Nonlinear compensation drive circuit

By designing a nonlinear compensation drive circuit and utilizing a combination of transconductance amplification, linear amplification, and nonlinear amplification units, the static nonlinearity problem of silicon photonic modulators was solved, achieving linear drive and nonlinear compensation in the current domain, thereby improving the linearity and reliability of signal transmission.

CN123122467APending Publication Date: 2026-11-17INST OF SEMICONDUCTORS - CHINESE ACAD OF SCI
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611622130.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-20
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Silicon photonic modulators have an inherent static nonlinearity problem, which causes the mapping characteristics between the output optical signal and the input electrical signal to deviate from a linear relationship. Existing driving circuits are unable to effectively compensate for static nonlinear distortion, resulting in a decrease in signal transmission quality.

Method used

Design a nonlinear compensation drive circuit, including a transconductance amplification unit, a linear amplification unit, and a nonlinear amplification unit. By converting the input voltage signal into a linear current signal and performing linear and nonlinear amplification in the current domain, a nonlinear compensation current signal is generated, which is finally converted into a nonlinear voltage signal to cancel the nonlinear distortion of the silicon photonic modulator.

Benefits of technology

It suppresses harmonic distortion over a wide dynamic range, improves the linearity and reliability of signal transmission in silicon photonic modulators, and enhances the modulation performance of multi-level signals.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123122467A_ABST
    Figure CN123122467A_ABST
Patent Text Reader

Abstract

This application provides a nonlinear compensation driving circuit, relating to the field of driving circuit technology, comprising: a transconductance amplification unit for amplifying an input voltage signal and converting it into a linear current signal; a linear amplification unit for linearly amplifying the linear current signal to obtain an amplified linear current signal; a nonlinear amplification unit for nonlinearly amplifying the linear current signal to generate a nonlinear compensation current signal whose amplitude is nonlinearly related to the input voltage signal; and a linear termination circuit, connected to the output terminals of the linear amplification unit and the nonlinear amplification unit respectively, for superimposing the amplified linear current signal and the nonlinear compensation current signal to obtain a superimposed current signal, converting the superimposed current signal into a nonlinear voltage signal, and outputting it to a silicon photonics modulator. This application solves the technical problem that silicon photonics modulator driving circuits are difficult to effectively compensate for static nonlinear distortion.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of drive circuit technology, and more specifically, to a nonlinear compensation drive circuit. Background Technology

[0002] In silicon photonic interconnect systems, silicon photonic modulators, as key devices for converting electrical signals to optical signals, are widely used in optical transceiver links. These include various implementations such as micro-ring modulators and Mach-Zehnder modulators (MZM).

[0003] However, silicon photonic modulators generally suffer from inherent static nonlinearity. The relationship between their voltage and effective refractive index typically exhibits a Lorentz-like curve or a sinusoidal curve, causing the mapping characteristics between the output optical signal and the input electrical signal to deviate from a linear relationship. Summary of the Invention

[0004] In view of this, this application provides a nonlinear compensation driving circuit to solve the problem that related silicon photonic modulator driving circuits are difficult to effectively compensate for static nonlinear distortion, which leads to a decrease in signal transmission quality.

[0005] One aspect of this application provides a nonlinear compensation driving circuit, comprising: a transconductance amplification unit for amplifying an input voltage signal and converting it into a linear current signal; a linear amplification unit connected to the output of the transconductance amplification unit for linearly amplifying the linear current signal to obtain an amplified linear current signal; a nonlinear amplification unit connected to the output of the transconductance amplification unit for nonlinearly amplifying the linear current signal to generate a nonlinear compensation current signal that is nonlinearly related to the amplitude of the input voltage signal; and a linear termination circuit connected to the outputs of the linear amplification unit and the nonlinear amplification unit, respectively, for superimposing the amplified linear current signal and the nonlinear compensation current signal to obtain a superimposed current signal, converting the superimposed current signal into a nonlinear voltage signal and outputting it to a silicon photonic modulator to compensate for the nonlinear distortion of the silicon photonic modulator.

[0006] According to an embodiment of this application, the transconductance amplification unit includes: a first PMOS transistor and a second PMOS transistor, wherein the gate of the first PMOS transistor receives a first differential input positive signal, and the gate of the second PMOS transistor receives a first differential input negative signal; the sources of the first PMOS transistor and the second PMOS transistor are connected together and are connected to a first current source; a first NMOS transistor and a second NMOS transistor, wherein the gate of the first NMOS transistor receives a second differential input positive signal, and the gate of the second NMOS transistor receives a second differential input negative signal; the sources of the first NMOS transistor and the second NMOS transistor are connected together and are connected to a second current source; wherein the first differential input positive signal and the first differential input negative signal constitute a first differential input signal, the second differential input positive signal and the second differential input negative signal constitute a second differential input signal, and the voltage amplitude of the first differential input signal is greater than the voltage amplitude of the second differential input signal.

[0007] According to an embodiment of this application, the transconductance amplification unit further includes: a bias circuit for providing bias voltage and bias current to the first PMOS transistor, the second PMOS transistor, the first NMOS transistor, and the second NMOS transistor, such that the first PMOS transistor, the second PMOS transistor, the first NMOS transistor, and the second NMOS transistor all operate in the saturation region, so as to amplify the input voltage signal and convert it into a linear current signal through transconductance gain.

[0008] According to an embodiment of this application, the output terminal of the transconductance amplification unit includes a first PMOS output node, a second PMOS output node, a first NMOS output node, and a second NMOS output node; the first PMOS output node and the second PMOS output node are respectively formed by the drains of the first PMOS transistor and the second PMOS transistor, and are respectively connected to the corresponding input terminals of the linear amplification unit and the nonlinear amplification unit; the first NMOS output node and the second NMOS output node are respectively formed by the drains of the first NMOS transistor and the second NMOS transistor, and are respectively connected to the corresponding input terminals of the linear amplification unit and the nonlinear amplification unit; wherein, the first PMOS output node, the second PMOS output node, the first NMOS output node, and the second NMOS output node are electrically isolated from each other, and the input impedance presented by the linear amplification unit is less than the input impedance presented by the nonlinear amplification unit at the corresponding output node.

