Signal processing method, signal transmission apparatus and interconnection interface
By dynamically adjusting the judgment level of the judgment feedback equalizer, the problem of inaccurate signal judgment in adaptive DFE in long-distance transmission and high-noise environments is solved, and the accuracy of signal transmission is improved.
Patent Information
- Application Number
- PCT/CN2024/087282
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-08
- Filing Date
- 2024-04-11
- Publication Date
- 2024-11-14
AI Technical Summary
In the prior art, due to the fixed judgment threshold, it is difficult for the adaptive DFE to accurately judge signals in long-distance transmission or high-noise environments, resulting in a decrease in signal transmission accuracy.
By obtaining the current noise tolerance value and prejudgment level in real time, dynamically adjust the judge level in the judgment feedback equalizer to ensure the correct judgment of the signal.
Improve the accuracy of signal transmission, avoid erroneous judgment problems caused by fixed judgment levels, and adapt to noise and signal amplitude changes.
Smart Images

Figure CN2024087282_14112024_PF_FP_ABST
Abstract
Description
Signal processing method, signal transmission device and interconnection interface
[0001] CROSS-REFERENCE TO RELATED APPLICATIONS
[0002] This application claims priority to the Chinese patent application filed with the China Patent Office on May 8, 2023, with application number 202310509047.X, and application name “Signal Processing Method, Signal Transmission Device and Interconnection Interface”, all contents of which are incorporated by reference into this application. Technical Field
[0003] The present application relates to the field of signal processing, and in particular to a signal processing method, a signal transmission device, and an interconnection interface. Background Art
[0004] As information rates continue to increase, signal integrity and bandwidth issues caused by channels are becoming increasingly serious. The combination of equalization technology and PAM4 (4-level Pulse Amplitude Modulation) signals has become one of the effective ways to solve this problem. Due to the unknown shape of channel characteristics, equalization technology usually uses an adaptive algorithm to update the tap coefficients in the equalizer, especially the decision feedback equalizer. Adaptive DFE (Decision Feedback Equalization) can only automatically track and compensate for changes in channel characteristics, but the decision threshold of its internal decider is fixed. The signal will be greatly distorted after long-distance transmission or in the presence of large noise. The fixed decision threshold will lead to erroneous decision result signals, reducing the accuracy of signal transmission.
[0005] Therefore, how to provide a solution to the above technical problems is a problem that those skilled in the art need to solve at present.
[0006] Summary of the Invention
[0007] The purpose of this application is to provide a signal processing method, a signal transmission device and an interconnection interface, which can track the changes in noise and signal amplitude in real time, automatically adjust the decision level of the decider in the feedback equalizer, ensure the correct decision of the signal, and improve the accuracy of signal transmission.
[0008] To solve the above technical problems, the present application provides a signal processing method, which includes:
[0009] When a current input signal is received, a current noise margin value and a current pre-determination level corresponding to the current input signal are obtained; the current input signal is the input signal received at the current moment;
[0010] Obtaining an actual decision level based on a current noise margin value and a current pre-decision level;
[0011] Updating the current decision level of the decision device in the decision feedback equalizer to the actual decision level, so that the decision device makes a decision on the current input signal based on the actual decision level to obtain a decision result signal;
[0012] The current output signal is obtained according to the judgment result signal.
[0013] In one embodiment of the present application, the process of obtaining the current noise margin value and the current pre-determination level corresponding to the current input signal includes:
[0014] Perform signal-noise separation on the current input signal to obtain a valid signal;
[0015] Get the amplitude of the effective signal;
[0016] The current noise margin value and the current pre-decision level are obtained based on the amplitude of the valid signal.
[0017] In one embodiment of the present application, before obtaining the current noise margin value and the current pre-determination level based on the amplitude of the valid signal, the signal processing method further includes:
[0018] Obtaining a preset relationship; the preset relationship is a corresponding relationship between a noise tolerance value, a pre-determined level, and an amplitude of a valid signal;
[0019] Accordingly, the process of obtaining the current noise margin value and the current pre-decision level based on the amplitude of the effective signal includes:
[0020] The current noise margin value and the current pre-determination level are obtained according to the amplitude of the effective signal and the preset relationship.
[0021] In one embodiment of the present application, the preset relationship is:
[0022] Among them, V q is the current noise tolerance value, V u is the amplitude, V p is the current pre-judgment level.
[0023] In one embodiment of the present application, the process of performing signal-to-noise separation processing on the current input signal to obtain a valid signal includes:
[0024] Perform signal-noise separation on the current input signal to obtain the effective signal and the noise signal;
[0025] Accordingly, the process of obtaining the actual decision level based on the current noise margin value and the current pre-decision level includes:
[0026] Obtain the noise difference between the amplitude of the noise signal and the current noise tolerance value;
[0027] The actual decision level is obtained based on the noise difference and the current pre-decision level.
[0028] In one embodiment of the present application, the process of obtaining the actual decision level based on the noise difference and the current pre-decision level includes:
[0029] The noise difference and the current pre-decision level are integrated to obtain the actual decision level.
[0030] In one embodiment of the present application, before obtaining the actual decision level based on the noise difference and the current pre-decision level, the signal processing method further includes:
[0031] Determining the positional relationship of each decision maker in the decision feedback equalizer;
[0032] Accordingly, the process of obtaining the actual decision level based on the noise difference and the current pre-decision level includes:
[0033] The actual decision level corresponding to each decider is obtained based on the noise difference, the current pre-decision level and the position relationship.
