LPDDR4X simulation method and system based on statistics theory

Abstract

The invention relates to an LPDDR4X simulation method and system based on a statistics theory. The method comprises the steps that a single-byte BYTE time domain simulation platform used for simulating LPDDR4X signal transmission characteristics is constructed; routing delay analysis is carried out on packaging and PCB channels corresponding to each signal in the single-byte BYTE, time sequence inconsistency in signal transmission is captured, and then the influence of phase crosstalk noise on eye height is captured; carrying out pulse response analysis on the signal, and quantifying ISI noise basic data caused by packaging and a PCB channel; independently analyzing the influence of crosstalk noise on each signal to obtain direct influence data of noise on eye height; processing ISI noise basic data and direct influence data by utilizing an MATLAB script, and extracting a main cursor voltage value, an ISI noise amplitude and crosstalk noise; according to the main cursor voltage value, the ISI noise amplitude and the crosstalk noise, the final eye pattern eye height is calculated. According to the method, the problem that a traditional simulation method is insufficient in eye pattern height evaluation is solved, and more accurate and efficient simulation support can be provided for high-speed low-power-consumption DDR design.

Description

Technical Field

[0001] The present invention belongs to the technical field of low-power and high-speed DDR design, and in particular relates to an LPDDR4X simulation method and system based on statistical theory. Background Art

[0002] With the rapid development of mobile devices, the Internet of Things, and edge computing, Low-Power Double Data Rate (LPDDR) memory has become a core component in high-performance System-on-Chip (SoC) designs. LPDDR4X, an enhanced version of LPDDR4, significantly increases bandwidth while maintaining low power consumption by lowering operating voltage (e.g., to 0.6V core voltage) and increasing data rates (up to 4266Mbps). This makes it widely used in smartphones, tablets, and embedded systems. However, its high speed and low voltage characteristics place higher demands on signal integrity, particularly regarding eye height sensitivity, posing new challenges for simulation and design verification.

[0003] In LPDDR4X design, the eye diagram, a key metric for evaluating signal quality, directly reflects the timing and voltage margins for data transmission. Compared to traditional DDR4 or DDR5, LPDDR4X's eye height swing is significantly reduced, supporting only two standards: VOH50 and VOH60, corresponding to output voltage ranges of 250mV and 300mV, respectively. In contrast, the eye height swings for DDR4 and DDR5 typically exceed 600mV. While this low-swing design reduces power consumption, it also makes the eye height more sensitive to transmission loss, noise interference, and signal distortion. For example, insertion loss, reflections, or crosstalk in PCB traces can significantly compress the eye height, leading to an increase in the bit error rate (BER). Therefore, accurate signal integrity simulation is essential for LPDDR4X design, especially during signoff verification during the SoC packaging and PCB post-simulation stages.

[0004] Traditional signal integrity simulation methods are primarily based on time-domain analysis, evaluating the timing margin of a signal by measuring the eye width and combining it with simple noise models (such as peak-to-peak noise) to predict performance. This approach is relatively effective in scenarios with larger swings, such as DDR3 or DDR4, but it has limitations in the high-speed, low-voltage environment of LPDDR4X. Because eye height is much more sensitive to loss and noise than eye width, traditional methods that focus only on timing and ignore voltage margins cannot accurately capture the full picture of signal quality. For example, in high-speed data transmission, crosstalk noise may cause the eye height to be compressed below the minimum threshold of the receiver (such as 200mV), but the eye width may still meet the specification requirements, thereby masking potential problems.

[0005] Furthermore, traditional crosstalk analysis tools also have shortcomings in LPDDR4X scenarios. Frequency domain analysis methods, such as FEXT (far-end crosstalk) and NEXT (near-end crosstalk), can quantify the coupling effects between different signal lines, but their results are usually presented in the form of frequency response, which makes it difficult to directly map them to the impact of time domain eye height. The peak-to-peak noise measurement commonly used in time domain analysis, while intuitive, ignores the statistical distribution characteristics of crosstalk and the superposition effect of skew (timing offset) in signal transmission. For example, in the multi-channel parallel transmission of LPDDR4X, the skew between channels may amplify the cumulative effect of crosstalk noise, further compressing the eye height, and traditional simulation methods cannot effectively model this complex interaction.

