FPGA (Field Programmable Gate Array) implementation method for distributed high-speed data transmission synchronization

By employing multi-level clock synchronization and synchronization signal delay compensation methods, combined with a checksum mechanism, efficient and reliable data transmission between distributed FPGA chips is achieved. This solves the synchronization problems caused by clock synchronization difficulties and path differences in traditional methods, thereby improving the system's stability and transmission efficiency.

CN120956374APending Publication Date: 2025-11-14NO 8511 RES INST OF CASIC
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202511004024.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-07-21
Publication Date
2025-11-14

AI Technical Summary

Technical Problem

In high-speed data transmission, traditional methods face difficulties in clock synchronization, low data transmission reliability, and poor anti-interference capabilities. In particular, in distributed multi-node transmission scenarios, the lack of path difference compensation mechanisms leads to system synchronization failures.

Method used

By employing multi-level clock synchronization, synchronization signal delay compensation, and data reception and transmission methods, combined with checksum mechanisms, automatic repeat request (ARQ), and hardware acceleration, the LVDS interface is used to transmit the synchronization clock. A fixed synchronization period and synchronization frame identifier are used, and the number of GTH interface branches is dynamically adjusted to ensure synchronous data transmission among multiple FPGA chips.

Benefits of technology

It achieves efficient and reliable data transmission between distributed FPGA chips, solves the synchronization problems caused by clock synchronization difficulties and path differences in traditional methods, and improves the stability and transmission efficiency of the system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120956374A_ABST
    Figure CN120956374A_ABST
Patent Text Reader

Abstract

The invention discloses an FPGA (Field Programmable Gate Array) implementation method for distributed high-speed data transmission synchronization, which belongs to the field of data processing, and comprises the following steps of: firstly, transmitting a processing clock of the last stage to each stage as a synchronous clock to ensure multi-stage clock synchronization of data; secondly, the data sending end inserts a synchronization frame under the processing of a synchronization clock for packaging, synchronously sends data according to a synchronization signal in a fixed period, and transmits the synchronization signal to the next level; thirdly, the data receiving end performs delay phase modulation on the synchronization signal to ensure that the time for the synchronization signal to arrive at the data receiving end is fixed; and finally, a data receiving end synchronously receives the multiple groups of high-speed serial port data by using the synchronous signal and the synchronous frame.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of data processing, specifically relating to an FPGA implementation method for distributed high-speed data transmission synchronization. Background Technology

[0002] In the field of high-speed data transmission, especially for high-speed intermediate frequency data transmission, traditional methods often face problems such as difficulty in clock synchronization, low data transmission reliability, and poor anti-interference capability. To address these issues, this invention proposes a distributed high-speed data transmission synchronization method based on FPGA. This method achieves high-speed synchronous data transmission through multi-level clock synchronization, synchronization signal transmission delay compensation, and multi-level data reception and transmission techniques. Furthermore, by incorporating checksum mechanisms, Automatic Repeat Request (ARQ), hardware acceleration, and efficient packetization and depacketization, the system's reliability and transmission efficiency are further improved.

[0003] Chinese patent CN202311489700.7 proposes a robust multi-fiber data synchronization method and apparatus based on FPGA. It employs link detection on multiple optical fibers, followed by a first-level anomaly handling process to filter out "fiber link disconnection" anomalies. Anomaly filtering within multi-fiber data frames is completed using a statistical threshold decision method. Synchronization is achieved when the number of decision pulses arriving simultaneously is greater than or equal to 80% of the total number of optical fibers. However, based on fiber link status detection and fixed threshold logic, if the total number of optical fibers decreases or multiple link anomalies exist (more than 20% fiber failure), the system may fail to synchronize. Furthermore, it lacks a compensation mechanism for transmission path differences in distributed multi-node transmission scenarios.

[0004] Chinese patent CN202111507435.1 proposes a method for implementing bidirectional high-speed data transmission between FPGA chips. The data transmission device of this method includes two FPGA chips and a GTH high-speed transceiver. In the transmitting FPGA, the transmitting control module splits the high-speed data into 8 channels according to a fixed clock cycle and sequentially sends them into 8 transmitting FIFO memories. Then, the first Aurora control module retrieves the data and sends it into the corresponding 8 first Aurora IP cores. The GTH high-speed transceiver transmits the data to the receiving FPGA. The second Aurora control module in the receiving FPGA sends the 8 channels of data received by the second Aurora IP core into 8 receiving FIFO memories. The receiving control module retrieves the 8 channels of data sequentially and combines them into 1 channel to achieve complete data recovery. The transmitting control module and the receiving control module are integrated into a single transceiver module to realize bidirectional data transmission between the two FPGA chips. Although they share the clock and reset module of the first GTH, the timing of the GTH channels in different QUADs may be offset due to differences in physical layout, affecting the synchronization of multi-channel data; it only supports 8 parallel GTH channels and cannot dynamically adjust the number of channels to adapt to different bandwidth requirements; if the clock domains of the bidirectional data streams of the transceiver module are not synchronized, signal competition or metastability may occur, reducing the stability of bidirectional transmission. Summary of the Invention

