A method and device for noise compensation of a transimpedance amplifier

By using feedforward equalizer and filter in the optical module to calculate and compensate the noise signal, the problem of noise introduced by the transimpedance amplifier affecting the performance of the optical module is solved, and the effect of improving transmission performance and reducing the pre-correction bit error rate is achieved.

CN118554892BActive Publication Date: 2024-12-03CORE TREND (ZHUHAI) TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411016104.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-27
Publication Date
2024-12-03
Estimated Expiration
2044-07-27

AI Technical Summary

Technical Problem

In high-speed optical modules, the noise introduced by the transimpedance amplifier causes the transmission performance of the optical module to degrade, and the prior art is difficult to effectively compensate for noise, especially in some gain settings, which seriously affects the performance of the optical module.

Method used

Using a combination of feedforward equalizer and filter, the analog front-end output signal is received for equalization processing, the error signal is calculated and filtered to estimate the noise signal, and compensation is performed before or after the feedforward equalizer according to the noise characteristics, using a specific Iterative algorithms and filter formulas calculate noise signals for accurate compensation.

Benefits of technology

It effectively improves the transmission performance of optical modules, reduces the forward error correction code rate and error symbol number, and is suitable for various TIA devices, improving the overall performance of the system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118554892B_ABST
    Figure CN118554892B_ABST
Patent Text Reader

Abstract

The embodiment of the present disclosure provides a noise compensation method and device for a transimpedance amplifier. The noise compensation method includes a feedforward equalizer receiving an output signal of an analog front end and outputting an equalized output signal; a symbol number 2 of a coding scheme based on pulse amplitude modulation i and the equalized output signal to calculate the error signal, where i is a positive integer; use a filter to filter the error signal to obtain an estimated noise signal; when compensation is performed before the feedforward equalizer, at the input end of the feedforward equalizer, the estimated noise signal is added to the output signal of the analog front end to serve as the corrected input signal of the feedforward equalizer; when compensation is performed after the feedforward equalizer, the estimated noise signal is added to the output signal of the feedforward equalizer to obtain the corrected output signal. The embodiments of the present disclosure have a compensation effect on TIA noise under different insertion loss channels, can improve the transmission performance of the system, and are suitable for various TIA devices.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to the field of communications, and in particular to a noise compensation method and device for a transimpedance amplifier. Background Art

[0002] As technologies such as AI and big models mature, cloud computing vendors have higher and higher requirements for bandwidth transmission in data centers. Optical modules are core devices for various interconnected scenarios in data centers. Currently, 800G optical modules have begun to be commercialized on a small scale. Transimpedance amplifiers (TIA) are one of the core components of optical modules. They are mainly used to convert current signals in optical modules into voltage signals and amplify them. However, during the amplification process, transimpedance amplifiers introduce noise, such as thermal noise, voltage noise, current noise, and so on. As the rate of optical modules increases, the noise introduced by transimpedance amplifiers will reduce the transmission performance of optical modules, so it is necessary to reduce or compensate for the noise of transimpedance amplifiers. Transimpedance amplifiers generally have an automatic gain adjustment function, that is, they automatically adjust the gain according to the set target voltage. However, the inventors found that the transimpedance amplifiers of mainstream manufacturers in high-speed optical modules will generate very large noise at certain gains, which will cause the performance of optical modules to seriously degrade at certain specific optical powers, so targeted compensation is needed. Summary of the invention

[0003] In order to compensate for the noise of a transimpedance amplifier, the present disclosure provides a method and device for compensating the noise of a transimpedance amplifier.

[0004] According to one aspect, an embodiment of the present disclosure provides a noise compensation method for a transimpedance amplifier, which includes a feedforward equalizer receiving an output signal of an analog front end and outputting an equalized output signal; a symbol number 2 based on a coding scheme of pulse amplitude modulation used; i and the equalized output signal to calculate the error signal, wherein i is a positive integer; the error signal is filtered using a filter to obtain an estimated noise signal; when compensation is performed before the feedforward equalizer, at the input end of the feedforward equalizer, the estimated noise signal is added to the output signal of the analog front end to serve as the corrected input signal of the feedforward equalizer; when compensation is performed after the feedforward equalizer, the estimated noise signal is added to the output signal of the feedforward equalizer to obtain a corrected output signal.

[0005] According to an embodiment of one aspect of the present disclosure, by compensating TIA noise, the transmission performance of the system can be effectively improved. At the same time, noise compensation can be performed before and / or after the feedforward equalizer according to the characteristics of the noise.

[0006] Optionally, according to the noise compensation method of the transimpedance amplifier of an embodiment of one aspect of the present disclosure, according to the symbol number 2 iand the equalized output signal to calculate the error signal, including:

[0007] If the number of symbols is equal to 2, calculate the decision value SZ using the following equation:

[0008]

[0009] where y is the output signal, k is the number of output signals y used to calculate the target value and the decision value, SZ is the average value of k output signals y; if i is greater than or equal to 2, perform i iterations of calculation. In the first iteration of calculation, calculate SZ according to the equation. In the jth iteration of calculation (where 1 < j ≤ i and j is a positive integer), use all the decision values obtained in the previous iterations to divide the numerical range of the output signal y into range intervals, calculate the average value of the output signals within each range interval to obtain new decision values. After i iterations of calculation, a total of 2 i - 1 decision values are obtained; divide all 2 i - 1 decision values into 2 i numerical ranges according to size, divide the k output signals y into 2 i groups according to the 2 i numerical ranges, calculate the average value of each group of output signals y, and use the 2 i average values as the 2 i target values; according to one of the 2 i numerical ranges where the output signal y is located, obtain the error signal, which is equal to the difference between the target value corresponding to the one numerical range and the output signal y.