[0034] In one embodiment of the present application, the decision feedback equalizer includes three decision devices. The process of obtaining the actual decision level corresponding to each decision device based on the noise difference, the current pre-decision level, and the position relationship includes:
[0035] determining a first decider, a second decider, and a third decider among the three deciders according to the positional relationship; an actual decision level corresponding to the first decider is a first decision level, an actual decision level corresponding to the second decider is a second decision level, and an actual decision level corresponding to the third decider is a third decision level;
[0036] Obtaining a first decision level based on the noise difference and the current pre-decision level;
[0037] A third decision level is obtained based on the first decision level; the first decision level and the third decision level are opposite numbers to each other;
[0038] A second decision level is obtained based on the first decision level and the current pre-decision level.
[0039] In one embodiment of the present application, obtaining the second decision level based on the first decision level and the current pre-decision level includes:
[0040] The second decision level is determined based on a level difference between the first decision level and the current pre-decision level.
[0041] In one embodiment of the present application, the second decision level is determined based on the level difference between the first decision level and the current pre-decision level.
[0042] Obtaining a standard decision level for a second decision level;
[0043] Determining a level difference between the first decision level and the current pre-decision level;
[0044] The second decision level is determined based on the sum of the standard decision level and the level difference.
[0045] In one embodiment of the present application, the process of performing signal-to-noise separation processing on the current input signal to obtain a valid signal includes:
[0046] The current input signal is separated from the noise by wavelet transform to obtain the effective signal and the noise signal.
[0047] In one embodiment of the present application, the process of performing signal-noise separation on the current input signal by wavelet transform to obtain the effective signal and the noise signal includes:
[0048] After performing wavelet transform on the current input signal, wavelet coefficients smaller than a preset threshold are determined as the first signal, and wavelet coefficients greater than or equal to the preset threshold are determined as the second signal;
[0049] reconstructing the first signal to obtain a noise signal;
[0050] The second signal is reconstructed to obtain a valid signal.
[0051] In one embodiment of the present application, the process of obtaining the current output signal according to the judgment result signal includes:
[0052] Delaying the decision result signal to obtain a delayed output signal;
[0053] The delayed output signal is decoded to obtain a current output signal, which is a non-return-to-zero signal.
[0054] In one embodiment of the present application, the process of delaying the decision result signal to obtain a delayed output signal includes:
[0055] Performing multiple delay processing on the decision result signal to obtain multiple delayed output signals;
[0056] The signal processing method further includes:
[0057] The feedback signal is updated according to the decision result signal, the multiple delayed output signals and the coefficients in the preset coefficient table, so as to compensate the current input signal at the next moment through the feedback signal.
[0058] In one embodiment of the present application, when a current input signal is received, a process of obtaining a current noise margin value and a current pre-determination level corresponding to the current input signal includes:
[0059] When the current input signal is received, the target input signal is obtained after compensating the current input signal through the feedback signal;
[0060] Get the current noise margin value and the current pre-decision level corresponding to the target input signal.
[0061] In one embodiment of the present application, the process of delaying the decision result signal to obtain a delayed output signal includes:
[0062] The decision result signal is delayed by thermometer code rules to obtain a delayed output signal.
[0063] To solve the above technical problems, the present application further provides a signal transmission device, including a decision feedback equalization main circuit and an adaptive circuit, wherein the decision feedback equalization main circuit includes a first processing module, a second processing module and a decision module;
[0064] an adaptive circuit configured to, upon receiving a current input signal, obtain a current noise margin value and a current pre-decision level corresponding to the current input signal, obtain an actual decision level based on the current noise margin value and the current pre-decision level, and output the actual decision level to a decision module; the current input signal being the input signal received at the current moment;
[0065] A first processing module is configured to transmit the current input signal to the decision module upon receiving the current input signal;
[0066] The decision module is used to make a decision on the current input signal according to the actual decision level to obtain a decision result signal;
[0067] The second processing module is used to obtain and output a current output signal according to the judgment result signal.
[0068] In one embodiment of the present application, the adaptive circuit includes:
[0069] The third processing module is used to perform signal-noise separation on the current input signal to obtain a valid signal and a noise signal;
[0070] A calculation module, configured to obtain the amplitude of the effective signal, and obtain the current noise margin value and the current pre-decision level based on the amplitude of the effective signal;
[0071] The fourth processing module is configured to obtain an actual decision level based on the current noise margin value and the current pre-decision level, and output the actual decision level to the decider module.
[0072] In one embodiment of the present application, the fourth processing module includes:
[0073] a comparison unit, configured to receive the amplitude of the noise signal and the current noise tolerance value, calculate the noise difference between the amplitude of the noise signal and the current noise tolerance value, and output the calculated difference;
[0074] The level control unit is used to obtain the actual decision level based on the noise difference and the current pre-decision level and output it to the decision module.
[0075] In one embodiment of the present application, the fourth processing module further includes:
[0076] The delay unit is used to delay the current pre-decision level for a preset time and send it to the level control unit, so that the level control unit receives the current pre-decision level and the noise difference at the same time.