[0006] In summary, the existing technologies for signal integrity simulation of LPDDR4X high-speed storage interfaces have the following major shortcomings:

[0007] 1. Ignoring Eye Height Assessment: Traditional simulation methods focus on eye width to assess timing margin, ignoring the impact of eye height. LPDDR4X eye height swings are only 250mV (VOH50) and 300mV (VOH60), significantly lower than the 600mV or more of DDR4 / 5. This makes it sensitive to losses, and traditional methods cannot accurately reflect this characteristic, resulting in significant deviations between simulation results and actual measurements.

[0008] 2. Insufficient quantification of crosstalk noise: Traditional tools such as frequency-domain FEXT / NEXT or time-domain peak-to-peak noise analysis cannot accurately quantify the impact of crosstalk on eye height, especially in low-swing scenarios, and fail to capture the statistical distribution characteristics of noise.

[0009] 3. Failure to consider skew differences: Existing methods ignore the superposition effect of skew (timing offset) in signal transmission, such as the accumulated noise in multi-channel transmission, which further weakens the ability to model eye height.

[0010] 4. Dependence on post-processing data, limited accuracy: Existing tools (such as HyperLynx) require complete netlist data support. When there is insufficient data in the early stages of design, the simulation results are inconsistent with actual measurements, making it difficult to guide optimization. Summary of the Invention

[0011] The purpose of this invention is to provide an LPDDR4X simulation method and system based on statistical theory. By constructing a time-domain simulation platform and combining pulse response and crosstalk noise analysis, the method accurately quantifies the factors affecting eye height and generates simulation results that are highly consistent with the measured height. This method addresses the problem that traditional simulation methods generally focus only on eye width while ignoring the influence of eye height. This leads to significant deviations between simulation results and measured data in high-speed LPDDR scenarios that are highly sensitive to eye diagrams. This method provides more accurate and efficient simulation support for high-speed, low-power DDR designs.

[0012] The present invention provides an LPDDR4X simulation method based on statistical theory, comprising the following steps:

[0013] Step 1: Build a time domain simulation platform for a single-byte BYTE for simulating the signal transmission characteristics of LPDDR4X; the signal in the single-byte BYTE includes a DQ data signal, a DM mask signal, and a DQS selection signal;

[0014] Step 2: Perform trace delay analysis on the package and PCB channel corresponding to each signal within a single byte to capture timing inconsistencies in signal transmission and, in turn, the impact of phase crosstalk noise on eye height.

[0015] Step 3: Perform impulse response analysis on the signal to quantify the ISI noise base caused by the package and PCB channel.

[0016] Step 4: Analyze the impact of crosstalk noise on each signal separately to obtain the direct impact of noise on eye height.

[0017] Step 5: Use MATLAB script to process the ISI noise basic data and direct impact data from steps 3 and 4, and extract the main cursor voltage value, ISI noise amplitude, and crosstalk noise;

[0018] Step 6: Calculate the final eye height of the eye diagram based on the main cursor voltage value, ISI noise amplitude, and crosstalk noise extracted in step 5.

[0019] Furthermore, the step 1 includes:

[0020] IO Buffer Modeling: Use IBIS models or SPICE models to describe the electrical characteristics of the transmitter and receiver to ensure that the simulation reflects the actual driving capability and impedance matching;

[0021] Channel modeling: Package and PCB channel scattering parameters are extracted through electromagnetic simulation, covering insertion loss, reflection loss and crosstalk effects.

[0022] Furthermore, the step 2 includes:

[0023] Delay measurement: Calculate the propagation delay of each signal from the transmitter to the receiver through time domain step waveform analysis;

[0024] Record differences: Record the timing offsets between signals in a table.

[0025] Furthermore, the step 3 includes:

[0026] Apply a single pulse excitation to the target signal with a pulse duration of one UI. For a 4266Mbps rate, one UI is 234.4ps. The rest of the signals remain in a static 0V state.

[0027] Record the time-voltage curve of the target signal with a sampling interval of 1 ps and a duration of 6 ns to ensure that all multiple reflection noise is covered, and generate a TXT file.

[0028] Furthermore, the step 4 includes:

[0029] The target signal remains static at 0V, and the other signals flip simultaneously, with the flip timing satisfying the timing offset difference recorded in step 2;

[0030] Record the time-voltage curve of the target signal with a sampling interval of 1 ps and a duration of 6 ns to ensure that all multiple reflection crosstalk noise is covered, and generate a TXT file.

[0031] Furthermore, the step 5 includes:

[0032] Import the TXT file of the impulse response in step 3 and the TXT file of the crosstalk noise in step 4, including the time-voltage curve;

[0033] Extract the main cursor voltage value from the TXT file of the impulse response, which represents the signal amplitude after loss;

[0034] Calculate the inter-symbol interference noise amplitude based on the pre-cursor and post-cursor in the TXT document based on the impulse response;

[0035] Based on the crosstalk noise waveform in the TXT document, use the main cursor position in impulse response analysis to calculate the noise values ​​of all pre-cursors and post-cursors, and add them together to form the statistical crosstalk noise.