[0010] In the embodiments of the present disclosure, in order to calculate the TIA noise signal, the error signal is first calculated. By using the above iterative method to calculate the error signal, the calculation result of the error signal can be made more accurate. This TIA noise signal compensation method can not only compensate for the TIA noise, but also compensate for the noise with a similar TIA noise distribution.

[0011] Optionally, according to the noise compensation method of the transimpedance amplifier according to an embodiment of one aspect of the present disclosure, when the coding scheme is PAM4, calculate the remaining 2 decision values in the second iteration of calculation according to the following equation:

[0012]

[0013] where mean represents the operation of taking the average. Among the k output signals y, calculate the average of the output signals greater than SZ and the average of the output signals less than SZ to obtain SH and SL respectively;

[0014] The four target values ​​are calculated according to the following equations:

[0015]

[0016] Wherein mean represents an average operation, TH3, TH1, TL1, and TL3 are four target values, the k output signals y are divided into four ranges by SH, SZ, and SL, and the output signals y in each range are averaged;

[0017] The error signal is calculated according to the following equation:

[0018]

[0019] Where error is the error signal.

[0020] Optionally, according to the noise compensation method of the transimpedance amplifier of an embodiment of one aspect of the present disclosure, when the coding scheme is PAM2, two target values ​​are calculated according to the following equations:

[0021]

[0022] Wherein mean represents an average operation, TH and TL are two target values, and among the k output signals y, the output signals y greater than SZ and the output signals y less than SZ are respectively averaged to obtain TH and TL;

[0023] The error signal is calculated according to the following equation:

[0024]

[0025] Where error is the error signal.

[0026] Optionally, according to the noise compensation method of the transimpedance amplifier of an embodiment of one aspect of the present disclosure, the filter is a first-order loop filter, and the estimated noise signal is obtained by the following equation:

[0027]

[0028] Wherein a is the tracking coefficient of the first-order loop filter, error is the error signal, z1 is the estimated noise signal, and n is a positive integer. The estimated noise signal is compensated after the feedforward equalizer.

[0029] Optionally, according to the noise compensation method of the transimpedance amplifier of an embodiment of one aspect of the present disclosure, the filter is a second-order filter, and the estimated noise signal is obtained by the following equation:

[0030]

[0031] where ki and kp are the tracking coefficients of the second-order loop filter, error is the error signal, d is the loop integral value, z2 is the estimated noise signal, m is a positive integer, and the estimated noise signal is compensated before the feed-forward equalizer.

[0032] According to an embodiment of another aspect of the present disclosure, there is provided a noise compensation device for a transimpedance amplifier, which includes a feed-forward equalizer that equalizes the output signal of the analog front-end to obtain an equalized output signal; an error estimation module, whose input terminal is connected to the output terminal of the feed-forward equalizer, and calculates an error signal based on the number of symbols 2 of the pulse amplitude modulation coding scheme used i and the equalized output signal, where i is a positive integer, and a filter is used to filter the error signal to obtain an estimated noise signal; a first transimpedance amplifier noise compensation module, whose input terminal is connected to the output terminal of the feed-forward equalizer and the output terminal of the error estimation module, and adds the output signal of the feed-forward equalizer and the estimated noise signal to obtain a corrected output signal.

[0033] Optionally, the noise compensation device for a transimpedance amplifier according to an embodiment of another aspect of the present disclosure further includes a second transimpedance amplifier noise compensation module, whose input terminal is connected to the output terminal of the error estimation module, and whose output terminal is connected to the input terminal of the feed-forward equalizer, and the error estimation module includes a second filter that filters the error signal to obtain a second estimated noise signal, and the second transimpedance amplifier noise compensation module adds the output signal of the analog front-end and the second estimated noise signal to obtain a corrected input signal of the feed-forward equalizer.

[0034] Optionally, for the noise compensation device for a transimpedance amplifier according to an embodiment of another aspect of the present disclosure, the error estimation module calculates the error signal based on the number of symbols 2 i and the equalized output signal, including

[0035] If the number of symbols is equal to 2, the decision value SZ is calculated using the following equation:

[0036]

[0037] where y is the output signal, k is the number of output signals y used to calculate the target value and the decision value, and SZ is the average value of k output signals y;

[0038] If i is greater than or equal to 2, then i iterations are performed. In the first iteration, SZ is calculated according to the equation. In the jth iteration (where 1 < j ≤ i, j is a positive integer), using all the decision values obtained in the previous iterations, the numerical range of the output signal y is divided into range intervals, and average the input signal in each range interval to obtain After i iterations, we get a total of 2 i -1 judgment value; set all 2 i -1 judgment value is divided into 2 by size i The k output signals y are divided into two groups according to 2 i The value range is divided into 2 i group, calculate the average value of each group of output signals y, and get 2 i The average value is 2 i target value; according to the 2 i A value range in the number of value ranges is used to obtain an error signal, which is equal to the difference between the target value corresponding to the number range and the output signal y.

[0039] Optionally, according to the noise compensation device of the transimpedance amplifier of the embodiment of another aspect of the present disclosure, the filter includes a first-order loop filter, which processes the error signal to obtain an estimated noise signal, and the estimated noise signal is output to the first transimpedance amplifier noise compensation module. The estimated noise signal is obtained by the following equation:

[0040]

[0041] Wherein a is the tracking coefficient of the first-order loop filter, error is the error signal, z1 is the estimated noise signal of the loop filter, and n is a positive integer.