[0077] In an embodiment of the present application, the level control unit is an integrator.
[0078] In one embodiment of the present application, the decision module includes three decision devices, each of which is configured to make a decision on a current input signal according to an actual decision level corresponding to itself to obtain a decision result signal.
[0079] In one embodiment of the present application, the second processing module includes three delay unit groups, each delay unit group is composed of a plurality of sub-delay units connected in series;
[0080] Each sub-delay unit is used to delay the signal to be processed it receives to obtain a delayed output signal; the signal to be processed is a decision result signal or a delayed output signal output by a sub-delay unit whose output end is connected to its input end.
[0081] In one embodiment of the present application, the first processing module includes:
[0082] a feedback processing unit, configured to store a preset coefficient table including tap coefficients, and output a feedback signal based on the decision result signal, the delayed output signal, and the tap coefficients;
[0083] an adder, configured to compensate a current input signal based on a feedback signal to obtain a target input signal;
[0084] Accordingly, each decider is specifically configured to make a decision on the target input signal according to an actual decision level corresponding to itself to obtain a decision result signal.
[0085] In one embodiment of the present application, the second processing module includes:
[0086] The decoder is used to decode the delayed output signal output by the last sub-delay unit to obtain a current output signal, which is a non-return-to-zero signal.
[0087] In one embodiment of the present application, the decoder is a thermometer decoder.
[0088] In order to solve the above technical problems, the present application also provides a chip interconnection interface, including a signal transmission device as described in any one of the above.
[0089] To solve the above technical problems, the present application also provides an electronic device, comprising: at least one processor and a memory;
[0090] Memory stores computer-executable instructions;
[0091] At least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the above method.
[0092] To solve the above technical problems, the present application also provides a non-volatile readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the above method is implemented.
[0093] The present application provides a signal processing method that first determines the current noise tolerance value and the current pre-decision level based on the current input signal received, and then updates the current decision level of the decision maker in the decision feedback equalizer based on the actual decision level obtained from the current noise tolerance value and the current pre-decision level. This method can track changes in noise and signal amplitude in real time, automatically adjust the decision level of the decision maker in the feedback equalizer, ensure correct signal judgment, and improve the accuracy of signal transmission. The present application also provides a signal transmission device and a core interconnection interface that have the same beneficial effects as the above-mentioned signal processing method. BRIEF DESCRIPTION OF THE DRAWINGS
[0094] In order to more clearly illustrate the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.
[0095] FIG1 is a flowchart of the steps of a signal processing method provided by the present application;
[0096] FIG2 is a schematic diagram of a two-level decision under different noise amplitudes provided by the present application;
[0097] FIG3 is a schematic diagram of a two-level decision in long-distance transmission provided by the present application;
[0098] FIG4 is a schematic structural diagram of a signal transmission device provided by the present application;
[0099] FIG5 is a schematic diagram of the structure of an adaptive circuit provided by the present application;
[0100] FIG6 is a schematic diagram of an electronic device provided by the present application;
[0101] FIG7 is a schematic diagram of a non-volatile readable storage medium provided in this application. DETAILED DESCRIPTION
[0102] The core of this application is to provide a signal processing method, a signal transmission device and an interconnection interface, which can track the changes in noise and signal amplitude in real time, automatically adjust the decision level of the decider in the feedback equalizer, ensure the correct decision of the signal, and improve the accuracy of signal transmission.
[0103] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.
[0104] In the first aspect, please refer to FIG1 , which is a flowchart of the steps of a signal processing method provided by the present application. The signal processing method includes:
[0105] S101: When a current input signal is received, a current noise margin value and a current pre-determination level corresponding to the current input signal are obtained; the current input signal is the input signal received at the current moment;
[0106] S102: Obtaining an actual decision level based on a current noise margin value and a current pre-decision level;
[0107] Taking into account that the current input signal received may be attenuated due to long-distance transmission, or include noise, these reasons will affect the subsequent signal judgment. In order to ensure accurate judgment of the current input signal, this step obtains the current noise tolerance value and the current pre-judgment level corresponding to the current input signal after obtaining the current input signal, and calculates the actual judgment level based on the current pre-judgment level and the current noise tolerance value. It can be understood that if the noise amplitude in the current input signal does not exceed the current noise tolerance value, it means that the current pre-judgment level can correctly judge the current input signal. If the noise amplitude in the current input signal exceeds the current noise tolerance value, it means that the current pre-judgment level cannot accurately judge the current input signal, and it is necessary to adjust the current pre-judgment level based on the current noise tolerance value and noise amplitude to obtain the actual judgment level.
[0108] S103: updating the current decision level of the decider in the decision feedback equalizer to the actual decision level, so that the decider makes a decision on the current input signal based on the actual decision level to obtain a decision result signal;
[0109] S104: Obtain a current output signal according to the judgment result signal.
[0110] Specifically, the decision feedback equalizer includes a decider, and the current decision level of the decider is updated according to the actual decision level obtained in the above steps, so that the decider uses the actual decision level to make a decision on the current input signal to obtain a decision result signal. The decision result signal can be processed to obtain the current output signal of the decision feedback equalizer. This embodiment can track the changes in noise and signal amplitude in real time, automatically adjust the decision level of the decider in the feedback equalizer, ensure the correct decision of the signal, and improve the accuracy of signal transmission.