[0036] Furthermore, the calculation formula for the eye diagram height in step 6 is as follows:

[0037] Eye diagram eye height = main cursor voltage value - ISI noise amplitude - crosstalk noise.

[0038] The present invention also provides an LPDDR4X simulation system based on statistical theory, including a simulation module, which executes the LPDDR4X simulation method based on statistical theory.

[0039] The present invention also provides a non-transitory computer-readable storage medium, which stores computer instructions. When the computer instructions are executed by a processor, the LPDDR4X simulation method based on statistical theory is implemented.

[0040] The present invention also provides an electronic device, comprising:

[0041] A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the LPDDR4X simulation method based on statistical theory by executing the computer instructions.

[0042] This approach, based on statistical theory, overcomes the limitations of traditional methods by developing a statistically-based LPDDR4X simulation method and system. This method analyzes the effect of crosstalk noise on eye diagram height using statistical methods and comprehensively considers skew differences in signal transmission. By integrating extensive measured data, this method accurately predicts and optimizes signal integrity, ensuring high consistency between simulation results and measured values. This method provides an efficient and accurate simulation tool for high-speed storage interface designs such as LPDDR4X and LPDDR5, and is widely applicable to signoff simulation in SOC packaging and PCB post-simulation stages.

[0043] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention and implement it according to the contents of the specification, the following is a detailed description of the preferred embodiments of the present invention with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 This is a flow chart of the LPDDR4X simulation method based on statistical theory of the present invention;

[0045] Figure 2 This is the fly-time of the routing timing in one embodiment of the present invention;

[0046] Figure 3 is an impulse response waveform in one embodiment of the present invention;

[0047] Figure 4 The crosstalk noise waveform in one embodiment of the present invention is:

[0048] Figure 5 The pulse response time-voltage sampling data in one embodiment of the present invention;

[0049] Figure 6 The crosstalk noise time-voltage sampling data in one embodiment of the present invention;

[0050] Figure 7 The MATLAB ISI noise analysis results in one embodiment of the present invention are shown below:

[0051] Figure 8 The MATLAB crosstalk noise analysis results in one embodiment of the present invention are as follows;

[0052] Figure 9 The figure is a fitting diagram of the simulation results and the measured results in one embodiment of the present invention; wherein, Figure a is the eye height margin parameter of the measured electronic eye diagram, and Figure b is the eye height margin parameter of the simulated eye diagram;

[0053] Figure 10 The figure is a schematic structural diagram of an electronic device of the present invention. DETAILED DESCRIPTION

[0054] The following embodiments of the present invention are described in further detail with reference to the accompanying drawings and examples. The following examples are used to illustrate the present invention but are not intended to limit the scope of the present invention.

[0055] Ginseng Figures 1 to 9 As shown in the figure, this embodiment provides an LPDDR4X simulation method based on statistical theory. It considers the impact of phase crosstalk noise on eye height, uses MATLAB scripts to integrate waveform data, extracts main cursor, ISI noise, and crosstalk noise parameters, calculates eye height based on statistical methods, accurately quantifies the impact of crosstalk noise on eye height, and distinguishes the contribution of different noise types to eye height. The specific steps are as follows:

[0056] Step S1, building a single BYTE simulation platform: building a time domain simulation platform for a single BYTE (including 8-bit data signal, 1-bit data mask signal, differential sampling signal DQS) for simulating the LPDDR4X signal transmission characteristics; the signals in the single BYTE include DQ data signal, DM mask signal and DQS selection signal. Including:

[0057] IO Buffer Modeling: Use the IBIS (Input / Output Buffer Information Specification) model or SPICE model to describe the electrical characteristics of the transmitter and receiver to ensure that the simulation reflects the actual driving capability and impedance matching;

[0058] Channel modeling: Package and PCB channel scattering parameters (S parameters) are extracted through electromagnetic simulation, covering insertion loss, reflection loss and crosstalk effects.

[0059] Step S2: Analyze signal timing differences: Perform trace delay analysis on the package and PCB channel corresponding to each signal (DQ, DM, DQS) within a single BYTE to capture timing inconsistencies in signal transmission and the impact of phase crosstalk noise on eye height. This includes:

[0060] Delay measurement: Calculate the propagation delay of each signal from the transmitter to the receiver through time domain step waveform analysis (e.g., 50 ps for DQ0 and 52 ps for DQ1).