[0009] According to an embodiment of this application, the linear amplification unit includes: a third NMOS transistor and a fourth NMOS transistor, the gates of both the third and fourth NMOS transistors being connected to an NMOS common bias voltage terminal, the sources of the third and fourth NMOS transistors being connected to a first NMOS output node and a second NMOS output node respectively, and the drains of the third and fourth NMOS transistors being connected to a first differential output node and a second differential output node respectively; and a third PMOS transistor and a fourth PMOS transistor, the gates of both the third and fourth PMOS transistors being connected to a PMOS common bias voltage terminal. The sources of the third PMOS transistor and the fourth PMOS transistor are connected to the first PMOS output node and the second PMOS output node, respectively. The drains of the third PMOS transistor and the fourth PMOS transistor are connected to the first differential output node and the second differential output node, respectively. The drains of the third PMOS transistor and the third NMOS transistor are connected to the first differential output node, and the drains of the fourth PMOS transistor and the fourth NMOS transistor are connected to the second differential output node. The third NMOS transistor, the fourth NMOS transistor, the third PMOS transistor, and the fourth PMOS transistor are all configured to operate in the saturation region within the signal swing range to transmit linear current signals.

[0010] According to embodiments of this application, the sources of the third NMOS transistor, the fourth NMOS transistor, the third PMOS transistor, and the fourth PMOS transistor respectively constitute a first NMOS low-impedance input node, a second NMOS low-impedance input node, a first PMOS low-impedance input node, and a second PMOS low-impedance input node; the drains of the third PMOS transistor and the third NMOS transistor together constitute a first differential output node, and the drains of the fourth PMOS transistor and the fourth NMOS transistor together constitute a second differential output node; the first differential output node and the second differential output node are connected to a linear termination circuit; wherein, the input impedance of each low-impedance input node is less than the output impedance of the first differential output node and the second differential output node.

[0011] According to embodiments of this application, the nonlinear amplification unit includes multiple NMOS nonlinear units and multiple PMOS nonlinear units; each NMOS nonlinear unit and each PMOS nonlinear unit includes an input transistor and a bias transistor connected in series; the gate of the input transistor of the PMOS nonlinear unit is connected to the corresponding PMOS output node, and the gate of the input transistor of the NMOS nonlinear unit is connected to the corresponding NMOS output node; the source of each input transistor receives an independently adjustable control voltage, and the drain is connected to the source of the corresponding bias transistor; the gate of each bias transistor receives an independently bias voltage, and the drain is connected to a linear termination circuit; each input transistor is configured to operate in the transition range from the subthreshold region, the weak inversion region to the strong inversion region, to generate a nonlinear compensation current signal when the input amplitude exceeds the equivalent turn-on start point; each bias transistor is configured to operate in the saturation region within a predetermined output swing range; the independently adjustable control voltage is used to adjust the equivalent turn-on start point without changing the intrinsic threshold voltage of the input transistor.

[0012] According to an embodiment of this application, the nonlinear amplification unit further includes multiple adjustable control ports; the adjustable control ports are connected one-to-one with the source of the input transistors, and the adjustable control ports are used to provide independently adjustable control voltages to the corresponding input transistors to independently adjust the current weight of each nonlinear unit and adapt to the compensation requirements of the silicon photonic modulator under different distortion levels.

[0013] According to an embodiment of this application, the linear termination circuit includes a resistive load. The first end of the resistive load is connected to the output terminal of the linear amplifier unit and the output terminal of the nonlinear amplifier unit, respectively, and the second end of the resistive load is connected to a reference voltage terminal. The resistive load is configured to perform ohmic conversion on the superimposed current signal flowing through it, so as to convert the superimposed current signal in current form into a nonlinear voltage signal in voltage form, which can be used as a voltage signal that can directly drive a silicon photonic modulator.

[0014] According to embodiments of this application, the linear termination circuit includes a passive polysilicon resistive load, or an active transimpedance switching circuit with a low input impedance current receiving end.

[0015] According to the nonlinear compensation drive circuit provided in this application, the input voltage signal is converted into a linear current signal by a transconductance amplification unit; the linear amplification unit linearly amplifies the linear current signal to construct the basic driving component required to drive the silicon photonic modulator, while the nonlinear amplification unit processes the same linear current signal in parallel to generate a nonlinear compensation current signal that has a specific nonlinear relationship with the input voltage amplitude, thereby mapping and inversely canceling the inherent nonlinear distortion characteristics of the silicon photonic modulator; the linear termination circuit superimposes the amplified linear current signal and the nonlinear compensation current signal and converts them into a nonlinear voltage signal output, using the nonlinear characteristics to pre-distort the response curve of the silicon photonic modulator, thereby suppressing harmonic distortion in a wide dynamic range and improving the linearity and reliability of the silicon photonic modulator signal transmission. Attached Figure Description

[0016] The above and other objects, features and advantages of this application will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0017] Figure 1 A schematic diagram of a nonlinear compensation drive circuit according to an embodiment of this application is shown.

[0018] Figure 2 The nonlinear current-mode logic drive circuit of a silicon photonic modulator according to an embodiment of this application is schematically shown;

[0019] Figure 3 A schematic diagram illustrates a nonlinear push-pull drive circuit for a silicon photonic modulator according to an embodiment of this application;

[0020] Figure 4 The diagram schematically illustrates the output photoelectric sensor of a micro-ring modulator nonlinear drive circuit according to an embodiment of this application.

[0021] Figure 5 This illustration schematically shows the uncompensated output optical eye diagram of a micro-ring modulator according to an embodiment of this application;

[0022] Figure 6 An overcompensated output optical eye diagram of a microring modulator according to an embodiment of this application is schematically shown;

[0023] Figure 7 The diagram illustrates the output optical eye diagram of a micro-ring modulator according to an embodiment of this application under appropriate compensation.