[0111] Based on the above embodiment:
[0112] As an embodiment, the process of obtaining the current noise margin value and the current pre-decision level corresponding to the current input signal includes:
[0113] Perform signal-noise separation on the current input signal to obtain a valid signal;
[0114] Get the amplitude of the effective signal;
[0115] The current noise margin value and the current pre-decision level are obtained based on the amplitude of the valid signal.
[0116] First, the current input signal V in Perform signal-noise separation to obtain the effective signal V u and noise signal V n , then the effective signal V u The amplitude of the signal is detected by the effective signal V u The current noise tolerance value V can be determined by the amplitude of q and the current pre-decision level V p .
[0117] As an embodiment, before obtaining the current noise margin value and the current pre-determination level based on the amplitude of the valid signal, the signal processing method further includes:
[0118] Obtaining a preset relationship; the preset relationship is a corresponding relationship between a noise tolerance value, a pre-determined level, and an amplitude of a valid signal;
[0119] Accordingly, the process of obtaining the current noise margin value and the current pre-decision level based on the amplitude of the effective signal includes:
[0120] The current noise margin value and the current pre-determination level are obtained according to the amplitude of the effective signal and the preset relationship.
[0121] As an embodiment, the preset relationship is:
[0122] Among them, V q is the current noise tolerance value, V u is the amplitude, V p is the current pre-judgment level.
[0123] Specifically, the corresponding relationship between the noise tolerance value, the pre-determined level and the amplitude of the effective signal is pre-established, that is, the preset relationship. u After the amplitude is calculated, the current noise tolerance value V can be calculated according to the preset relationship. q and the current pre-decision level V p .
[0124] As an embodiment, the process of performing signal-to-noise separation processing on the current input signal to obtain a valid signal includes:
[0125] Perform signal-noise separation on the current input signal to obtain the effective signal and the noise signal;
[0126] Accordingly, the process of obtaining the actual decision level based on the current noise margin value and the current pre-decision level includes:
[0127] Obtain the noise difference between the amplitude of the noise signal and the current noise tolerance value;
[0128] The actual decision level is obtained based on the noise difference and the current pre-decision level.
[0129] Specifically, the noise signal V n The comparator compares the two and inputs the noise difference ΔV between them. Then, the current pre-decision level V is determined based on the noise difference ΔV. p Adjust to get the actual decision level V c The sign of the noise difference ΔV represents the adjustment direction, i.e., increasing the current pre-decision level Vp or decreasing the current pre-decision level V p , the absolute value of the noise difference ΔV and the current pre-decision level V p Together they determine the size of the adjustment step.
[0130] As an embodiment, the process of obtaining the actual decision level based on the noise difference and the current pre-decision level includes:
[0131] The noise difference and the current pre-decision level are integrated to obtain the actual decision level.
[0132] As an embodiment, before obtaining the actual decision level based on the noise difference and the current pre-decision level, the signal processing method further includes:
[0133] Determining the positional relationship of each decision maker in the decision feedback equalizer;
[0134] Accordingly, the process of obtaining the actual decision level based on the noise difference and the current pre-decision level includes:
[0135] The actual decision level corresponding to each decider is obtained based on the noise difference, the current pre-decision level and the position relationship.
[0136] As an embodiment, the decision feedback equalizer includes three decision devices. The process of obtaining the actual decision level corresponding to each decision device based on the noise difference, the current pre-decision level, and the position relationship includes:
[0137] determining a first decider, a second decider, and a third decider among the three deciders according to the positional relationship; an actual decision level corresponding to the first decider is a first decision level, an actual decision level corresponding to the second decider is a second decision level, and an actual decision level corresponding to the third decider is a third decision level;
[0138] Obtaining a first decision level based on the noise difference and the current pre-decision level;
[0139] A third decision level is obtained based on the first decision level; the first decision level and the third decision level are opposite numbers to each other;
[0140] A second decision level is obtained based on the first decision level and the current pre-decision level.
[0141] As an embodiment, the second decision level is determined based on the difference between the first decision level and the current pre-decision level. Specifically, a standard decision level for the second decision level can be obtained; the level difference between the first decision level and the current pre-decision level can be determined; and the second decision level can be determined based on the sum of the standard decision level and the level difference.
[0142] Specifically, the decision feedback equalizer in this embodiment includes three decision devices, namely a first decision device, a second decision device and a third decision device. The first decision level of the first decision device is V t The second decision level of the second decision device is V m , the third decision level of the third decision device is V b , according to the position relationship of the standard decision level, assuming that V b =-V t , V t and V b The update is based on the actual decision level V c , Vt=V c , V b =-V c , V m The standard decision level is 0, V m The update is based on the actual decision level V c With the pre-decision level V pThe difference ΔV, V m is 0±ΔV.
[0143] As an embodiment, the process of performing signal-to-noise separation processing on the current input signal to obtain a valid signal includes:
[0144] The current input signal is separated from the noise by wavelet transform to obtain the effective signal and the noise signal.
[0145] As an embodiment, the process of performing signal-noise separation on the current input signal through wavelet transform to obtain a valid signal and a noise signal includes:
[0146] After performing wavelet transform on the current input signal, wavelet coefficients smaller than a preset threshold are determined as the first signal, and wavelet coefficients greater than or equal to the preset threshold are determined as the second signal;
[0147] Reconstructing the first signal to obtain a noise signal;
[0148] The second signal is reconstructed to obtain a valid signal.