[0061] Record differences: Record the timing skew between signals in a table (such as TXT format).

[0062] This step captures timing inconsistencies in signal transmission and accurately captures the impact of phase crosstalk noise on eye height.

[0063] Step S3, impulse response analysis: Perform impulse response analysis on the signal to quantify the basic data of ISI (Inter-Symbol Interference) noise caused by the package and PCB channel. This includes:

[0064] Test conditions: A single pulse excitation is applied to the target signal with a pulse duration of one UI. For a 4266 Mbps rate, one UI is 234.4 ps. The remaining signals remain in a static 0 V state.

[0065] Data recording: Record the time-voltage curve of the target signal with a sampling interval of 1 ps and a duration of 6 ns to ensure that all multiple reflection noise is covered, and generate a TXT file.

[0066] Step S4, crosstalk noise analysis: Analyze the impact of crosstalk noise on each signal individually (evaluate the multi-signal interaction effect) to obtain data on the direct impact of noise on eye height. This includes:

[0067] Test conditions: The target signal remains static at 0V, while the remaining signals flip simultaneously (for example, from 0V to 0.25V at a rate of 4266Mbps). The flip timing meets the timing skew difference recorded in step 2.

[0068] Data recording: Record the time-voltage curve of the target signal with a sampling interval of 1 ps and a duration of 6 ns to ensure that all multiple reflection crosstalk noise is covered, and generate a TXT file.

[0069] This step simulates the crosstalk scenario in real transmission and provides data on the direct impact of noise on eye height.

[0070] Step S5, MATLAB waveform analysis: Use MATLAB scripts to process the ISI noise basic data and direct impact data (waveform data) from steps S3 and S4, and extract the main cursor voltage value, ISI noise amplitude, and crosstalk noise. This includes:

[0071] Input data: Import the TXT file of the impulse response in step S3 and the TXT file of the crosstalk noise in step S4, including the time-voltage curve.

[0072] Parameter extraction:

[0073] Main Cursor Voltage: Extracts the main cursor voltage value from the impulse response TXT file, representing the signal amplitude after loss.

[0074] ISI Noise: Calculates the intersymbol interference (ISI) noise amplitude based on the pre-cursor and post-cursor in the TXT document of the impulse response.

[0075] Crosstalk noise: Based on the crosstalk noise waveform in a TXT document, use the main cursor position in impulse response analysis to calculate the noise values ​​of all pre-cursors and post-cursors, and add them together to form the statistical crosstalk noise.

[0076] Step S6, eye diagram height calculation: Calculate the final eye diagram eye height based on the main cursor voltage value, ISI noise amplitude, and crosstalk noise extracted in step S5. The calculation formula is as follows:

[0077] Eye height = Main Cursor voltage - ISI noise amplitude - Crosstalk noise.

[0078] This simulation method uses statistical methods to analyze the effect of crosstalk noise on eye diagram height and comprehensively considers skew differences in signal transmission, overcoming the limitations of traditional methods. By integrating a large amount of measured data, this method can accurately predict and optimize signal integrity, ensuring that simulation results are highly consistent with measured values. This method provides an efficient and accurate simulation tool for high-speed storage interface designs such as LPDDR4X and LPDDR5, and is widely applicable to signoff simulation in the SOC packaging and PCB post-simulation stages. The specific results are as follows:

[0079] 1) Improving eye height assessment accuracy: Statistically analyzing the impact of crosstalk noise on eye height compensates for the traditional method's neglect of eye height, ensuring that simulation results accurately reflect the signal integrity of LPDDR4X.

[0080] 2) Quantifying crosstalk and skew effects: By comprehensively considering the statistical characteristics of crosstalk and skew differences, we accurately model their effects on eye height, overcoming the limitations of traditional quantification tools.

[0081] 3) Enhanced consistency between simulation and actual measurement: Optimized simulation prediction capabilities by combining a large amount of actual measured data, providing a reliable basis for LPDDR4X and LPDDR5 design.

[0082] 4) Support early design optimization: Provide efficient tools in the SOC packaging and PCB post-simulation stages, reduce dependence on post-process data, and improve signoff simulation efficiency.

[0083] This embodiment also provides an LPDDR4X simulation system based on statistical theory, including a simulation module, which executes the LPDDR4X simulation method based on statistical theory.