[0024] Figure 8 The schematic diagram illustrates the output power spectrum of the microring under different bias voltages according to embodiments of this application;

[0025] Figure 9 A schematic diagram of a nonlinear NMOS driving unit according to an embodiment of this application is shown;

[0026] Figure 10 A schematic diagram of a non-linear drive operating mode according to an embodiment of this application is shown. Detailed Implementation

[0027] The embodiments of this application will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of this application. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of this application for ease of explanation. However, it will be apparent that one or more embodiments may be implemented without these specific details. Furthermore, descriptions of well-known structures and technologies are omitted in the following description to avoid unnecessarily obscuring the concepts of this application.

[0028] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. The terms “comprising,” “including,” etc., as used herein indicate the presence of features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0029] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0030] With the rapid development of Artificial Intelligence (AI), High Performance Computing (HPC), and large-scale models, computing centers are evolving from clusters with tens of thousands of GPUs to clusters with hundreds of thousands of GPUs. Traditional electrical interconnects are limited by bandwidth and power consumption in high-speed transmission, making it difficult to meet the needs of future large-scale intelligent computing systems. Silicon-based optical interconnect technology, due to its high bandwidth density, low power consumption, and good integration, has become an important development direction for realizing the interconnection of high-performance computing systems.

[0031] However, silicon photonic modulators generally suffer from inherent static nonlinearity. The relationship between their voltage and effective refractive index typically exhibits a Lorentz-like or sinusoidal curve, causing the mapping characteristics between the output optical signal and the input electrical signal to deviate from a linear relationship. When using multi-level signal formats such as four-level pulse amplitude modulation (PAM4), the nonlinearity of silicon photonic modulators leads to uneven output level spacing, signal distortion, and eye diagram closure, thereby increasing the bit error rate and limiting system speed and transmission performance.

[0032] In practical applications, existing silicon photonic modulator driving circuits struggle to simultaneously provide linear drive components and configurable nonlinear compensation components in the current domain. Consequently, they are unable to effectively compensate for static nonlinear distortion, leading to a decline in the quality of multilevel signal transmission. How to compensate for the static nonlinearity of silicon photonic modulators at the circuit level in a low-cost, low-power, and easily integrated manner remains a key unsolved technical problem.

[0033] If a controllable nonlinear electrical signal output can be directly achieved at the driver end, so that the optical signal modulated by the silicon photonics modulator has a linear level interval distribution, the modulation performance of the system can be significantly improved, and the linearity and bit error rate performance of high-speed PAM4 and higher-order multilevel signals can be enhanced.

[0034] To address the inherent static nonlinearity problem in silicon photonic modulators during modulation, this application discloses a nonlinear compensation drive circuit for high-speed optical communication and optical interconnection.

[0035] Figure 1 A schematic diagram of a nonlinear compensation drive circuit according to an embodiment of this application is shown. The nonlinear compensation drive circuit includes: a transconductance amplification unit 110, a linear amplification unit 120, a nonlinear amplification unit 130, and a linear termination circuit 140.

[0036] The transconductance amplifier unit is used to amplify the input voltage signal and convert it into a linear current signal.

[0037] The linear amplifier unit, connected to the output of the transconductance amplifier unit, is used to linearly amplify the linear current signal to obtain the amplified linear current signal.

[0038] The nonlinear amplifier unit, connected to the output of the transconductance amplifier unit, is used to perform nonlinear amplification based on the linear current signal to generate a nonlinear compensated current signal whose amplitude is nonlinearly related to the input voltage signal.

[0039] The linear termination circuit is connected to the output terminals of the linear amplifier unit and the nonlinear amplifier unit, respectively. It is used to superimpose the amplified linear current signal with the nonlinear compensation current signal to obtain a superimposed current signal. The superimposed current signal is then converted into a nonlinear voltage signal and output to the silicon photonic modulator to compensate for the nonlinear distortion of the silicon photonic modulator.

[0040] In this embodiment, the transconductance amplification unit serves as a signal input stage, receiving an external input voltage signal. Its internal bias design ensures that the core amplification device operates in the saturation region, thereby obtaining a stable transconductance gain. It amplifies the input voltage signal and converts it into a linear current signal that is linearly related to the voltage amplitude. This signal serves as the common input basis for the subsequent two amplification stages.

[0041] Transconductance gain is a circuit parameter that characterizes the ability of an electronic device (such as a transistor) to convert changes in input voltage into changes in output current. It is defined as the ratio of the change in output current to the change in input voltage.

[0042] The linear amplifier unit is connected to the output of the transconductance amplifier unit. It adopts a low input impedance structure (e.g., impedance is lower than a preset value) to linearly buffer and amplify the input linear current signal. Without introducing additional nonlinear distortion, it obtains an amplified linear current signal with enhanced driving capability and transmits it to the linear termination circuit.

[0043] The nonlinear amplifier unit is also connected to the output of the transconductance amplifier unit and processes the same linear current signal in parallel with the linear amplifier unit. Internally, through a device operating in the transition region between the cutoff region and the saturation region, it performs amplitude-dependent nonlinear amplification of the input signal to generate a nonlinear compensation current signal that is nonlinearly related to the amplitude of the input voltage signal. Moreover, this nonlinear characteristic can be configured to be opposite to the inherent nonlinear distortion trend of the silicon photonic modulator.

[0044] The linear termination circuit is connected to the output terminals of the linear amplification unit and the nonlinear amplification unit respectively. It directly superimposes the received amplified linear current signal and the nonlinear compensation current signal in the current domain to obtain a superimposed current signal containing linear and nonlinear components. Then, through ohmic conversion or transimpedance conversion, the superimposed current signal is converted into a nonlinear voltage signal with predistortion characteristics, and finally output to the silicon photonic modulator. The nonlinear voltage signal is used to cancel the nonlinear transmission characteristics of the silicon photonic modulator, thereby realizing static nonlinear compensation.