[0149] In this embodiment, wavelet transform and reconstruction use sampling and interpolation methods to separate the effective signal and noise signal in the current input signal according to a preset threshold. After the current input signal is transformed by wavelet, the wavelet coefficients less than the preset threshold are regarded as noise signals, while the wavelet coefficients greater than the preset threshold are regarded as effective signals. In this way, the separation of signal and noise is achieved, and wavelet reconstruction can be performed to obtain the filtered effective signal V u and noise signal V n Wavelet transform and signal reconstruction technology are used to replace the resistor voltage divider technology in traditional PAM4 DFE, thus avoiding voltage ripple and voltage deviation caused by resistance.
[0150] As an embodiment, the process of obtaining the current output signal according to the judgment result signal includes:
[0151] Delaying the decision result signal to obtain a delayed output signal;
[0152] The delayed output signal is decoded to obtain a current output signal, which is a non-return-to-zero signal.
[0153] As an embodiment, the process of delaying the decision result signal to obtain a delayed output signal includes:
[0154] Performing multiple delay processing on the decision result signal to obtain multiple delayed output signals;
[0155] The signal processing method further includes:
[0156] The feedback signal is updated according to the decision result signal, the multiple delayed output signals and the coefficients in the preset coefficient table, so as to compensate the current input signal at the next moment through the feedback signal.
[0157] As an embodiment, when a current input signal is received, a process of obtaining a current noise margin value and a current pre-determination level corresponding to the current input signal includes:
[0158] When the current input signal is received, the target input signal is obtained after compensating the current input signal through the feedback signal;
[0159] Get the current noise margin value and the current pre-decision level corresponding to the target input signal.
[0160] As an embodiment, the process of delaying the decision result signal to obtain a delayed output signal includes:
[0161] The decision result signal is delayed by thermometer code rules to obtain a delayed output signal.
[0162] Specifically, when the feedback signal V fed When the initial level is 0, the target input signal V x With the current input signal V in Same, the target input signal V x After the judgment of 3 judges, if V x If the value is higher than the decision level, the decision device outputs a high level 1. If the value is lower than the decision level, the decision device outputs a low level 0, forming a 3-bit decision result signal, including the output data [V t,1 , V m,1 , V b,1 ], every time a clock cycle passes, data is continuously sampled and transmitted. After j clock cycles, the i-th delayed output signal includes the output data [V t,j , V m,j , V b,j ](j≤i).
[0163] Then, the output data of the decision maker and delay unit [V t,i , V m,i , V b,i ](i=1,2,3) and the coefficients in the coefficient table are weighted to update the feedback signal V fed , and at the next moment transmit data V in To compensate, the target input signal V x =V in -V fed .
[0164] For example, the adaptive decision implementation process is described with reference to FIG2 and FIG3 , and there are two implementation scenarios: one is different noise amplitudes, as shown in FIG2 ; the other is long-distance transmission, as shown in FIG3 .
[0165] In the first case: with a value close to V u Taking the effective signal level of / 3 as an example, when the noise amplitude is less than the noise tolerance value and the noise amplitude is positive, as shown in ① in Figure 2, after the effective signal level and the noise are superimposed, the signal amplitude is still less than the pre-judgment level, then the input signal can be correctly judged, that is, the judgment value is V u / 3; If the noise amplitude is greater than the noise tolerance value and the noise amplitude is positive, as shown in ② in Figure 2, the signal amplitude after the effective signal level and the noise are superimposed is greater than the pre-decision level. If the decision level is not changed, the input signal is wrongly judged, that is, the decision value is V u By adjusting the difference ΔV between the noise margin value and the noise amplitude and the sign of the difference ΔV, the decision level is raised by ΔV on the basis of the pre-decision level. Then the input signal can be correctly judged, that is, the decision value is V u / 3.
[0166] Close to V u Taking the effective signal level as an example, when the noise amplitude is greater than the noise tolerance value and the noise amplitude is negative, as shown in ③ in Figure 2, the signal amplitude after the effective signal level and the noise are superimposed is less than the pre-decision level. The difference ΔV between the noise tolerance value and the noise amplitude and the sign of the difference ΔV make the decision level increase or decrease by ΔV based on the pre-decision level. Then the input signal can be correctly judged, that is, the decision value is V u .
[0167] In the second case, when the noise amplitude remains unchanged, long-distance transmission will cause the effective signal amplitude to decrease, resulting in a decrease in the signal-to-noise margin, as shown in Figure 3, the change from dotted line ① to solid line ②.
[0168] Close to V u Taking the effective signal level as an example, when the effective signal amplitude drops to a certain extent and the noise amplitude is negative, as shown in Figure 3, the signal amplitude after the effective signal level and the noise are superimposed is less than the pre-decision level. The difference ΔV between the noise tolerance value and the noise amplitude and the sign of the difference ΔV make the decision level increase or decrease by ΔV based on the pre-decision level. Then the input signal can be correctly judged, that is, the decision value is V u .
[0169] In summary, the present application not only avoids the influence of voltage ripple and the erroneous judgment problem caused by the fixed judgment level under the traditional method, but also can track the changes in signal amplitude or noise amplitude in real time to realize automatic adjustment of the judgment level, thereby improving the accuracy of signal transmission.