[0042] Optionally, according to the noise compensation device of the transimpedance amplifier of the embodiment of another aspect of the present disclosure, the second filter includes a second-order loop filter, which processes the error signal to obtain a second estimated noise signal, and the second estimated noise signal is obtained by the following equation:

[0043]

[0044] Wherein ki and kp are the tracking coefficients of the second-order loop filter, error is the error signal, d is the loop integral value, z2 is the second estimated noise signal, and n is a positive integer.

[0045] Optionally, the noise compensation device for a transimpedance amplifier according to an embodiment of another aspect of the present disclosure further includes a decision feedback equalizer which receives the sum of the estimated noise signal of the transimpedance amplifier and the output of the feedforward equalizer.

[0046] According to an embodiment of another aspect of the present disclosure, there is provided a noise compensation device for a transimpedance amplifier, which includes a feedforward equalizer that equalizes the output signal of the analog front end to obtain an equalized output signal; an error estimation module, whose input end is connected to the output end of the feedforward equalizer, and calculates an error signal based on the number of symbols 2 of the pulse amplitude modulation coding scheme used, where i is a positive integer, and a filter is used to filter the error signal to obtain an estimated noise signal; a transimpedance amplifier noise compensation module, whose input end is connected to the output end of the error estimation module, and whose output end is connected to the input end of the feedforward equalizer, and adds the output signal of the analog front end and the estimated noise signal to obtain a corrected input signal of the feedforward equalizer. i And the equalized output signal to calculate an error signal, including

[0047] Optionally, for the noise compensation device of the transimpedance amplifier according to an embodiment of another aspect of the present disclosure, the error estimation module calculates an error signal based on the number of symbols 2 i And the equalized output signal, including

[0048] If the number of symbols is equal to 2, the decision value SZ is calculated using the following equation:

[0049]

[0050] where y is the output signal, k is the number of output signals y used to calculate the target value and the decision value, and SZ is the average value of k output signals y;

[0051] If i is greater than or equal to 2, then i iterations are performed. In the first iteration, SZ is calculated according to the equation. In the jth iteration (where 1 < j ≤ i, j is a positive integer), using all the decision values obtained in the previous iterations, the numerical range of the output signal y is divided into range intervals, and the average value of the output signals in each range interval is calculated to obtain new decision values. After i iterations, a total of 2 i -1 decision values are obtained; all 2 i -1 decision values are divided into 2 i numerical ranges according to size, and the k output signals y are divided into 2 i groups according to 2 i numerical ranges. The average value of each group of output signals y is calculated, and the 2 i average values obtained are used as 2 i target values; according to one of the 2 i numerical ranges where the output signal y is located, an error signal is obtained, which is equal to the difference between the target value corresponding to the one numerical range and the output signal y.

[0052] Optionally, the noise compensation device of the transimpedance amplifier according to another aspect of an embodiment of the present disclosure further includes a decision feedback equalizer, which receives the output signal of the feedforward equalizer, and the input signal of the feedforward equalizer is the sum of the output signal of the analog front end and the estimated noise signal of the transimpedance amplifier.

[0053] The embodiments of the present disclosure have a compensation effect on TIA noise under different insertion loss channels, can effectively improve the transmission performance of the system, and are suitable for various TIA devices. At the same time, the embodiments of the present disclosure can not only reduce the pre-FEC bit error rate, but also significantly reduce the number of FEC symbol errors, thereby improving the transmission performance of the system.

[0054] It is not necessary to achieve all the advantages described above at the same time for any device or method to implement the present disclosure. Other features and advantages of the present disclosure will be described in the subsequent description embodiments, and partly become apparent from the description embodiments, or be understood by implementing the present disclosure. The purposes and advantages of the embodiments of the present disclosure can be realized and obtained by the structures indicated in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings of the embodiments will be briefly introduced below. Obviously, the drawings in the following description only relate to some embodiments of the present disclosure, but are not intended to limit the present disclosure.

[0056] Figure 1 is a structural block diagram of the receiving side of an optical module according to an embodiment of the present disclosure;

[0057] Figure 2 is a structural block diagram of an optical module receiving side according to another embodiment of the present disclosure;

[0058] Figure 3 is a structural block diagram of an optical module receiving side according to yet another embodiment of the present disclosure;

[0059] Figure 4 is a flow chart of a noise compensation method for a transimpedance amplifier according to an embodiment of the present disclosure;

[0060] Figure 5 is a flow chart of a method for calculating an error signal according to an embodiment of the present disclosure;

[0061] Figure 6 This is a comparison curve diagram of gain effects under different channel insertion losses based on a conventional solution and an embodiment of the present disclosure;

[0062] Figure 7 This is a typical block diagram of the receiving side structure of an optical module. DETAILED DESCRIPTION

[0063] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution of the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings of the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. The various different embodiments can be combined with each other to form other embodiments not shown in the following description. Based on the described embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0064] Unless otherwise defined, the technical terms or scientific terms used herein shall have the usual meanings understood by persons of ordinary skill in the field to which the present disclosure belongs. The words "first", "second" and similar words used in the present disclosure specification and claims do not indicate any order, quantity or importance, but are only used to distinguish different components. Similarly, words such as "one" or "one" do not necessarily indicate a quantity restriction. Words such as "include" or "comprise" mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Words such as "connect" or "connected" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0065] Figure 7 This is a block diagram of a typical optical module receiving side structure. After the optical signal is transmitted through the optical fiber, it is first converted into a current signal by the optical module's detector (PD), and then the current signal is converted into a voltage signal and amplified by the transimpedance amplifier (TIA). The amplified voltage signal enters the DSP chip and is processed by the analog front end (AFE), feedforward equalizer (FFE), decision feedback equalizer (DFE) and other modules to restore the original transmission signal. Figure 7 Only the above-mentioned partial modules in the DSP chip for restoring the original transmitted signal are shown, and other structures of the DSP chip are not shown.