[0045] The nonlinear compensation drive circuit in this embodiment combines a nonlinear amplification unit with a linear amplification unit. By superimposing a linear current signal with an adjustable nonlinear compensation current signal, it effectively corrects the static nonlinearity of the silicon photonic modulator, thereby improving the modulation linearity and high-speed transmission performance. This circuit can be extended to a push-pull structure to compensate for the static nonlinearity of the silicon photonic modulator under various conditions.

[0046] In this embodiment, the input voltage signal is converted into a linear current signal by a transconductance amplification unit. The linear amplification unit amplifies the linear current signal linearly to construct the basic driving component required to drive the silicon photonic modulator. Meanwhile, the nonlinear amplification unit processes the same linear current signal in parallel to generate a nonlinear compensation current signal that has a specific nonlinear relationship with the input voltage amplitude, thereby mapping and inversely canceling the inherent nonlinear distortion characteristics of the silicon photonic modulator. The linear termination circuit superimposes the amplified linear current signal and the nonlinear compensation current signal and converts them into a nonlinear voltage signal for output. The nonlinear characteristics are used to pre-distort the response curve of the silicon photonic modulator, thereby suppressing harmonic distortion in a wide dynamic range and improving the linearity and reliability of the signal transmission of the silicon photonic modulator.

[0047] In some embodiments, the transconductance amplification unit includes: a first PMOS transistor and a second PMOS transistor, wherein the gate of the first PMOS transistor receives a first differential input positive signal, and the gate of the second PMOS transistor receives a first differential input negative signal; the sources of the first PMOS transistor and the second PMOS transistor are connected together and are connected to a first current source; a first NMOS transistor and a second NMOS transistor, wherein the gate of the first NMOS transistor receives a second differential input positive signal, and the gate of the second NMOS transistor receives a second differential input negative signal; the sources of the first NMOS transistor and the second NMOS transistor are connected together and are connected to a second current source; wherein the first differential input positive signal and the first differential input negative signal constitute a first differential input signal, the second differential input positive signal and the second differential input negative signal constitute a second differential input signal, and the voltage amplitude of the first differential input signal is greater than the voltage amplitude of the second differential input signal.

[0048] In this embodiment, the transconductance amplification unit adopts a Class A linear amplifier structure, and by setting two sets of differential input pairs in the high voltage domain and the low voltage domain, it takes into account both the requirements of high swing and linearity.

[0049] In Class A, all transistors in the amplifier always operate in the saturation region, providing a continuous and stable linear current output.

[0050] For transistors, the saturation state is entirely determined by the relative magnitudes of their gate-source and drain-source voltages. A transistor operates in the saturation region when the drain-source voltage is not less than the overdrive voltage. In this state, the drain current is primarily controlled by the gate-source voltage, exhibiting a relatively constant current characteristic, commonly used for signal amplification or as a constant current source.

[0051] The high-voltage domain differential input pair is composed of a first PMOS transistor and a second PMOS transistor. The gate of the first PMOS transistor receives the first differential input positive signal, and the gate of the second PMOS transistor receives the first differential input negative signal. Together, they constitute the first differential input signal and have a high voltage amplitude. The sources of the two transistors are connected and then connected to a first current source, which is specifically a tail current source, used to provide a stable DC bias current for the PMOS transistors to ensure that the differential pair operates in a balanced state.

[0052] The low-voltage domain differential input pair consists of a first NMOS transistor and a second NMOS transistor. The gate of the first NMOS transistor receives the second differential input positive signal, and the gate of the second NMOS transistor receives the second differential input negative signal. Together, they constitute the second differential input signal, and its voltage amplitude is less than that of the first differential input signal. The sources of the two transistors are connected and then connected to a second current source. Specifically, the second current source is a pull-up bias current source, which is used to provide a suitable bias voltage and current path for the NMOS transistor to ensure that the NMOS transistor operates in the saturation region.

[0053] Through the embodiments of this application, the transconductance amplifier unit can convert two sets of differential input voltage signals into linear current signals over a wide input voltage range, providing a unified current domain processing basis for subsequent linear and nonlinear amplifier units. Furthermore, the symmetrical design of the two differential pairs helps to suppress common-mode noise and improve signal integrity.

[0054] In some embodiments, the transconductance amplification unit further includes: a bias circuit, which is connected to the drain and gate of the first PMOS transistor, the second PMOS transistor, the first NMOS transistor, and the second NMOS transistor, respectively, and provides a bias voltage and a bias current, so that the first PMOS transistor, the second PMOS transistor, the first NMOS transistor, and the second NMOS transistor all operate in the saturation region, so as to amplify the input voltage signal and convert it into a linear current signal through transconductance gain.

[0055] In this embodiment, the transconductance amplification unit further includes a bias circuit, which provides precise and stable bias voltage and bias current to each transistor by being connected to the first PMOS transistor, the second PMOS transistor, the first NMOS transistor, and the second NMOS transistor, respectively.

[0056] The bias circuit typically employs a current mirror structure combined with a resistor divider network to ensure that all four input transistors are biased into the saturation region. When a transistor operates in the saturation region, its drain current is proportional to the square of the gate-source voltage. Through preset size design and bias point selection, the transconductance amplifier unit can exhibit a constant transconductance gain within a set input voltage range, thereby linearly amplifying the input voltage signal and converting it into a linear current signal. This avoids signal compression or distortion caused by the transistor entering the cutoff or linear region, ensuring the linearity and accuracy of the signal conversion.

[0057] In some embodiments, the output terminal of the transconductance amplification unit includes a first PMOS output node, a second PMOS output node, a first NMOS output node, and a second NMOS output node;

[0058] The first PMOS output node and the second PMOS output node are respectively formed by the drains of the first PMOS transistor and the second PMOS transistor, and are respectively connected to the corresponding input terminals of the linear amplifier unit and the nonlinear amplifier unit.