[0005] This invention proposes an FPGA implementation method for distributed high-speed data transmission synchronization. This method is primarily used for high-speed transmission of raw intermediate frequency (IF) data. Taking a 2400MHz AD sampling rate and a 14-bit data width as an example, combined with 16 independent GTH interfaces, it achieves a total data transmission rate of 33.6Gbps. Through techniques such as data phase splitting, packetization, synchronization frame insertion, and high-speed serial port transmission, synchronous data transmission among multiple distributed FPGA chips is ensured.

[0006] The technical solution for implementing this invention is: an FPGA implementation method for distributed high-speed data transmission synchronization, comprising the following steps:

[0007] Step 1, Multi-level Clock Synchronization: Sort the N distributed data terminals according to their physical location from the master end to the farthest end in ascending order, including the first-level transmitter, the second-level data terminals to the (N-1)th-level data terminals, and the Nth-level receiver, and determine the cascading relationship; transmit the processing clock of the last-level (Nth-level) data terminal to the other levels through the LVDS interface as the synchronization clock sync_clock to ensure clock synchronization, where the value of N ranges from 2 to 20; Proceed to Step 2;

[0008] Step 2: First-level data transmission: The first-level data transmitter uses a synchronous clock sync_clock to generate a synchronization signal sync_signal with a fixed period of M. This signal is transmitted synchronously to the next level via the P-group GTH high-speed serial port, while simultaneously transmitting the synchronization signal. Here, P ranges from 1 to 32, and M is an integer multiple of P, with a value range of 1 to 65536. Proceed to Step 3.

[0009] Step 3: Intermediate-level data terminal synchronous reception: The data terminals from level 2 to level N-1 receive the synchronization signal sync_signal from the previous level, generate sync_signal_delay by delaying it by an integer multiple of M periods, buffer P groups of GTH data streams, and parse the synchronization frame to extract intermediate-level data; proceed to step 4.

[0010] Step 4: Intermediate-level data synchronous transmission: Combine sync_signal_delay to clock-align intermediate-level data, repackage and insert synchronization frames to generate transmission data, and serially transmit to the next level through P-group GTH, synchronously transmitting sync_signal_delay; proceed to step 5.

[0011] Step 5: Nth-level receiver synchronous reception: The Nth-level receiver receives the intermediate-level data and delayed synchronization signal sent by the N-1th-level data terminal, and performs final synchronization processing to ensure that the data and delayed synchronization signal are aligned for subsequent digital signal processing.

[0012] Compared with the prior art, the significant advantages of this invention are:

[0013] (1) Multi-level clock synchronization and path delay compensation: By using the processing clock of the last-stage receiver as a unified synchronization source and transmitting it to all nodes through the LVDS interface, clock source consistency is ensured. Intermediate nodes perform integer-cycle delay compensation (an integer multiple of the synchronization period M) on the received synchronization signal, thereby eliminating the transmission delay difference caused by different physical distances. This solves the problem that traditional methods cannot compensate for path differences, as well as the problem that existing technologies (such as CN2023) cannot synchronize when the number of optical fibers decreases.

[0014] (2) High-efficiency and reliable data transmission guarantee: A fixed synchronization period M is adopted in combination with the start and end markers of the synchronization frame (such as 55AA and AA55) to ensure reliable detection and alignment of data frames. By dynamically configuring P groups of GTH high-speed serial ports (P ranges from 1 to 32), the transmission bandwidth can be flexibly adjusted, solving the problem of fixed branching in existing technologies (such as CN2021). At the same time, combined with hardware verification and ARQ technologies, the anti-interference capability and data reliability are improved.