[0170] In the second aspect, please refer to FIG4 , which is a schematic structural diagram of a signal transmission device provided by the present application. The signal transmission device includes a decision feedback equalization main circuit 41 and an adaptive circuit 42. The decision feedback equalization main circuit 41 includes a first processing module, a second processing module, and a decision module.
[0171] Adaptive circuit 42 is configured to, upon receiving a current input signal, obtain a current noise margin value and a current pre-decision level corresponding to the current input signal, determine an actual decision level based on the current noise margin value and the current pre-decision level, and output the actual decision level to the decision module; the current input signal is the input signal received at the current moment;
[0172] A first processing module is configured to transmit the current input signal to the decision module upon receiving the current input signal;
[0173] The decision module is used to make a decision on the current input signal according to the actual decision level to obtain a decision result signal;
[0174] The second processing module is used to obtain and output a current output signal according to the judgment result signal.
[0175] It can be seen that in this embodiment, the current noise tolerance value and the current pre-judgment level are first determined based on the current input signal received, and the current decision level of the decider in the signal transmission device is updated based on the actual decision level obtained based on the current noise tolerance value and the current pre-judgment level. It can track the changes in noise and signal amplitude in real time, automatically adjust the decision level of the decider in the feedback equalizer, ensure the correct judgment of the signal, and improve the accuracy of signal transmission.
[0176] As an embodiment, please refer to FIG5 , which is a schematic diagram of the structure of an adaptive circuit provided by the present application, including:
[0177] The third processing module 421 is used to perform signal-noise separation on the current input signal to obtain a valid signal and a noise signal;
[0178] A calculation module 422 is configured to obtain an amplitude of a valid signal, and obtain a current noise margin value and a current pre-decision level based on the amplitude of the valid signal;
[0179] The fourth processing module is configured to obtain an actual decision level based on the current noise margin value and the current pre-decision level, and output the actual decision level to the decider module.
[0180] As an embodiment, the fourth processing module includes:
[0181] a comparing unit 423, configured to receive the amplitude of the noise signal and the current noise tolerance value, calculate a noise difference between the amplitude of the noise signal and the current noise tolerance value, and output the calculated difference;
[0182] The level control unit 424 is configured to obtain an actual decision level based on the noise difference and the current pre-decision level and output the actual decision level to the decision module.
[0183] As an embodiment, the fourth processing module further includes:
[0184] The delay unit 425 is configured to delay the current pre-determined level for a preset time and send it to the level control unit 424 so that the level control unit 424 receives the current pre-determined level and the noise difference at the same time.
[0185] As an embodiment, the level control unit 424 is an integrator.
[0186] Specifically, the adaptive circuit includes a third processing module 421, a calculation module 422, a comparison unit 423, a level control unit 424 and a delay unit 425. First, the third processing module 421 uses sampling and interpolation methods to separate the effective signal and the noise signal according to the set threshold. After the current input signal is transformed by wavelet, when the wavelet coefficient is less than the threshold, it is regarded as a noise signal, and the wavelet coefficient is greater than the threshold, it is regarded as a valid signal. In this way, the separation of signal and noise is achieved, and then wavelet reconstruction can be performed to obtain the filtered effective signal V u and noise signal V n Then, the calculation module 422 performs amplitude detection on the filtered signal and calculates the amplitude of the signal according to the pre-determined level V p , noise tolerance value V q With the effective signal V u The corresponding relationship, output V p and V q . (Pre-decision level V p , noise tolerance value V q With the effective signal V u The relationship is V q =V u / 3 and V p =2V u / 3), the comparison unit 423 compares V q and V n The amplitude is compared and the difference ΔV between the two is output. The decision level is regulated based on the pre-decision level V c , thereby obtaining the actual decision level of each decision device in the decision feedback equalization main circuit 41. The sign of the difference ΔV represents the coefficient adjustment direction of the level control module, and the absolute value of the difference ΔV and V p Together they determine the size of the adjustment stride.
[0187] As an embodiment, the decision module includes three decision units, each of which is used to make a decision on the current input signal according to its own actual decision level to obtain a decision result signal. t Indicates the actual decision level of the first decision maker, expressed in V m Indicates the actual decision level of the second decision maker, expressed in V b Represents the actual decision level of the third decision maker, which is used to realize the weighted signal V x level judgment.
[0188] As an embodiment, the second processing module includes three delay unit groups, each of which is composed of a plurality of sub-delay units connected in series. In FIG. 4 , each sub-delay unit is represented by T.
[0189] Each sub-delay unit is used to delay the signal to be processed it receives to obtain a delayed output signal; the signal to be processed is a decision result signal or a delayed output signal output by a sub-delay unit whose output end is connected to its input end.
[0190] As an embodiment, the first processing module includes:
[0191] A feedback processing unit 411 is configured to store a preset coefficient table including tap coefficients and output a feedback signal based on the decision result signal, the delayed output signal, and the tap coefficients;
[0192] an adder 412 for compensating the current input signal based on the feedback signal to obtain a target input signal;
[0193] Accordingly, each decider is specifically configured to make a decision on the target input signal according to an actual decision level corresponding to itself to obtain a decision result signal.