[0059] The first NMOS output node and the second NMOS output node are respectively formed by the drains of the first NMOS transistor and the second NMOS transistor, and are respectively connected to the corresponding input terminals of the linear amplifier unit and the nonlinear amplifier unit.

[0060] Among them, the first PMOS output node, the second PMOS output node, the first NMOS output node and the second NMOS output node are electrically isolated from each other, and the input impedance presented by the linear amplifier unit at each output node is less than the input impedance presented by the nonlinear amplifier unit at the corresponding output node.

[0061] In this embodiment, regarding the specific structure of the output terminal, the first PMOS output node is led out from the drain of the first PMOS transistor, and the second PMOS output node is led out from the drain of the second PMOS transistor. Similarly, the first NMOS output node is led out from the drain of the first NMOS transistor, and the second NMOS output node is led out from the drain of the second NMOS transistor.

[0062] In the actual circuit layout, these four output nodes are kept independent of each other in terms of physical layer and electrical connection. That is, there is no direct short circuit connection between the first PMOS output node, the second PMOS output node, the first NMOS output node and the second NMOS output node, so as to ensure the independence of signal transmission.

[0063] In terms of connectivity, the first PMOS output node and the first NMOS output node are both connected to the corresponding input terminals of the linear amplifier unit, forming either the inverting or non-inverting input of the differential signal; the second PMOS output node and the second NMOS output node are both connected to the corresponding input terminals of the nonlinear amplifier unit, forming another set of differential signal inputs. This connection method allows the current signal generated by the transconductance amplifier unit to be distributed to two different signal processing paths in the subsequent stage as needed.

[0064] Regarding impedance matching design, linear amplifier units are designed to present relatively low input impedance at each output node. For example, the input stage of a linear amplifier unit can employ a common-source configuration or an amplifier structure with local feedback, making its input impedance significantly lower than that of a nonlinear amplifier unit. Conversely, nonlinear amplifier units present relatively high input impedance at each output node.

[0065] By utilizing the low input impedance of the linear amplifier unit, it can effectively absorb the main current component from the transconductance amplifier unit, thereby dominating the gain in the low-frequency or linear region; while the high-impedance nonlinear amplifier unit is not sensitive to current changes and is mainly used to extract the nonlinear component in the output voltage swing of the transconductance amplifier unit or to perform specific nonlinear transformations, and will not significantly affect the signal processing of the linear amplifier unit due to the current shunting effect.

[0066] Through the embodiments of this application, the impedance differentiation design not only ensures the high bandwidth transmission performance of the linear amplifier unit, but also provides a stable signal sampling basis for the nonlinear amplifier unit, realizing the coordinated optimization of the main signal path and the compensation path.

[0067] In some embodiments, the linear amplification unit includes:

[0068] The gates of the third NMOS transistor and the fourth NMOS transistor are both connected to the NMOS common bias voltage terminal. The sources of the third NMOS transistor and the fourth NMOS transistor are respectively connected to the first NMOS output node and the second NMOS output node. The drains of the third NMOS transistor and the fourth NMOS transistor are respectively connected to the first differential output node and the second differential output node.

[0069] The third PMOS transistor and the fourth PMOS transistor have their gates connected to the PMOS common bias voltage terminal, their sources connected to the first PMOS output node and the second PMOS output node, and their drains connected to the first differential output node and the second differential output node, respectively.

[0070] The drains of the third PMOS transistor and the third NMOS transistor are connected together to the first differential output node, and the drains of the fourth PMOS transistor and the fourth NMOS transistor are connected together to the second differential output node; the third NMOS transistor, the fourth NMOS transistor, the third PMOS transistor and the fourth PMOS transistor are all configured to operate in the saturation region within the signal swing range to transmit linear current signals.

[0071] In the embodiments of this application, the linear amplification unit is implemented using a cascaded common gate structure, which is composed of a third NMOS transistor, a fourth NMOS transistor, a third PMOS transistor, and a fourth PMOS transistor to form a push-pull amplification path.

[0072] This linear amplifier unit specifically includes a pair of NMOS current-transfer transistors (i.e., the third and fourth NMOS transistors) and a pair of PMOS current-transfer transistors (i.e., the third and fourth PMOS transistors). In terms of bias design, the gates of the third and fourth NMOS transistors are connected to the common NMOS bias voltage terminal to obtain a stable quiescent operating point; similarly, the gates of the third and fourth PMOS transistors are connected to the common PMOS bias voltage terminal. This common-gate bias method ensures that the two transistors have consistent transconductance characteristics, which helps to improve the amplifier's common-mode rejection ratio and linearity.

[0073] In terms of connectivity, the source of the third NMOS transistor is connected to the first NMOS output node, and the source of the fourth NMOS transistor is connected to the second NMOS output node, thus directly receiving the current signal from the NMOS differential pair of the transconductance amplifier unit. Simultaneously, the source of the third PMOS transistor is connected to the first PMOS output node, and the source of the fourth PMOS transistor is connected to the second PMOS output node, to receive the current signal from the PMOS differential pair of the transconductance amplifier unit. The drains of these transistors are cross-coupled to the subsequent load or output stage: the drains of the third NMOS transistor and the third PMOS transistor are jointly connected to the first differential output node, and the drains of the fourth NMOS transistor and the fourth PMOS transistor are jointly connected to the second differential output node. This push-pull connection structure allows the PMOS current source and the NMOS current source to be superimposed at the output, jointly driving the differential voltage swing between the first and second differential output nodes.

[0074] In terms of operating mode, the third NMOS transistor, the fourth NMOS transistor, the third PMOS transistor, and the fourth PMOS transistor are all configured to operate in the saturation region throughout the signal swing range. By directly connecting the input terminal of this linear amplifier unit (i.e., the source of each transistor) to the output node of the transconductance amplifier unit, and utilizing the weak source-following characteristic of the MOS transistor in the saturation region, the unit presents a low input impedance at each output node. This low input impedance characteristic can effectively absorb the current signal generated by the transconductance amplifier unit and convert it into a voltage signal output, while avoiding additional nonlinear load effects on the preceding transconductance amplifier unit caused by large fluctuations in node voltage.