[0015] (3) Stable multi-level cascaded transmission architecture: Intermediate nodes ensure strict synchronization of data at each level through data caching, parsing, repackaging, and synchronization alignment. The final synchronization processing at the last-level receiver reduces metastability caused by clock domain asynchrony, improving system stability and transmission efficiency, which is superior to the bidirectional transmission module design in traditional methods. Attached Figure Description

[0016] Figure 1 This is a flowchart of an FPGA implementation method for distributed high-speed data transmission synchronization.

[0017] Figure 2 This is a schematic diagram of synchronous clock transmission. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] The technical solutions of the various embodiments of the present invention can be combined with each other, but only if they can be implemented by those skilled in the art. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.

[0020] The following section will further introduce the specific implementation method, as well as the technical difficulties and inventive points of this invention, using this design example as an example.

[0021] Combination Figure 1 and Figure 2 An FPGA implementation method for distributed high-speed data transmission synchronization includes the following steps:

[0022] Step 1, Multi-level Clock Synchronization: Sort the N distributed data terminals in order of physical location from the master end to the farthest end, including the first-level transmitter, the second-level data terminals to the (N-1)th-level data terminals, and the Nth-level receiver, and determine the cascading relationship; transmit the processing clock of the Nth-level data terminal to the other levels through the LVDS interface as the synchronization clock sync_clock to ensure clock synchronization, where the value of N ranges from 2 to 20.

[0023] Multi-level clock synchronization specifically includes the following steps:

[0024] Step 11: Sort the N distributed data terminals according to their physical location from the master terminal to the farthest terminal, and determine the data cascading relationship of each data terminal.

[0025] Step 12: Transmit the processing clock of the last stage, i.e. the Nth stage receiver, to each stage through the LVDS interface as the synchronization clock sync_clock, to ensure clock synchronization during multi-stage data transmission.

[0026] Step 13: Other data terminals at all levels receive the synchronization clock sync_clock and use global clock resources for clock processing to ensure the stability of the multi-level synchronization clock quality and clock synchronization.

[0027] Proceed to step 2.

[0028] Step 2, First-level data transmission: The first-level data transmitter uses a synchronous clock sync_clock to generate a synchronization signal sync_signal with a fixed period of M, which is transmitted synchronously to the next level through the P-group GTH high-speed serial port. At the same time, the synchronization signal is transmitted. Among them, P ranges from 1 to 32, and M is an integer multiple of P, with a value range of 1 to 65536.

[0029] The first-level data sender synchronously transmits data, specifically including the following steps:

[0030] Step 21: Use the synchronous clock sync_clock to generate a synchronization signal sync_signal with a fixed period of M, and transmit the synchronization signal to the next stage;

[0031]

[0032] Where i represents the clock tick.

[0033] Step 22: Generate a corresponding synchronization frame sync_frame based on the synchronization signal sync_signal with a fixed period of M, and then insert the synchronization frame sync_frame into the data sending end under the processing of the synchronization clock;

[0034]

[0035] Where 55AA is a hexadecimal number, representing the start value of the synchronization frame; AA55 represents the end value of the synchronization frame.

[0036] Step 23: Pack the data to be sent according to the specified data format to form the dataout to be sent.

[0037] Step 24: Use the P-group GTH high-speed serial port to send the dataout to the next-level data terminal, and at the same time transmit the synchronization signal sync_signal to the next-level data terminal.

[0038] Proceed to step 3.

[0039] Step 3: Intermediate-level data terminal synchronous reception: The data terminals from level 2 to level N-1 receive the synchronization signal sync_signal from the previous level, generate sync_signal_delay by delaying it by an integer multiple of M periods, buffer P groups of GTH data streams, and parse the synchronization frame to extract intermediate-level data.

[0040] The intermediate data receiving stage synchronously includes the following steps:

[0041] Step 31: Use the second-level data terminal to the (N-1)th-level data terminal as the intermediate-level data terminal. The intermediate-level data terminal receives the synchronization signal sync_signal, and then delays the received synchronization signal sync_signal_signal to form a delayed synchronization signal sync_signal_delay. This ensures that the time interval between the delayed synchronization signal sync_signal_delay arriving at the data receiving end and the synchronization signal sync_signal arriving at the sending end is a fixed integer multiple of the synchronization period M.

[0042] Step 32: The data receiving end buffers the P group GTH high-speed serial port data to obtain buffered data dataout_rec_tmp, and parses it according to the specified data format to extract the synchronization frame rec_syncframe;

[0043] Step 33: Parse the received cached data dataout_rec_tmp based on the received synchronization frame rec_syncframe to obtain the intermediate data dataout_rec.

[0044] Proceed to step 4.