[0066] As the speed of optical modules increases, the bandwidth required by TIA increases, and the noise generated becomes more and more obvious. The processing modules in the current DSP, such as FFE and DFE, can only process the insertion loss and crosstalk of the transmission channel, and cannot effectively compensate for the rapidly changing noise generated by TIA, which will lead to the decline of optical module receiving performance.

[0067] Figure 1It is a structural block diagram of the receiving side of an optical module according to an embodiment of the present disclosure. After the optical signal is transmitted through the optical fiber, it is converted into an amplified voltage signal through the first PD 110 and the first TIA 120 of the optical module and input into the DSP unit 130. The DSP unit 130 includes a first AFE 131, a first FFE 132, a first TIA noise compensation module 133, a first DFE 134 and a first noise estimation module 135, wherein the first AFE 131, the first FFE 132, the first TIA noise compensation module 133, and the first DFE 134 are connected in sequence, and the first noise estimation module 135 is connected between the first FFE 132 and the first TIA noise compensation module 133.

[0068] The first AFE 131 is used to receive the voltage signal sent by the first TIA 120 and convert the analog signal into a digital signal. The first FFE 132 is used to perform an equalization algorithm on the received signal to reduce inter-symbol interference. The first noise estimation module 135 receives the output of the first FFE 132 and outputs its estimated noise signal to the first TIA noise compensation module 133. The first TIA noise compensation module 133 inputs the sum of the output of the first FFE 132 and the noise signal estimated by the first noise estimation module 135 to the first DFE 134. The first TIA noise compensation module 133 may include an adder or a circuit that implements the adder function. The compensation process of TIA noise is described in detail below.

[0069] The signal input to the DSP unit 130 is pre-processed by the first AFE 131, and the output signal x of the first AFE is filtered by the first FFE 132 to output the filtered signal y, where

[0070] (Equation 1)

[0071] where h is the coefficient of the first FFE 132, Optionally, the first FFE 132 may adopt an iterative algorithm to update the coefficients to filter the input signal.

[0072] To obtain an estimated noise signal, the first noise estimation module 135 receives the signal y and first calculates an error signal.

[0073] The number of symbols used in the Pulse Amplitude Modulation (PAM) coding scheme is 2 i (i is an integer) to determine the number of target values ​​and decision values. The number of target values ​​is equal to the number of symbols 2 i , the number of decision values ​​is 1 less than the number of target values, that is, 2 i -1 pc.

[0074] First, calculate a decision value SZ based on the signal y output by the first FFE 132. As shown in Equation (2), accumulate a certain number of signals y and calculate the average value to obtain SZ, that is:

[0075] (Equation 2)

[0076] Among them, k is the number of signals y used to calculate the target value and the decision value, and the value of k can be determined in advance.

[0077] If the number of symbols 2 i equals 2, that is, i = 1, the decision value is SZ. If i is greater than or equal to 2, in order to calculate the decision value, i iterations of calculation will be performed. First, calculate SZ in the first iteration. At the jth iteration (1 < j ≤ i, j is an integer), using all the decision values obtained in the previous iterations, divide the numerical range of the signal y into range intervals, calculate the average value of the input signals in each range interval to obtain new decision values. After i iterations of calculation, a total of 2 i -1 decision values are obtained.

[0078] Based on all 2 i -1 decision values (including SZ), calculate 2 i target values. Among them, divide the k signals y into 2 i -1 numerical ranges according to the magnitudes of the decision values, and then divide these 2 i numerical ranges into 2 i groups, and then obtain the average values of the 2 i groups of signals y. The obtained 2 i average values are used as the 2 i target values.

[0079] Then, according to one of the 2 i numerical ranges where the current signal y is located, obtain an error signal, which is equal to the difference between the target value corresponding to this one numerical range and the signal y.

[0080] In this embodiment, by obtaining the numerical range corresponding to the output signal y through an iterative method to calculate the error signal, a more accurate error signal can be obtained, so that the subsequent estimation of the TIA noise is more accurate.

[0081] In one embodiment, when the PAM4 coding scheme is used, the number of symbols is 4, i=2, and 2 iterations are required. First, the error estimation module 135 uses the signal y to obtain the decision value SZ according to equation (2), and then uses the signal y and the decision value SZ to calculate the four target values ​​TH3, TH1, TL1 and TL3 of the error signal and the remaining two decision values ​​SH and SL. Specifically, the signal y greater than SZ and the signal less than SZ are found from the k signals y, and then the signals are averaged to obtain SH and SL respectively. In the following formula, mean represents the averaging operation.

[0082] (Equation 3)

[0083] Next, the target value is calculated based on the three decision values, where SH>SZ>SL. The three decision values ​​can be divided into four value ranges, and the k signals y are divided into four groups, where the k signals y are averaged to obtain TH3; the k signals y are averaged to obtain TH1; the k signals y are averaged to obtain TL1; the k signals y are averaged to obtain TL3 (the above range division method can also use greater than or equal to replace greater than or less than or equal to replace less than, which has little effect on the final calculation result, so no distinction is made. This also applies to the demarcation of various ranges below). As shown in equation (4), mean represents the operation of averaging.