[0075] In some embodiments, the sources of the third NMOS transistor, the fourth NMOS transistor, the third PMOS transistor, and the fourth PMOS transistor respectively constitute the first NMOS low-resistance input node, the second NMOS low-resistance input node, the first PMOS low-resistance input node, and the second PMOS low-resistance input node.

[0076] The drains of the third PMOS transistor and the third NMOS transistor together form the first differential output node, and the drains of the fourth PMOS transistor and the fourth NMOS transistor together form the second differential output node. The first differential output node and the second differential output node are connected to the linear termination circuit.

[0077] The input impedance of each low-impedance input node is less than the output impedance of the first differential output node and the second differential output node.

[0078] In terms of connection, the first NMOS low-impedance input node is connected to the first NMOS output node of the transconductance amplifier unit, the second NMOS low-impedance input node is connected to the second NMOS output node of the transconductance amplifier unit, the first PMOS low-impedance input node is connected to the first PMOS output node of the transconductance amplifier unit, and the second PMOS low-impedance input node is connected to the second PMOS output node of the transconductance amplifier unit. This one-to-one connection allows the four current signals output from the transconductance amplifier unit to be injected into the four low-impedance input nodes of the linear amplifier unit. Since the source (for NMOS) or source (for PMOS) of the MOS transistor exhibits a low AC impedance when operating in the saturation region, each low-impedance input node constitutes a low-impedance load relative to the transconductance amplifier unit, effectively absorbing the linear current signal generated by the transconductance amplifier unit without causing significant voltage swing loss.

[0079] In the output configuration, the drains of the third PMOS transistor and the third NMOS transistor are connected together to form the first differential output node; the drains of the fourth PMOS transistor and the fourth NMOS transistor are connected together to form the second differential output node. These two differential output nodes serve as the voltage output terminals of the linear amplifier unit and are connected to the subsequent linear termination circuit. In this embodiment, both the first and second differential output nodes exhibit high output impedance, significantly higher than the input impedance of each low-impedance input node. Specifically, the input impedance of each low-impedance input node is less than the output impedance of the first and second differential output nodes.

[0080] The embodiments of this application optimize the signal transmission characteristics of the linear amplifier unit, ensuring both the performance of the transconductance amplifier unit and the driving requirements of the linear termination circuit.

[0081] In some embodiments, the nonlinear amplification unit includes a plurality of NMOS nonlinear units and a plurality of PMOS nonlinear units;

[0082] Each NMOS nonlinear unit and each PMOS nonlinear unit includes an input transistor and a bias transistor connected in series;

[0083] The gate of the input transistor of the PMOS nonlinear unit is connected to the corresponding PMOS output node, and the gate of the input transistor of the NMOS nonlinear unit is connected to the corresponding NMOS output node; the source of each input transistor receives an independently adjustable control voltage, and the drain is connected to the source of the corresponding bias transistor; the gate of each bias transistor receives an independent bias voltage, and the drain is connected to the linear termination circuit.

[0084] Each input transistor is configured to operate in the transition range from the subthreshold region, weak inversion region to strong inversion region, to generate a nonlinear compensation current signal when the input amplitude exceeds the equivalent turn-on start point; each bias transistor is configured to operate in the saturation region within a predetermined output swing range; an independently adjustable control voltage is used to adjust the equivalent turn-on start point without changing the intrinsic threshold voltage of the input transistor.

[0085] In this embodiment, the nonlinear amplification unit is implemented using a Class B structure, consisting of multiple NMOS nonlinear units and multiple PMOS nonlinear units connected in parallel, to cover nonlinear compensation requirements of different polarities and amplitudes.

[0086] In this context, Class B structure refers to a transistor in an amplifier that operates in the transition region between the cutoff and saturation regions, turning on only when the input signal exceeds a threshold to generate a nonlinear current.

[0087] For the PMOS nonlinear unit, the gate of its input transistor (PMOS) is connected to the corresponding PMOS output node of the transconductance amplifier unit (e.g., the first PMOS output node or the second PMOS output node) to receive the voltage signal from the preceding stage. For the NMOS nonlinear unit, the gate of its input transistor (NMOS) is connected to the corresponding NMOS output node of the transconductance amplifier unit (e.g., the first NMOS output node or the second NMOS output node). The source of each input transistor receives an independently adjustable control voltage to flexibly adjust the equivalent conduction characteristics of the transistor; the drain of each input transistor is connected to the source of the corresponding bias transistor. The gate of each bias transistor receives an independent bias voltage to set the static operating point of the nonlinear unit, and its drain is connected to the linear termination circuit of the subsequent stage to output the synthesized nonlinear compensation current signal.

[0088] In terms of operating mode configuration, each input transistor is specifically configured to operate in the transition range from the subthreshold region, weak inversion region to strong inversion region. Within this specific operating region, the transconductance of the transistor changes exponentially or superlinearly with the gate-source voltage. The nonlinear amplifier unit utilizes this characteristic to generate a nonlinear compensation current signal when the input voltage amplitude exceeds the equivalent conduction start point. This equivalent conduction start point is not determined by the intrinsic threshold voltage of the input transistor itself, but is adjusted by an independently adjustable control voltage applied to the source. By adjusting the magnitude of this control voltage, the voltage threshold at which the input transistor begins to generate a significant nonlinear current can be set, thus flexibly adapting to different amplitude input signals or different compensation requirements without changing the process parameters (i.e., without changing the intrinsic threshold voltage of the input transistor). Simultaneously, each bias transistor is configured to always operate in the saturation region within a predetermined output signal swing range, making it equivalent to a stable current source. This provides a constant source load to the input transistor and ensures a high impedance characteristic at the output node, avoiding load pulling effects on the output of the linear amplifier unit.