[0045] Step 4: Intermediate-level data synchronous transmission: Combine sync_signal_delay to clock-align intermediate-level data, repackage and insert synchronization frames to generate transmission data, and serially transmit it to the next level through P-group GTH, synchronously transmitting sync_signal_delay.

[0046] The intermediate-level data terminal synchronously sends data, specifically including the following steps:

[0047] Step 41: Combine the delay synchronization signal sync_signal_delay with the synchronization clock sync_clock to delay and synchronize the received intermediate data dataout_rec to obtain the synchronized data dataout_sync, ensuring that dataout_sync is synchronized with the synchronization signal sync_delay.

[0048] Step 42: Insert the synchronization data dataout_sync into the synchronization frame for packaging, and transmit it to the next-level data terminal through the P group GTH high-speed serial port.

[0049] Step 43: Transmit the synchronization signal sync_delay to the next level data terminal.

[0050] Proceed to step 5.

[0051] Step 5: Nth-level receiver synchronous reception: The Nth-level receiver receives the intermediate-level data and delayed synchronization signal sent by the N-1th-level data terminal, and performs final synchronization processing to ensure that the data and delayed synchronization signal are aligned for subsequent digital signal processing.

[0052] The synchronous reception at the Nth level receiver specifically includes the following steps:

[0053] Step 51: The Nth-level receiver synchronously receives the intermediate-level data and delayed synchronization signal sent by the N-1th-level data terminal. First, it receives the synchronization signal sync_signal N sent by the N-1th-level data terminal; then, it delays and modulates the phase of the received synchronization signal sync_signal N to form a delayed synchronization signal sync_signal_delay_N, ensuring that the time interval between the delayed synchronization signal sync_signal_delay_N arriving at the Nth-level data receiver and the synchronization signal sync_signal_N of the N-1th-level data terminal is a fixed integer multiple of the synchronization period M.

[0054] Step 52: The Nth-level receiver obtains the synchronization frame rec_syncframe_N according to the specified data format. Based on the received synchronization frame rec_syncframe_N, it synchronously receives the data dataout_N sent by the previous level from the P-group GTH high-speed serial port, and obtains the Nth-level data dataout_rec_N.

[0055] Step 53: Combine the delay synchronization signal sync_N_delay with the synchronization clock sync_clock to delay and synchronize the received N-1 level data dataout_rec_N, and obtain synchronized data dataout_sync_N. Ensure that dataout_sync_N is synchronized with the synchronization signal sync_delay_N, which will be used for subsequent digital signal processing.

[0056] Example 1

[0057] An FPGA implementation method for distributed high-speed data transmission synchronization is presented, primarily used for high-speed transmission of raw intermediate frequency (IF) data. Taking an AD sampling rate of 2400MHz, a data width of 14 bits, and an FPGA chip with 16 independent GTH interfaces as an example, the data rate is 2400MHz * 14 bits = 33.6Gbps. The 2400MHz raw IF data is divided into 16 phases, with the number of phases matching the number of GTH interfaces. Each phase clock is 150MHz. The data from each phase is packaged every 32 clock cycles (32 data points in total), resulting in an effective data width of 448 bits. A 16-bit header, a 16-bit synchronization frame, a 16-bit checksum, and a 16-bit trailer are added to complete the data packaging, forming a complete 512-bit data packet. Therefore, the total data rate is 2400MHz / 32 * 512 = 38.4Gbps. It adopts 16X independent GTH interfaces of FPGA chip, each 1X has a transmission rate of 3.125Gbps, and the total transmission rate of 16X parallel transmission is 50Gbps, which meets the data transmission requirement of 38.4Gbps and realizes the synchronous transmission of intermediate frequency data to multiple distributed FPGA chips.

[0058] Baotou Data area Synchronization Frame Checksum wrap tail Bit width 16 bits 448 bits 16 bits 16 bits 16 bits content 0x7E7E 32-point data sync_frame Verification results from packet header to synchronization frame 0xE7E7

[0059] Combination Figure 1 An FPGA implementation method for distributed high-speed data transmission synchronization, the steps of which are as follows:

[0060] Step 1, multi-level 150MHz clock synchronization, specifically includes the following steps:

[0061] Step 11: Sort the N distributed data terminals and determine the data cascading relationship between each data terminal;

[0062] Implementation Details: In a distributed system, the N data endpoints first need to be sorted, including the first-level sender, the second-level data endpoints up to the (N-1)th-level data endpoints, and the Nth-level receiver, to determine their cascading relationships and the data cascading relationships between them. This is typically determined based on the system topology and data flow. In a chain-like structure, data starts from the first-level data endpoint and is sequentially passed to the Nth-level receiver. Simultaneously, a Clock Data Recovery (CDR) circuit is introduced at each node to regenerate a high-quality clock signal, reducing clock skew caused by jitter and transmission delay.