[0084] (Equation 4)

[0085] Then the error signal error is calculated according to the following formula:

[0086] (Equation 5)

[0087] In another embodiment, if PAM2 modulation coding is used, the number of symbols is 2, i=1, and only one calculation is required to obtain one decision value SZ. SZ is used to calculate two target values ​​TH and TL. The calculation method of SZ is the same as that of PAM4 modulation coding, see equation (2). The calculation methods of TH and TL are as follows:

[0088] (Equation 6)

[0089] Mean in equation (6) represents the operation of averaging. Next, the error signal error is calculated according to equation (7):

[0090] (Equation 7)

[0091] The error signal will be used to estimate the TIA noise signal. The first noise estimation module 135 can filter the error signal error using a loop filter and generate an estimated TIA noise signal. The first noise estimation module 135 can include a first-order loop filter to perform a sliding average process on the error signal, and the estimated noise signal is calculated by the following equation:

[0092] (Equation 8)

[0093] Wherein a is a preset tracking coefficient of the first-order loop filter, which can be configured according to the speed of noise change, such as 0.1, 0.01, etc. Generally speaking, the faster the noise changes, the larger the value of a; z1 is the noise signal estimated by the loop filter, and n is a positive integer. The first TIA noise compensation module 133 uses the calculated estimated noise signal to add to the output signal y of the first FFE 132 to compensate for the adverse effects of TIA noise on output performance.

[0094] Figure 2 2 is a block diagram of the optical module receiving side structure according to another embodiment of the present disclosure. After the optical signal is transmitted through the optical fiber, it is converted into an amplified voltage signal by the second detector (PD) 210 and the second transimpedance amplifier (TIA) 220 of the optical module and input to the DSP unit 230. The second PD 210 and the second TIA 220 are connected to the Figure 1 The first PD 110 shown is the same as or similar to the first TIA 120. The DSP unit 230 includes a second AFE 231, a second FFE 232, a second TIA noise compensation module 233, a second DFE 234, and a second noise estimation module 235, wherein the second AFE 231, the second TIA noise compensation module 233, the second FFE 232, and the second DFE 234 are connected in sequence, and the second noise estimation module 235 is connected between the second TIA noise compensation module 233 and the second FFE 232, receives the output of the second FFE 232, and outputs the TIA noise signal calculated by it to the second TIA noise compensation module 233. The second TIA noise compensation module 233 inputs the sum of the estimated noise signal and the output signal of the second AFE 231 to the second FFE 232.

[0095] and Figure 1Similar to the embodiment shown, the signal input to the DSP unit 230 is pre-processed by the second AFE 231. The output signal x of the second AFE is output to the second FFE 232 for filtering, and the filtered signal y is output. The initial relationship between x and y is defined by equation (1). The second noise estimation module 235 calculates the error signal in the same way as the first noise estimation module 135, that is, according to the received signal and the number of symbols used in the pulse amplitude modulation coding scheme 2 i To determine the decision value and the target value, and then calculate the error signal according to the target value, refer to the description of the error signal calculation process in the above text, and the above specific examples of error signal calculation in the PAM4 and PAM2 coding schemes are also applicable to the second noise estimation module 235. For example, when the coding scheme is PAM4, the error signal calculated by the second noise estimation module 235 is obtained by equation (5), and when the coding scheme is PAM2, the error signal calculated by the second noise estimation module 235 is obtained by equation (7), which will not be repeated here.

[0096] The second noise estimation module 235 may filter the error signal using a loop filter and generate an estimated noise signal. Figure 2 In the illustrated embodiment, the second noise estimation module 235 may include a second-order loop filter or a high-order loop filter and compensate for TIA noise before the second FFE. When a second-order loop filter is used, the noise signal may be calculated by the following equation:

[0097] (Equation 9)

[0098] Where ki and kp are the tracking coefficients of the second-order loop filter, which can be configured according to the speed of noise change, such as ki=1 / 16, kp=1 / 256, etc., d is the loop integral value, z2 is the noise signal estimated by the loop filter, and n is a positive integer. The second TIA noise compensation module uses the calculated estimated noise signal to add to the output signal x of the second AFE to compensate for the adverse effects of TIA noise.

[0099] Figure 1 and Figure 2 The first and second TIA noise compensation modules and the noise estimation module are implemented by hardware circuits to ensure the calculation speed. The DSP unit represents a device for implementing the DSP function, which can be implemented by a DSP chip, an ASIC chip, or an FPGA chip, etc. Correspondingly, the TIA noise compensation module and the noise estimation module can be implemented on hardware such as a DSP chip, an ASIC chip, or an FPGA chip.

[0100] Figure 1 The illustrated embodiment compensates the output signal of the FFE. Figure 2The embodiment shown is to compensate the input signal of FFE. Both implementations can compensate TIA noise. Figure 1 Compared to the embodiment shown, Figure 2 The illustrated embodiment is more suitable for situations where the noise amplitude is relatively large and changes relatively slowly. Figure 1 and Figure 2 The first and second TIA noise compensation modules and the noise estimation module shown can be set in the DSP unit separately or simultaneously. When the first and second TIA noise compensation modules and the noise estimation module are set in the DSP unit at the same time, a gating module corresponding to the two schemes can be set to adopt one of the above compensation schemes or adopt two compensation schemes at the same time as needed to compensate the TIA noise signal. Figure 3 An embodiment is shown that employs two compensation schemes simultaneously.

[0101] Figure 3 is a structural block diagram of the receiving side of an optical module according to another embodiment of the present disclosure, wherein Figure 1 The second TIA noise compensation module 233 is added to the embodiment of the present invention, and the first noise estimation module 135 is modified into a third noise estimation module 335, which is connected to the first and second TIA noise compensation modules. The third noise estimation module 335 includes the functions of the first noise estimation module 135 and the second noise estimation module 235. The structure can also use both the first noise estimation module 135 and the second noise estimation module 235 to replace the third noise estimation module 335. Figure 1 and Figure 2 For the same parts, see the above description. Figure 1 and Figure 2 In the embodiment shown, the third noise estimation module 335 can use different filters to estimate the noise signal for the output signal and the input signal of the FFE, that is, it can include the first and second filters. The third noise estimation module 335 calculates the error signal, and filters the error signal through the first and second filters to obtain the first estimated noise signal and the second estimated noise signal. They are compensated by the first TIA noise compensation module 133 and the second TIA noise compensation module 233 after and before the FFE module, respectively. The process of calculating the error signal and the process of estimating the noise signal are similar to Figure 1 and Figure 2 The methods in the embodiments are the same.