[0089] Through the embodiments of this application, the nonlinear amplification unit can utilize its high input impedance characteristics (derived from the high output impedance in the saturation region of the bias transistor and the gate connection of the input transistor) to extract and amplify the nonlinear components in the signal without significantly affecting the linear operation of the transconductance amplification unit. The application of an independently adjustable control voltage enables this nonlinear compensation mechanism to possess flexible tunability.

[0090] In some embodiments, the nonlinear amplification unit further includes a plurality of adjustable control ports;

[0091] The adjustable control port is connected one-to-one with the source of the input transistor. The adjustable control port is used to provide an independently adjustable control voltage to the corresponding input transistor to independently adjust the current weight of each nonlinear unit and adapt to the compensation requirements of the silicon photonic modulator under different distortion levels.

[0092] In this embodiment, the nonlinear amplification unit also integrates multiple adjustable control ports. Each adjustable control port is connected to the source of the input transistor in at least one nonlinear unit, and is used to provide an independently adjustable control voltage to the corresponding input transistor.

[0093] Adjustable control ports are typically implemented by on-chip digital control circuitry driving a digital-to-analog converter, enabling the output of differentiated control voltages for different nonlinear units. By raising or lowering the potential of the input transistor's source, the bias point of the transistor can be precisely adjusted, thereby changing its transconductance characteristics and turn-on threshold, achieving independent and fine adjustment of the current weights of each nonlinear unit.

[0094] Through the embodiments of this application, the independently adjustable mechanism enables the nonlinear amplification unit to be flexibly configured with various nonlinear compensation curves such as compressed, extended, or piecewise linear, thereby adaptively matching the specific nonlinear distortion characteristics of silicon photonic modulators under different process, voltage, and temperature conditions, significantly improving the applicability and compensation accuracy of the driving circuit.

[0095] In some embodiments, the linear termination circuit includes a resistive load, the first end of which is connected to the output of the linear amplifier unit and the output of the nonlinear amplifier unit, respectively, and the second end of which is connected to the reference voltage terminal.

[0096] The resistive load is configured to perform ohmic conversion on the superimposed current signal flowing through it, so as to convert the superimposed current signal in current form into a nonlinear voltage signal in voltage form, which can be used as a voltage signal to directly drive the silicon photonic modulator.

[0097] In this embodiment, the linear terminal circuit uses a resistive load as the core conversion element. Its first end serves as a bus node, which is connected to the output of the linear amplifier unit and the output of the nonlinear amplifier unit, respectively. It is used to receive the amplified linear current signal and the nonlinear compensation current signal, and to directly superimpose the two in the current domain to form a superimposed current signal.

[0098] The second terminal of the resistive load is connected to a reference voltage terminal, which is typically the power supply voltage, ground potential, or a specific bias voltage, to provide a DC operating point for the output signal.

[0099] A resistive load converts a superimposed current signal into a nonlinear voltage signal. This conversion process not only achieves the transformation from current to voltage, but also appropriately limits and shapes the signal bandwidth through the impedance characteristics of the resistor. This ensures that the nonlinear voltage signal output to the silicon photonics modulator has both a suitable voltage swing and good signal integrity, thereby effectively driving the silicon photonics modulator and compensating for its nonlinear distortion.

[0100] In some embodiments, the linear termination circuit includes a passive polysilicon resistive load or an active transimpedance converter circuit with a low input impedance current receiver.

[0101] In the embodiments of this application, the linear termination circuit can be implemented using a passive polysilicon resistive load, or it can be implemented using an active transimpedance conversion circuit with a low input impedance current receiving end.

[0102] For example, the linear termination circuit specifically employs a passive polysilicon resistor load. The first and second differential output nodes are each connected to a preset power supply voltage terminal or reference ground terminal through one or more series-connected passive polysilicon resistors. Polysilicon resistors have the advantages of good process compatibility, controllable temperature coefficient, and low parasitic capacitance, effectively converting the superimposed current signal at the first and second differential output nodes (i.e., the vector sum of the amplified linear current signal and the nonlinear compensation current signal) into a differential voltage signal. This differential voltage signal is directly applied to the differential input terminal of the silicon photonics modulator. Utilizing the linear volt-ampere characteristic of the resistor, the linear conversion relationship between the output voltage and input current is ensured while simplifying the circuit structure, thereby accurately compensating for the nonlinear distortion of the silicon photonics modulator. Furthermore, the high resistance of the polysilicon resistor helps to improve the voltage swing, meeting the drive amplitude requirements of the silicon photonics modulator.

[0103] For example, the linear termination circuit specifically employs an active transimpedance converter circuit with a low input impedance current receiver. This active transimpedance converter circuit typically includes a transimpedance amplifier and a feedback impedance network. The input terminal of the transimpedance amplifier (i.e., the low input impedance current receiver) is connected to the first differential output node and the second differential output node, respectively, to receive the superimposed current signals from the linear amplification unit and the nonlinear amplification unit. Because this input terminal exhibits extremely low input impedance, it can effectively suppress the influence of the output impedance of the preceding circuit on signal synthesis, ensuring high-fidelity current superposition of the linear current signal and the nonlinear compensation current signal without significant voltage fluctuations. Subsequently, the transimpedance amplifier, utilizing its high-gain characteristics, converts the superimposed current signal into a voltage signal and outputs it through the feedback impedance network. The advantage of this active implementation is that it can provide a larger dynamic range, higher conversion gain, and stronger driving capability, making it particularly suitable for applications requiring high-speed, large-signal-swing driving of silicon photonic modulators.

[0104] Through the embodiments of this application, both implementation methods can superimpose the current signals output by the linear amplification unit and the nonlinear amplification unit and convert them into the required nonlinear voltage signal to adapt to different system integration and performance requirements.

[0105] refer to Figure 2 , Figure 2 The diagram illustrates a nonlinear current-mode logic drive circuit for a silicon photonic modulator according to an embodiment of this application.