[0063] Step 12: Divide the 2400MHz raw intermediate frequency data into 16 phases, with each phase having a data rate of 150MHz, i.e., a processing clock of 150MHz; transmit the 150MHz processing clock of the last stage (Nth stage) data terminal to the other N-1 stages via the LVDS interface as a synchronization clock sync_clock, to ensure clock source synchronization during multi-stage data transmission, where the value of N ranges from 2 to 20;

[0064] Implementation Details: After determining the data cascading relationship, a final data stage (Nth stage) needs to be selected as the synchronization clock source. The processing clock of this stage node is transmitted to all other nodes via the LVDS (Low Voltage Differential Signaling) interface as the global synchronization clock `sync_clock`. This method ensures that all nodes use the same clock source, reduces clock skew, and ensures high-speed and low-noise transmission. During clock transmission, a checksum mechanism is added to ensure the correct transmission of the synchronization clock; cables or channels with high anti-interference capabilities are used to reduce signal attenuation and noise interference during transmission.

[0065] Step 13: Other data terminals at all levels receive the synchronization clock sync_clock and use the global clock resource for clock processing. At this time, the clock source processed by each distributed FPGA is the same 150MHz clock, ensuring the stability of the multi-level synchronization clock quality and clock synchronization.

[0066] Implementation details: After all other nodes receive the synchronization clock (sync_clock), they need to process it to ensure the stability and consistency of the clock signal. A PLL is used within the node to process the received sync_clock, regenerating a high-quality clock and reducing jitter.

[0067] The nodes use global clock resources for clock processing, enabling nodes distributed across different FPGA chips to share the same clock source and ensuring that the local clocks of all nodes are consistent with the global clock resources.

[0068] Step 2, the first-level data sender synchronously sends data, specifically including the following steps:

[0069] Step 21: Use the synchronous clock sync_clock to generate a synchronization signal sync_signal with a fixed period of 8192, and transmit the synchronization signal to the next stage;

[0070]

[0071] Implementation details: The first-level data transmitter uses the global synchronization clock sync_clock as the reference clock to generate a synchronization signal sync_signal with a fixed period of 8192. sync_signal is then transmitted to the next-level node via a high-speed channel for subsequent data synchronization.

[0072] Step 22: Generate a corresponding synchronization frame sync_frame based on the synchronization signal sync_signal with a fixed period of 8192, and then insert the synchronization frame into the data transmission end for packaging under the synchronization clock processing.

[0073]

[0074] Implementation details: A synchronization frame (sync_frame) conforming to a specific format is generated based on sync_signal to identify the start position of the data packet. Under the control of sync_clock, the data sending end inserts the sync_frame into the predetermined position of the data packet and packages the data to form a complete data frame.

[0075] Step 23: For the intermediate frequency data divided into 16 phases, take 32 data points from each phase to form a data packet, and package them independently according to the specified data format to form 16 transmit data packets (dataout).

[0076] Baotou Data area Synchronization Frame Checksum wrap tail Bit width 16 bits 448 bits 16 bits 16 bits 16 bits content 0x7E7E 32-point data sync_frame Verification results from packet header to synchronization frame 0xE7E7

[0077] Implementation details: The raw data to be sent is divided into several data packets and packaged according to a specified data format. The data format typically includes a header, data area, synchronization frame, checksum, and trailer. Each data packet is packaged according to the above format to ensure data integrity and standardization.

[0078] Step 24: 16 data packets will be generated. Each data packet will be sent to the next-level data terminal using a high-speed serial port. At the same time, the synchronization signal sync_signal will be transmitted to the next-level data terminal.

[0079] Implementation details: The packaged data is sent to the next-level node via 16 high-speed serial ports (such as GTH interfaces). Each data packet is sent through an independent high-speed serial port, ensuring parallel data transmission and improving overall transmission efficiency. Simultaneously, the synchronization signal `sync_signal` is transmitted for data synchronization in the next-level node, ensuring that the synchronization signal and data signal are in the same timing sequence and avoiding data misunderstandings caused by clock deviations.