[0102] Figure 4 The flowchart of the noise compensation method of the transimpedance amplifier according to an embodiment is shown. In step 41, the input signal is equalized using FFE in the DSP unit to obtain the output signal y of the FFE. The relationship between the input and output signals can be seen in equation 1.

[0103] In step 42, an error signal error is calculated based on the number of symbols used in the pulse amplitude modulation (PAM) coding scheme and the output signal y of the FFE.

[0104] Specifically, according to the number of symbols 2 used in the PAM coding scheme i (where i is an integer), the number of target values and decision values is determined. The number of target values is equal to the number of symbols 2 i , and the number of decision values is one less than the number of target values, i.e., 2 i - 1.

[0105] Figure 5 FIG. shows a flowchart of a method for calculating an error signal according to an embodiment of the present disclosure. Refer to Figure 5 , in step 421, a decision value SZ is calculated based on the signal y output by the FFE. As shown in Equation (2), a certain number of signal y values are accumulated and averaged to obtain SZ.

[0106] In step 422, if the number of symbols in the coding scheme is equal to 2, the decision value is SZ; if the number of symbols 2 i is greater than 2, i iterations are performed. In the first iteration, SZ is calculated. In the jth iteration (1 < j ≤ i, j is a positive integer), using all the decision values obtained in the previous iterations, the numerical range of the signal y is divided into range intervals, and the average value of the input signals in each range interval is calculated to obtain new decision values. After i iterations, a total of 2 i - 1 decision values are obtained.

[0107] In step 423, based on all 2 i - 1 decision values (including SZ), 2 i target values are calculated. Among them, k signal y values are divided into 2 i - 1 groups according to the 2 i numerical ranges divided by the decision values in ascending order, and the average value of each group of signal y values is obtained. The 2 i obtained average values are used as the 2 i target values. i

[0108] Next, in step 424, according to the numerical range in which the signal y is located among the 2 i numerical ranges, the error signal error is obtained, which is equal to the difference between the target value corresponding to the numerical range and the signal y.

[0109] More specific steps for obtaining the error signal using the PAM4 and PAM2 coding schemes can be found in the description of Equations 2 - 7 above.

[0110] Back to Figure 4 In step 43, based on the error signal error, a filter is used to filter the error signal to obtain an estimated noise signal. Optionally, compensation can be performed before and / or after the FFE module. Optionally, when compensation is performed after the FFE module, a first-order loop filter can be used to filter the error signal, and the estimated noise signal after filtering can be expressed by equation (8). The tracking coefficient of the loop filter is configured according to the speed of noise change.

[0111] Alternatively, when compensation is performed before the FFE module, a second-order loop filter or a high-order loop filter can be used to filter the error signal. When a second-order loop filter is used, the estimated noise signal after filtering can be expressed by equation (9). The tracking coefficient of the second-order or high-order loop filter can be configured according to the speed of noise change.

[0112] In step 44, the input and / or output signal of the FFE is corrected by adding the estimated noise signal to the input and / or output signal of the FFE module to compensate for the TIA noise. Specifically, when compensation is selected before the FFE, the estimated noise signal is added to the input signal of the FFE; when compensation is selected after the FFE, the estimated noise signal is added to the output signal of the FFE, and the compensated output signal generated thereby is input to the DFE. Optionally, compensation can also be selected both before and after the FFE. The above method can be used to compensate for TIA noise and noise with similar distribution characteristics to TIA noise.

[0113] The following is based on the above Figure 1 Simulation and experimental results of the scheme shown. Figure 6 The results of simulating a 106Gbps PAM4 signal are shown, which show the gain effect of the conventional solution and the solution of the present invention under different channel insertion losses. During the simulation process, different insertion losses were added to the channel, and then TIA noise was added. From the comparison of the graphs, it can be seen that the solution of the present invention has a compensation effect for TIA noise under different insertion loss channels, which can effectively improve the transmission performance of the system.

[0114] In addition, Table 1 shows the statistical analysis results of the 106 Gbps optical module experimental data, in which the pre-correction bit error rate (pre-FEC BER) and the number of KP4 FEC (KP4 forward error correction technology) symbol errors of the conventional method and the solution of the present invention are statistically analyzed.

[0115] Pre-FEC BER KP4 FEC symbol error rate Conventional solution 2.30E-05 11 Solution of the present invention 3.80E-06 3

[0116] Table 1

[0117] From the above comparison results, we can see that this scheme can not only reduce the bit error rate before FEC correction, but also significantly reduce the number of FEC symbol errors, thereby improving the transmission performance of the system.

[0118] The above description is merely an exemplary embodiment of the present disclosure and is not intended to limit the protection scope of the present disclosure. The protection scope of the present disclosure is determined by the appended claims.