[0194] Specifically, the adder 412 implements the input signal V in and feedback compensation signal V fed The preset coefficient table stores the tap coefficients, and the feedback processing unit 411 implements weighted processing of the delay unit output signal and the coefficients.
[0195] As an embodiment, the second processing module includes:
[0196] The decoder 413 is configured to decode the delayed output signal output by the last sub-delay unit to obtain a current output signal, which is a non-return-to-zero signal.
[0197] As an embodiment, the decoder 413 is a thermometer decoder.
[0198] Specifically, the thermometer decoder implements a 3b-2b decoding function, converting the delayed output signals output by the three-way delay unit group into two-way signals D0 and D1 for output.
[0199] First, when the feedback compensation signal V fed When the initial level is 0, the output signal V x With the current input signal V in Same. x After the judgment of 3 judges, if V x If the value is higher than the decision level, the decision device outputs a high level 1. If the value is lower than the decision level, the decision device outputs a low level 0, forming the initial data of 3 bits [V t ,1,V m ,1,V b ,1], every time a clock cycle passes, data is continuously sampled and transmitted. After 3 clock cycles (j≤i), the output data of the third delay unit is [V t ,3,V m ,3,V b ,3].
[0200] Then, the output data of the decision maker and the sub-delay unit [V t ,i,V m ,i,V b ,i](i=1,2,3) and the coefficients in the coefficient table are weighted to update the feedback signal V fed . And at the next moment, the transmission data V in To compensate, the output signal of the adder V x =V in -V fed .
[0201] In summary, this application designs a high-speed, three-tap PAM4 DFE with adaptive decision levels based on CMOS technology using wavelet transform, signal reconstruction, and threshold adjustment techniques. Wavelet transform and signal reconstruction replace the resistor divider technology used in traditional PAM4 DFE, avoiding voltage ripple and voltage deviation caused by resistors. Furthermore, threshold adjustment technology tracks changes in noise and signal amplitude, automatically adjusting the decision level to ensure accurate signal decisions and thus suppress DFE error propagation. This makes it fully applicable to various chiplet interconnect interfaces, such as C2C and D2D.
[0202] In a third aspect, the present application also provides a chip interconnection interface, including a signal transmission device as described above.
[0203] Among them, the chiplet interconnection interface includes but is not limited to various chiplet interconnection interfaces such as C2C (Chip-to-Chip, chip to chip), D2D (Die-to-Die, bare die) and so on.
[0204] For an introduction to a chip interconnection interface provided in this application, please refer to the above embodiment, and this application will not go into details here.
[0205] The chip interconnection interface provided in this application has the same beneficial effects as the above-mentioned signal processing method.
[0206] FIG6 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device 60 includes at least one processor 61 and a memory 62 .
[0207] Memory stores computer-executable instructions;
[0208] At least one processor executes the computer-implemented instructions stored in the memory, causing the at least one processor to perform a method such as signal processing.
[0209] 7 , an embodiment of the present application provides a non-volatile readable storage medium. The non-volatile readable storage medium 700 stores computer-executable instructions 701 . When a processor executes the computer-executable instructions, a signal processing method is implemented.
[0210] It should also be noted that, in this specification, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus comprising the element.
[0211] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A signal processing method, characterized in that: The signal processing method comprises: When a current input signal is received, a current noise margin value and a current pre-determination level corresponding to the current input signal are obtained; the current input signal is an input signal received at the current moment; Obtaining an actual decision level based on the current noise tolerance value and the current pre-decision level; Updating the current decision level of the decision device in the decision feedback equalizer to the actual decision level, so that the decision device makes a decision on the current input signal based on the actual decision level to obtain a decision result signal; A current output signal is obtained according to the decision result signal.
2. The signal processing method according to claim 1, characterized in that: The process of obtaining the current noise margin value and the current pre-decision level corresponding to the current input signal comprises: Performing signal-to-noise separation processing on the current input signal to obtain a valid signal; Acquiring the amplitude of the effective signal; A current noise margin value and a current pre-decision level are acquired based on the amplitude of the effective signal.
3. The signal processing method according to claim 2, characterized in that: Before acquiring the current noise margin value and the current pre-determination level based on the amplitude of the effective signal, the signal processing method further includes: Acquire a preset relationship; the preset relationship is a corresponding relationship between a noise tolerance value, a pre-determined level, and an amplitude of a valid signal; Correspondingly, the process of obtaining the current noise tolerance value and the current pre-decision level based on the amplitude of the effective signal includes: A current noise margin value and a current pre-determination level are obtained according to the amplitude of the effective signal and the preset relationship.
4. The signal processing method according to claim 3, characterized in that: The preset relationship is: Among them, V q is the current noise tolerance value, V u is the amplitude, V p is the current pre-decision level.
5. The signal processing method according to claim 2, characterized in that: The process of performing signal-to-noise separation processing on the current input signal to obtain a valid signal includes: Performing signal-noise separation on the current input signal to obtain a valid signal and a noise signal; Accordingly, the process of obtaining the actual decision level based on the current noise tolerance value and the current pre-decision level includes: Obtaining a noise difference between the amplitude of the noise signal and the current noise tolerance value; An actual decision level is obtained based on the noise difference and the current pre-decision level.