[0080] Step 3: Intermediate-level data terminal synchronous reception, specifically including the following steps:

[0081] Step 31: Use the second-level data terminal to the (N-1)th-level data terminal as the intermediate-level data terminal. The intermediate-level data terminal receives the synchronization signal sync_signal, and then delays the received synchronization signal sync_signal_signal to form a delayed synchronization signal sync_signal_delay. This ensures that the time interval between the delayed synchronization signal sync_signal_delay arriving at the data receiving end and the synchronization signal sync_signal arriving at the sending end is a fixed multiple of the synchronization period 8192.

[0082] Implementation details: The intermediate data terminals (from level 2 to level N-1) first receive the synchronization signal `sync_signal` from the previous level node. Then, a delay line or delay module is used to delay `sync_signal`, generating a delayed synchronization signal `sync_signal_delay`. This ensures that the time interval between the delayed `sync_signal_delay` and the original `sync_signal` is fixed, guaranteeing data synchronization consistency across nodes and preventing data misalignment due to inconsistent transmission delays.

[0083] Step 32: The data receiving end buffers 16 sets of high-speed serial port data to obtain buffered data dataout_rec_tmp, and parses it according to the specified data format to extract the synchronization frame rec_syncframe;

[0084] Implementation details: The receiving end stores the received high-speed serial port data in a FIFO (First-In-First-Out) buffer, forming temporary received data `dataout_rec_tmp`, ensuring data continuity and integrity. Subsequently, based on the data packet format, the synchronization frame `rec_syncframe` is identified and extracted to ensure accurate data parsing. The buffered data is parsed according to the specified data format to extract the synchronization frame `rec_syncframe` for subsequent data synchronization.

[0085] Step 33: Parse the received cached data dataout_rec_tmp based on the received synchronization frame rec_syncframe to obtain the intermediate data dataout_rec.

[0086] Implementation details: The parsing module is used to parse dataout_rec_tmp and extract the synchronization frame rec_syncframe. The starting point for reading data is adjusted according to rec_syncframe. The cached dataout_rec_tmp is parsed to ensure the correctness and consistency of the data. The parsed data is verified to ensure data integrity and consistency, and finally the intermediate data dataout_rec is obtained.

[0087] Step 4, intermediate-level data transmission, specifically includes the following steps:

[0088] Step 41: Combine the delay synchronization signal sync_signal_delay with the synchronization clock sync_clock to delay and synchronize the received intermediate data dataout_rec to obtain the synchronized data dataout_sync, ensuring that dataout_sync is synchronized with the synchronization signal sync_delay.

[0089] Implementation details: At the intermediate data level, the received intermediate data `dataout_rec` needs to undergo delay synchronization processing. Timing adjustments for `dataout_rec` are performed using `sync_clock` and `sync_signal_delay`. This ensures that the synchronized data `dataout_sync` remains synchronized with the synchronization signal `sync_delay`.

[0090] Step 42: Insert the synchronization data dataout_sync into the synchronization frame for packaging, and transmit it to the next-level data terminal through 16 high-speed serial ports;

[0091] Implementation details: The synchronized `dataout_sync` is re-inserted into the synchronization frame, forming a new data packet according to the specified data format. Data is transmitted in parallel through 16 high-speed serial ports to the next-level node. Error correction coding is added to the data transmission to improve reliability.

[0092] Step 43: Transmit the synchronization signal sync_signal_delay to the next level data terminal.

[0093] Implementation details: The sync_signal_delay is transmitted through a high-speed channel to ensure that its arrival time at the next level node is consistent with the transmission sequence of the previous level node.

[0094] Step 5, the Nth level receiver synchronizes the reception, specifically including the following steps:

[0095] Step 51: The Nth-level receiver synchronously receives the data output from the (N-1)th level. First, it receives the synchronization signal sync_signal N sent by the (N-1)th level; then, it delays and modulates the received synchronization signal sync_signal N to form a delayed synchronization signal sync_signal_delay_N, ensuring that the time interval between the delayed synchronization signal sync_signal_delay_N arriving at the Nth-level data receiver and the synchronization signal sync_signal_N of the (N-1)th-level transmitter is a fixed integer multiple of the synchronization period M.

[0096] Implementation details: The Nth-level receiver first receives the synchronization signal sync_signal_N sent by the (N-1)th-level node, and uses a delay processing module to delay sync_signal_N. This ensures that the time interval between the generated sync_signal_delay_N and the original sync_signal_N is fixed, avoiding data misalignment caused by transmission delay.