Claims

1. A noise compensation method for a transimpedance amplifier, comprising: The feedforward equalizer receives the output signal of the analog front end and outputs an equalized output signal; Number of symbols based on the pulse amplitude modulation coding scheme used2 i and the equalized output signal to calculate the error signal, where i is a positive integer, which includes If the number of symbols is equal to 2, the average value of k output signals y is used as the decision value, where k is the number of output signals y used to calculate the target value and the decision value; if i is greater than or equal to 2, i iterations are performed to obtain a total of 2 i -1 judgment value; set all 2 i -1 judgment value is divided into 2 by size i The k output signals y are divided into two groups according to 2 i The value range is divided into 2 i group, calculate the average value of each group of output signals y, and get 2 i The average value is 2 i target value; according to the 2 i A value range is a value range in the value range, and an error signal is obtained; Using a filter to filter the error signal to obtain an estimated noise signal; When compensation is performed before the feedforward equalizer, the estimated noise signal is added to the output signal of the analog front end at the input end of the feedforward equalizer to serve as the corrected input signal of the feedforward equalizer; when compensation is performed after the feedforward equalizer, the estimated noise signal is added to the output signal of the feedforward equalizer to obtain a corrected output signal.

2. The noise compensation method for a transimpedance amplifier according to claim 1, characterized in that If i is greater than or equal to 2, then perform i iterations to obtain a total of 2 i -1 judgment value includes In the first iteration, the decision value SZ is calculated according to the following equation: , SZ is the average value of k output signals y. During the j-th iteration calculation, using all the decision values obtained from the previous j - 1 iterations, the numerical range of the output signal y is divided into 2 j-1 range intervals. The average value of the output signal y within each range interval is calculated to obtain 2 j-1 new decision values. After i iterations of calculation, a total of 2 i - 1 decision values are obtained, where 1 < j ≤ i and j is a positive integer; and The error signal is equal to the difference between the target value corresponding to the one numerical range and the output signal y.

3. The noise compensation method for a transimpedance amplifier according to claim 2, characterized in that: When the coding scheme is PAM4, the remaining two decision values ​​are calculated in the second iteration calculation according to the following equation: , Wherein mean represents an average operation, and among the k output signals y, the output signals greater than SZ and the output signals less than SZ are averaged to obtain SH and SL respectively; The four target values ​​are calculated according to the following equations: , Wherein mean represents an average operation, TH3, TH1, TL1, and TL3 are four target values, the k output signals y are divided into four ranges by SH, SZ, and SL, and the output signals y in each range are averaged; The error signal is calculated according to the following equation: , Where error is the error signal.

4. The noise compensation method for a transimpedance amplifier according to claim 1, characterized in that: When the coding scheme is PAM2, two target values ​​are calculated according to the following equations: , Wherein SZ is the decision value, mean represents the operation of averaging, TH and TL are two target values, and among the k output signals y, the output signals y greater than SZ and the output signals y less than SZ are averaged to obtain TH and TL respectively; The error signal is calculated according to the following equation: , Where error is the error signal.

5. The noise compensation method for a transimpedance amplifier according to claim 1, characterized in that: The filter is a first-order loop filter, and the estimated noise signal is obtained by the following equation: , Wherein a is the tracking coefficient of the first-order loop filter, error is the error signal, z1 is the estimated noise signal, and n is a positive integer. The estimated noise signal is compensated after the feedforward equalizer.

6. The noise compensation method for a transimpedance amplifier according to claim 1, characterized in that: The filter is a second-order loop filter, and the estimated noise signal is obtained by the following equation: , Wherein ki and kp are the tracking coefficients of the second-order loop filter, error is the error signal, d is the loop integral value, z2 is the estimated noise signal, n is a positive integer, and the estimated noise signal is compensated before the feedforward equalizer.

7. A noise compensation device for a transimpedance amplifier, comprising: A feedforward equalizer that performs equalization processing on the output signal of the analog front end to obtain an equalized output signal; The error estimation module, whose input is connected to the output of the feedforward equalizer, is based on the number of symbols of the pulse amplitude modulation coding scheme used. i and the equalized output signal to calculate an error signal, wherein i is a positive integer, and the error signal is filtered using a filter to obtain an estimated noise signal; a first transimpedance amplifier noise compensation module, whose input end is connected to the output end of the feedforward equalizer and the output end of the error estimation module, and adds the output signal of the feedforward equalizer and the estimated noise signal to obtain a corrected output signal; The number of symbols in the coding scheme based on the pulse amplitude modulation used is 2 i and the equalized output signal to calculate the error signal including if the number of symbols is equal to 2, taking the average value of k output signals y as the decision value, wherein k is the number of output signals y used to calculate the target value and the decision value; If i is greater than or equal to 2, then perform i iterations to obtain a total of 2 i -1 decision value; Set all 2 i -1 judgment value is divided into 2 by size i The k output signals y are divided into two groups according to 2 i The value range is divided into 2 i group, calculate the average value of each group of output signals y, and get 2 i The average value is 2 i target value; According to the output signal y where the 2 i An error signal is obtained by selecting a value range from a value range.

8. The noise compensation device for a transimpedance amplifier according to claim 7, characterized in that It also includes a second transimpedance amplifier noise compensation module, whose input end is connected to the output end of the error estimation module, and whose output end is connected to the input end of the feedforward equalizer, and the error estimation module includes a second filter, which filters the error signal to obtain a second estimated noise signal, and the second transimpedance amplifier noise compensation module adds the output signal of the analog front end and the second estimated noise signal to obtain a corrected input signal of the feedforward equalizer.

9. The noise compensation device for a transimpedance amplifier according to claim 7 or 8, characterized in that If i is greater than or equal to 2, then perform i iterations to obtain a total of 2 i -1 judgment value includes In the first iteration, the decision value SZ is calculated according to the following equation: , SZ is the average value of k output signals y. During the j-th iteration calculation, using all the decision values obtained from the previous j - 1 iterations, the numerical range of the output signal y is divided into 2 j-1 range intervals. The average value of the output signal y within each range interval is calculated to obtain 2 j-1 new decision values. After i iterations of calculation, a total of 2 i - 1 decision values are obtained, where 1 < j ≤ i and j is a positive integer; and The error signal is equal to the difference between the target value corresponding to the one numerical range and the output signal y.