[0084] This embodiment also provides a non-transitory computer-readable storage medium, which stores computer instructions. When the computer instructions are executed by a processor, they implement the LPDDR4X simulation method based on statistical theory.

[0085] Ginseng Figure 10 As shown, this embodiment further provides an electronic device, including:

[0086] The memory 201 and the processor 202 are communicatively connected to each other, the memory 201 stores computer instructions, and the processor 202 executes the LPDDR4X simulation method based on statistical theory by executing the computer instructions.

[0087] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A LPDDR4X simulation method based on statistical theory, characterized in that: The steps include: Step 1: Build a time domain simulation platform for a single-byte BYTE for simulating the signal transmission characteristics of LPDDR4X; the signal in the single-byte BYTE includes a DQ data signal, a DM mask signal, and a DQS selection signal; Step 2: Perform trace delay analysis on the package and PCB channel corresponding to each signal within a single byte to capture timing inconsistencies in signal transmission and, in turn, the impact of phase crosstalk noise on eye height. Step 3: Perform impulse response analysis on the signal to quantify the ISI noise base caused by the package and PCB channel. Step 4: Analyze the impact of crosstalk noise on each signal separately to obtain the direct impact of noise on eye height. Step 5: Use MATLAB script to process the ISI noise basic data and direct impact data from steps 3 and 4, and extract the main cursor voltage value, ISI noise amplitude, and crosstalk noise; Step 6: Calculate the final eye height of the eye diagram based on the main cursor voltage value, ISI noise amplitude, and crosstalk noise extracted in step 5.

2. The LPDDR4X simulation method based on statistical theory according to claim 1, characterized in that: The step 1 comprises: IO Buffer Modeling: Use IBIS models or SPICE models to describe the electrical characteristics of the transmitter and receiver to ensure that the simulation reflects the actual driving capability and impedance matching; Channel modeling: Package and PCB channel scattering parameters are extracted through electromagnetic simulation, covering insertion loss, reflection loss and crosstalk effects.

3. The LPDDR4X simulation method based on statistical theory according to claim 2, characterized in that: The step 2 includes: Delay measurement: Calculate the propagation delay of each signal from the transmitter to the receiver through time domain step waveform analysis; Record differences: Record the timing offsets between signals in a table.

4. The LPDDR4X simulation method based on statistical theory according to claim 3, characterized in that: The step 3 comprises: Apply a single pulse excitation to the target signal with a pulse duration of one UI. For a 4266Mbps rate, one UI is 234.4ps. The rest of the signals remain in a static 0V state. Record the time-voltage curve of the target signal with a sampling interval of 1 ps and a duration of 6 ns to ensure that all multiple reflection noise is covered, and generate a TXT file.

5. The LPDDR4X simulation method based on statistical theory according to claim 4, characterized in that: The step 4 comprises: The target signal remains static at 0V, and the other signals flip simultaneously, with the flip timing satisfying the timing offset difference recorded in step 2; Record the time-voltage curve of the target signal with a sampling interval of 1 ps and a duration of 6 ns to ensure that all multiple reflection crosstalk noise is covered, and generate a TXT file.

6. The LPDDR4X simulation method based on statistical theory according to claim 5, characterized in that: The step 5 comprises: Import the TXT file of the impulse response in step 3 and the TXT file of the crosstalk noise in step 4, including the time-voltage curve; Extract the main cursor voltage value from the TXT file of the impulse response, which represents the signal amplitude after loss; Calculate the inter-symbol interference noise amplitude based on the pre-cursor and post-cursor in the TXT document based on the impulse response; Based on the crosstalk noise waveform in the TXT document, use the main cursor position in impulse response analysis to calculate the noise values ​​of all pre-cursors and post-cursors, and add them together to form the statistical crosstalk noise.

7. The LPDDR4X simulation method based on statistical theory according to claim 6, characterized in that: The eye height calculation formula in step 6 is as follows: Eye diagram eye height = main cursor voltage value - ISI noise amplitude - crosstalk noise.

8. A LPDDR4X simulation system based on statistical theory, characterized in that: The method comprises a simulation module, wherein the simulation module executes the LPDDR4X simulation method based on statistical theory as described in any one of claims 1 to 7.

9. A non-transitory computer-readable storage medium, characterized in that The non-transitory computer-readable storage medium stores computer instructions, which, when executed by a processor, implement the LPDDR4X simulation method based on statistical theory according to any one of claims 1 to 7.

10. An electronic device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the LPDDR4X simulation method based on statistical theory as described in any one of claims 1 to 7 by executing the computer instructions.

Images