[0097] Step 52: The data receiving end obtains the synchronization frame rec_syncframe_N according to the specified data format, and synchronously receives the data out_N sent by the previous level from the P-group high-speed serial port data based on the received synchronization frame rec_syncframe_N to obtain the Nth level data dataout_rec_N.

[0098] Implementation details: The received data is parsed using a parsing module, and the starting point for reading the data is adjusted according to rec_syncframe_N. The parsed data is then validated to ensure data integrity and consistency.

[0099] Step 53: After receiving dataout_rec_N, the Nth-level node combines the delay synchronization signal sync_N_delay with the global synchronization clock sync_clock to perform delay synchronization processing on the data, generating the final synchronized data dataout_sync_N. This ensures that dataout_sync_N remains synchronized with the synchronization signal sync_delay_N, providing a reliable data source for subsequent digital signal processing.

[0100] Implementation details: Timing adjustments to dataout_rec_N are performed using sync_clock and sync_N_delay. This ensures that the synchronized dataout_sync_N remains synchronized with the sync_delay_N signal.

[0101] Typical applications

[0102] The distributed high-speed data transmission synchronization method of this invention has a wide range of applications. Here are some typical applications:

[0103] (1) Electronic Information System

[0104] In electronic information systems, high-speed data transmission synchronization methods can be used in multi-level signal processing and distributed data acquisition systems to ensure data synchronization and efficient transmission between nodes, providing a reliable data source for subsequent signal analysis and processing.

[0105] (2) Data Centers and Cloud Computing

[0106] In large-scale data centers and cloud computing environments, the method of this invention can be used for high-speed data synchronization and transmission between distributed servers, supporting high concurrency and large-scale data processing needs, and improving the overall throughput and response speed of the system.

[0107] (3) Real-time control system

[0108] In systems requiring high-precision real-time control, such as autonomous driving and industrial automation, the method of this invention can ensure data synchronization and timeliness between control nodes, thus guaranteeing the stability and reliability of the system.

[0109] (4) Scientific research and experimentation

[0110] In scientific research and experimental fields, such as high-energy physics experiments and astronomical observations, the method of this invention can be used in distributed data acquisition systems to support the acquisition and transmission of multi-point synchronous data, providing accurate data support for scientific analysis.

Claims

1. An FPGA implementation method for distributed high-speed data transmission synchronization, characterized in that, Includes the following steps: Step 1, Multi-level Clock Synchronization: Sort the N distributed data terminals according to their physical location from the master end to the farthest end in ascending order, including the first-level transmitting end, the second-level data terminals to the (N-1)th-level data terminals, and the Nth-level receiving end, and determine the cascading relationship; transmit the processing clock of the Nth-level data terminal to the other levels through the LVDS interface as the synchronization clock sync_clock to ensure clock synchronization, where the value of N ranges from 2 to 20; Proceed to Step 2; Step 2: First-level data transmission: The first-level data transmitter uses a synchronous clock sync_clock to generate a synchronization signal sync_signal with a fixed period of M. This signal is transmitted synchronously to the next level via the P-group GTH high-speed serial port, while simultaneously transmitting the synchronization signal. Here, P ranges from 1 to 32, and M is an integer multiple of P, with a value range of 1 to 65536. Proceed to Step 3. Step 3: Intermediate-level data terminal synchronous reception: The data terminals from level 2 to level N-1 receive the synchronization signal sync_signal from the previous level, generate sync_signal_delay by delaying it by an integer multiple of M periods, buffer P groups of GTH data streams, and parse the synchronization frame to extract intermediate-level data; proceed to step 4. Step 4: Intermediate-level data synchronous transmission: Combine sync_signal_delay to clock-align intermediate-level data, repackage and insert synchronization frames to generate transmission data, and serially transmit to the next level through P-group GTH, synchronously transmitting sync_signal_delay; Proceed to step 5; Step 5: Nth-level receiver synchronous reception: The Nth-level receiver receives the intermediate-level data and delayed synchronization signal sent by the N-1th-level data terminal, and performs final synchronization processing to ensure that the data and delayed synchronization signal are aligned for subsequent digital signal processing.

2. The FPGA implementation method for distributed high-speed data transmission synchronization according to claim 1, characterized in that, Step 1, multi-level clock synchronization, specifically includes the following steps: Step 11: Sort the N distributed data terminals according to their physical location from the master terminal to the farthest terminal, and determine the data cascading relationship of each data terminal. Step 12: Transmit the processing clock of the last stage, i.e. the Nth stage receiver, to each stage through the LVDS interface as the synchronization clock sync_clock to ensure clock synchronization during multi-stage data transmission; Step 13: Other data terminals at all levels receive the synchronization clock sync_clock and use global clock resources for clock processing to ensure the stability of the multi-level synchronization clock quality and clock synchronization.