6. The signal processing method according to claim 5, characterized in that: The process of obtaining the actual decision level based on the noise difference and the current pre-decision level includes: The noise difference and the current pre-decision level are integrated to obtain an actual decision level.
7. The signal processing method according to claim 5, characterized in that: Before obtaining the actual decision level based on the noise difference and the current pre-decision level, the signal processing method further includes: Determining the positional relationship of each of the decision devices in the decision feedback equalizer; Correspondingly, the process of obtaining the actual decision level based on the noise difference and the current pre-decision level includes: An actual decision level corresponding to each of the decision devices is obtained based on the noise difference, the current pre-decision level and the position relationship.
8. The signal processing method according to claim 7, characterized in that: The decision feedback equalizer includes three decision devices; The process of obtaining the actual decision level corresponding to each of the decision devices based on the noise difference, the current pre-decision level and the position relationship includes: Determine a first determiner, a second determiner, and a third determiner among the three determiners according to the position relationship; the actual decision level corresponding to the first determiner is a first decision level, the actual decision level corresponding to the second determiner is a second decision level, and the actual decision level corresponding to the third determiner is a third decision level; Obtaining the first decision level based on the noise difference and the current pre-decision level; The third decision level is obtained based on the first decision level; the first decision level and the third decision level are opposite numbers to each other; The second decision level is obtained based on the first decision level and the current pre-decision level.
9. The signal processing method according to claim 8, characterized in that: The obtaining the second decision level based on the first decision level and the current pre-decision level comprises: The second decision level is determined based on a level difference between the first decision level and the current pre-decision level.
10. The signal processing method according to claim 9, characterized in that: The second decision level is determined based on a level difference between the first decision level and the current pre-decision level. Acquire a standard decision level of the second decision level; Determining a level difference between the first decision level and the current pre-decision level; The second decision level is determined based on the sum of the standard decision level and the level difference.
11. The signal processing method according to claim 2, characterized in that: The process of performing signal-to-noise separation processing on the current input signal to obtain a valid signal includes: The current input signal is subjected to signal-noise separation by wavelet transformation to obtain a valid signal and a noise signal.
12. The signal processing method according to claim 11, characterized in that: The process of performing signal-noise separation on the current input signal by wavelet transform to obtain a valid signal and a noise signal comprises: After performing wavelet transform on the current input signal, wavelet coefficients less than a preset threshold are determined as the first signal, and wavelet coefficients greater than or equal to the preset threshold are determined as the second signal; Reconstructing the first signal to obtain a noise signal; The second signal is reconstructed to obtain a valid signal.
13. The signal processing method according to any one of claims 1 to 12, characterized in that: The process of obtaining the current output signal according to the judgment result signal includes: Performing delay processing on the decision result signal to obtain a delayed output signal; The delayed output signal is decoded to obtain the current output signal, where the current output signal is a non-return-to-zero signal.
14. The signal processing method according to claim 13, characterized in that: The process of delaying the decision result signal to obtain a delayed output signal includes: Performing multiple delay processing on the decision result signal to obtain multiple delayed output signals; The signal processing method further includes: The feedback signal is updated according to the decision result signal, the plurality of delayed output signals and the coefficients in the preset coefficient table, so as to compensate the current input signal at the next moment through the feedback signal.
15. The signal processing method according to claim 14, characterized in that: When a current input signal is received, the process of obtaining a current noise margin value and a current pre-decision level corresponding to the current input signal includes: When a current input signal is received, the current input signal is compensated by the feedback signal to obtain a target input signal; A current noise margin value and a current pre-decision level corresponding to the target input signal are obtained.
16. The signal processing method according to claim 13, characterized in that: The process of delaying the decision result signal to obtain a delayed output signal includes: The decision result signal is delayed by using a thermometer code rule to obtain a delayed output signal.
17. A signal transmission device, characterized in that: It includes a decision feedback equalization main circuit and an adaptive circuit, wherein the decision feedback equalization main circuit includes a first processing module, a second processing module and a decision module; The adaptive circuit is used for, when receiving a current input signal, obtaining a current noise tolerance value and a current pre-decision level corresponding to the current input signal, obtaining an actual decision level based on the current noise tolerance value and the current pre-decision level, and outputting the actual decision level to the decision module; the current input signal is an input signal received at the current moment; The first processing module is used for transmitting the current input signal to the decision module when receiving the current input signal; The decision module is used to make a decision on the current input signal according to the actual decision level to obtain a decision result signal; The second processing module is used to obtain and output a current output signal according to the judgment result signal.
18. A chip interconnection interface, characterized in that: Comprising a signal transmission device as described in any one of claim 17.
19. An electronic device, characterized in that: include: at least one processor and memory; The memory stores computer-executable instructions; The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the method according to any one of claims 1 to 16.
20. A non-volatile readable storage medium, characterized in that: The non-volatile readable storage medium stores computer-executable instructions, and when the processor executes the computer-executable instructions, the method according to any one of claims 1 to 16 is implemented.
Citation Information
Patent Citations
Method for optimization regulation of receiver decision level and apparatus thereof
CN101355389A
Adaptive PAM4 decision feedback equalization circuit
CN113595949A
Signal judgment equalization method and device
CN115550115A
Signal processing method, signal transmission device and interconnection interface
CN116232816A
Decision-feedback equalizer with alarm ciruit
EP0732831A2