10. The noise compensation device for a transimpedance amplifier according to claim 9, characterized in that: When the coding scheme is PAM4, the remaining two decision values ​​are calculated in the second iteration calculation according to the following equation: , Wherein mean represents an average operation, and among the k output signals y, the output signals greater than SZ and the output signals less than SZ are averaged to obtain SH and SL respectively; The four target values ​​are calculated according to the following equations: , Wherein mean represents an average operation, TH3, TH1, TL1, and TL3 are four target values, the k output signals y are divided into four ranges by SH, SZ, and SL, and the output signals y in each range are averaged; The error signal is calculated as follows: , Where error is the error signal.

11. The noise compensation device for a transimpedance amplifier according to claim 7 or 8, characterized in that: When the coding scheme is PAM2, two target values ​​are calculated according to the following equations: , Wherein SZ is the decision value, mean represents the operation of averaging, TH and TL are two target values, and among the k output signals y, the output signals y greater than SZ and the output signals y less than SZ are averaged to obtain TH and TL respectively; The error signal is calculated as follows: , Where error is the error signal.

12. The noise compensation device for a transimpedance amplifier according to claim 7, characterized in that The filter includes a first-order loop filter, which processes the error signal to obtain an estimated noise signal, and the estimated noise signal is output to the first transimpedance amplifier noise compensation module. The estimated noise signal is obtained by the following equation: , Wherein a is the tracking coefficient of the first-order loop filter, error is the error signal, z1 is the estimated noise signal of the loop filter, and n is a positive integer.

13. The noise compensation device for a transimpedance amplifier according to claim 8, characterized in that The second filter includes a second-order loop filter, which processes the error signal to obtain a second estimated noise signal, and the second estimated noise signal is obtained by the following equation: , Wherein ki and kp are the tracking coefficients of the second-order loop filter, error is the error signal, d is the loop integral value, z2 is the second estimated noise signal, and n is a positive integer.

14. A noise compensation device for a transimpedance amplifier, comprising: A feedforward equalizer that performs equalization processing on the output signal of the analog front end to obtain an equalized output signal; The error estimation module, whose input is connected to the output of the feedforward equalizer, is based on the number of symbols of the pulse amplitude modulation coding scheme used. i and the equalized output signal to calculate an error signal, wherein i is a positive integer, and the error signal is filtered using a filter to obtain an estimated noise signal; a transimpedance amplifier noise compensation module, whose input end is connected to the output end of the error estimation module, and whose output end is connected to the input end of the feedforward equalizer, and adds the output signal of the analog front end and the estimated noise signal to obtain a corrected input signal of the feedforward equalizer, The number of symbols in the coding scheme based on the pulse amplitude modulation used is 2 i And the equalized output signal to calculate the error signal includes If the number of symbols is equal to 2, the average value of k output signals y is used as the decision value, where k is the number of output signals y used to calculate the target value and the decision value; If i is greater than or equal to 2, then perform i iterations to obtain a total of 2 i -1 decision value; Set all 2 i -1 judgment value is divided into 2 by size i The k output signals y are divided into two groups according to 2 i The value range is divided into 2 i group, calculate the average value of each group of output signals y, and get 2 i The average value is 2 i target value; According to the output signal y where the 2 i An error signal is obtained by selecting a value range from a value range.

15. The noise compensation device for a transimpedance amplifier according to claim 14, characterized in that If i is greater than or equal to 2, then perform i iterations to obtain a total of 2 i -1 decision value is included in the first iteration calculation according to the following equation to calculate the decision value SZ, , SZ is the average value of k output signals y. During the j-th iteration calculation, using all the decision values obtained from the previous j - 1 iterations, the numerical range of the output signal y is divided into 2 j-1 range intervals. The average value of the output signal y within each range interval is calculated to obtain 2 j-1 new decision values. After i iterations of calculation, a total of 2 i - 1 decision values are obtained, where 1 < j ≤ i and j is a positive integer; and The error signal is equal to the difference between the target value corresponding to the one numerical range and the output signal y.

16. The noise compensation device for a transimpedance amplifier according to claim 15, characterized in that: When the coding scheme is PAM4, the remaining two decision values ​​are calculated in the second iteration calculation according to the following equation: , Wherein mean represents an average operation, and among the k output signals y, the output signals greater than SZ and the output signals less than SZ are averaged to obtain SH and SL respectively; The four target values ​​are calculated according to the following equations: , Wherein mean represents an average operation, TH3, TH1, TL1, and TL3 are four target values, the k output signals y are divided into four ranges by SH, SZ, and SL, and the output signals y in each range are averaged; The error signal is calculated as follows: , Where error is the error signal.

17. The noise compensation device for a transimpedance amplifier according to claim 14, characterized in that: When the coding scheme is PAM2, two target values ​​are calculated according to the following equations: , Wherein SZ is the decision value, mean represents the operation of averaging, TH and TL are two target values, and among the k output signals y, the output signals y greater than SZ and the output signals y less than SZ are averaged to obtain TH and TL respectively; The error signal is calculated as follows: , Where error is the error signal.

18. The noise compensation device for a transimpedance amplifier according to claim 14, characterized in that The filter includes a second-order loop filter, which processes the error signal to obtain an estimated noise signal, and the estimated noise signal is obtained by the following equation: , Where ki and kp are the tracking coefficients of the second-order loop filter, error is the error signal, d is the loop integral value, z2 is the estimated noise signal, and n is a positive integer.