frequency multiplication unit is configured to multiply a reference clock from 30.72MHz to 409.6MHz by a phase-locked loop; A zero-crossing detection unit comprises a signal shaping filter and a hysteresis comparator, and is configured to convert an input signal into a square wave signal; A digital logic operation unit comprises an XOR gate and a precision counter, and is configured to perform logic operation on two square wave signals and output a pulse width modulation signal; A data processing unit is configured to perform continuous sampling and data screening, suppress random errors by using a multiple measurement average algorithm, and output accurate phase difference parameters with a confidence index.