3. The FPGA implementation method for distributed high-speed data transmission synchronization according to claim 1, characterized in that, In step 2, the first-level data sender synchronously sends data, specifically including the following steps: Step 21: Use the synchronous clock sync_clock to generate a synchronization signal sync_signal with a fixed period of M, and transmit the synchronization signal to the next stage; Where i represents the clock tick; Step 22: Generate a corresponding synchronization frame sync_frame based on the synchronization signal sync_signal with a fixed period of M, and then insert the synchronization frame sync_frame into the data sending end under the processing of the synchronization clock; Where 55AA is a hexadecimal number, representing the start value of the synchronization frame; AA55 represents the end value of the synchronization frame. Step 23: Pack the data to be sent according to the specified data format to form the data to be sent, dataout; Step 24: Use the P-group GTH high-speed serial port to send the dataout to the next-level data terminal, and at the same time transmit the synchronization signal sync_signal to the next-level data terminal.

4. The FPGA implementation method for distributed high-speed data transmission synchronization according to claim 1, characterized in that, In step 3, the intermediate data terminal synchronously receives data, specifically including the following steps: Step 31: Use the second-level data terminal to the (N-1)th-level data terminal as the intermediate-level data terminal. The intermediate-level data terminal receives the synchronization signal sync_signal, and then delays the received synchronization signal sync_signal_signal to form a delayed synchronization signal sync_signal_delay. This ensures that the time interval between the delayed synchronization signal sync_signal_delay arriving at the data receiving end and the synchronization signal sync_signal arriving at the sending end is a fixed integer multiple of the synchronization period M. Step 32: The data receiving end buffers the P group GTH high-speed serial port data to obtain buffered data dataout_rec_tmp, and parses it according to the specified data format to extract the synchronization frame rec_syncframe; Step 33: Parse the received cached data dataout_rec_tmp based on the received synchronization frame rec_syncframe to obtain the intermediate data dataout_rec.

5. The FPGA implementation method for distributed high-speed data transmission synchronization according to claim 1, characterized in that, In step 4, the intermediate data terminal synchronously sends data, specifically including the following steps: Step 41: Combine the delay synchronization signal sync_signal_delay with the synchronization clock sync_clock to delay and synchronize the received intermediate data dataout_rec to obtain the synchronized data dataout_sync, ensuring that dataout_sync is synchronized with the synchronization signal sync_delay. Step 42: Insert the synchronization data dataout_sync into the synchronization frame, package it, and transmit it to the next-level data terminal through the P group GTH high-speed serial port; Step 43: Transmit the synchronization signal sync_delay to the next level data terminal.

6. The FPGA implementation method for distributed high-speed data transmission synchronization according to claim 1, characterized in that, In step 5, the Nth-level receiver synchronously receives data, specifically including the following steps: Step 51: The Nth-level receiver synchronously receives the intermediate-level data and delayed synchronization signal sent by the N-1th-level data terminal. First, it receives the synchronization signal sync_signal N sent by the N-1th-level data terminal; then, it delays and modulates the phase of the received synchronization signal sync_signal N to form a delayed synchronization signal sync_signal_delay_N, ensuring that the time interval between the delayed synchronization signal sync_signal_delay_N arriving at the Nth-level data receiver and the synchronization signal sync_signal_N of the N-1th-level data terminal is a fixed integer multiple of the synchronization period M. Step 52: The Nth level receiver obtains the synchronization frame rec_syncframe_N according to the specified data format. Based on the received synchronization frame rec_syncframe_N, it synchronously receives the data out_N sent by the previous level from the P group of GTH high-speed serial port data to obtain the Nth level data dataout_rec_N. Step 53: Combine the delay synchronization signal sync_N_delay with the synchronization clock sync_clock to delay and synchronize the received N-1 level data dataout_rec_N, and obtain synchronized data dataout_sync_N. Ensure that dataout_sync_N is synchronized with the synchronization signal sync_delay_N, which will be used for subsequent digital signal processing.

Citation Information

Patent Citations

  • Implementation method of inter-chip bidirectional high-speed data transmission based on FPGA (Field Programmable Gate Array)

    CN114297121A

  • Robust multichannel optical fiber data synchronization method and device based on FPGA (Field Programmable Gate Array)

    CN117639921A