[0106] The circuit includes a differential pair consisting of a first NMOS transistor MN1 and a second NMOS transistor MN2, whose gates receive the second differential input positive signal VIPL and the second differential input negative signal VINL, respectively, and whose sources are connected to the second current source I1; a common-gate structure consisting of a third NMOS transistor MN3 and a fourth NMOS transistor MN4, whose gates are both connected to the common bias voltage terminal Vbias1, whose sources are connected to the output node of the differential pair, and whose drains are connected to the output node; and multiple parallel nonlinear compensation paths, each consisting of an input transistor and a bias transistor connected in series, such as MNLNx, MNLN1, and MN. LPx and MNLP1 are bias transistors such as MNCNx, MNCN1, MNCPx, and MNCP1 connected in series. The gate of the input transistor is connected to the output node of the differential pair, and the source receives an adjustable control voltage, such as VCNLx, VCNL1, VCPLx, and VCPL1. The gate of the bias transistor receives an independent bias voltage VbiasN, and the drain is connected to the output node. The first terminal of the resistive load RT is connected to the output terminals of the common-gate structure and the nonlinear compensation path, respectively. The second terminal serves as a reference voltage terminal to convert the superimposed current signal into a voltage signal, which is finally output to the silicon photonic modulator through the differential output terminals VOP and VON.

[0107] Figure 3 The schematic diagram illustrates a nonlinear push-pull drive circuit for a silicon photonic modulator according to an embodiment of this application.

[0108] The circuit adopts a symmetrical push-pull architecture, mainly composed of a high-voltage input stage, a low-voltage input stage, a nonlinear compensation unit, and a common-gate output stage. The high-voltage input stage consists of a first PMOS transistor MP1 and a second PMOS transistor MP2. Their gates receive the first differential input positive signal VIPH and the first differential input negative signal VINH, respectively. Their sources are connected to the first current source I2, i.e., the pull-up bias current source. The drains are connected to the differential output terminals VOP and VON through high-voltage coupling / compensation capacitors Cp1 and Cp2. The low-voltage input stage consists of a first NMOS transistor MN1 and a second NMOS transistor MN2. Their gates receive the second differential input positive signal VIPL and the second differential input negative signal VINL, respectively. Their sources are connected to the second current source I1, i.e., the tail current source. The drains are connected to the differential output terminals VOP and VON through low-voltage coupling / compensation capacitors Cn1 and Cn2. The output stage consists of a common-gate linear path formed by the third NMOS transistor MN3, the fourth NMOS transistor MN4, the third PMOS transistor MP3, and the fourth PMOS transistor MP4. The gates of all four transistors are connected to the common bias voltage terminal Vbias1. The drains of MP3 and MP4 are connected to the high-voltage output node of the differential output terminal VOP, while the drains of MN3 and MN4 are connected to the low-voltage output node of the differential output terminal VON. The first terminals of resistive loads RT1 and RT2 are connected to the high-voltage output node of the differential output terminal VOP and the low-voltage output node of the differential output terminal VON, respectively, while the second terminals are grounded as reference voltage terminals. To provide nonlinear compensation, two parallel nonlinear units are set up in the circuit: the PMOS nonlinear unit NLPU within the dashed box above contains multiple parallel compensation branches. Each branch consists of an input transistor and a bias transistor connected in series. Examples of input transistors include MPHNx, MPHN1, MPHPx, and MPHP1, and examples of bias transistors include MPCNx, MPCN1, MPCPx, and MPCP1. The gate of the input transistor is connected to the drain of the high-voltage input stage, and the source receives the branch control terminal voltage, such as VCNHx, VCNH1, VCPHx, and VCPH1. The bias transistor... The gate of the NLNU receives an independent bias voltage VbiasP; the NMOS nonlinear unit NLNU structure within the dashed box below is symmetrical to the NLPU and includes an input transistor and a bias transistor. The input transistors are, for example, MNLNx, MNLN1, MNLPx, and MNLP1, and the bias transistors are, for example, MNCNx, MNCN1, MNCPx, and MNCP1. The gate of the input transistor is connected to the drain of the low-voltage input stage, and the source receives the branch control terminal voltage, for example, VCNLx, VCNL1, VCPLx, and VCPL1. The gate of the bias transistor receives an independent bias voltage VbiasN.In addition, the circuit includes multiple bias resistors, namely Rbias1 to Rbias4, and bias decoupling capacitors, namely Cb1 to Cb4, to provide stable DC bias and noise suppression for each transistor. The ellipsis indicates an expandable parallel branch of the same type.

[0109] Figure 4 The diagram illustrates the output photoelectric diagram of a micro-ring modulator nonlinear drive circuit according to an embodiment of this application.

[0110] This figure shows the differential voltage eye diagram of a 64 Gb / s PAM4 signal after passing through a nonlinear drive circuit. The horizontal axis represents Time (ps), and the vertical axis represents Voltage (mV). Multiple superimposed tracks indicate different code pattern conversions. The gradually changing eye height from high to low level is used to create electrical predistortion that is opposite to the static nonlinearity of the micro-ring modulator.

[0111] Figure 5 The diagram illustrates the uncompensated output optical eye diagram of a micro-ring modulator according to an embodiment of this application.

[0112] This figure shows the output optical eye diagram obtained by directly modulating a microring with a 64Gb / s PAM4 signal without nonlinear compensation. The horizontal axis represents Time (ps), and the vertical axis represents Optical Power (µW). Each horizontal trajectory corresponds to one of the four optical power levels of PAM4, and the intersection of the trajectories indicates the transition between adjacent levels. The non-uniform three-level eye height reflects the static nonlinearity of the microring modulator.

[0113] Figure 6 An overcompensated output optical eye diagram of a microring modulator according to an embodiment of this application is illustrated schematically.

[0114] This figure shows the output optical eye diagram of a 64 Gb / s PAM4 when the micro-ring modulator is overcompensated by a nonlinear drive circuit. The horizontal axis represents Time (ps), and the vertical axis represents Optical Power (µW). Each horizontal trajectory corresponds to four optical power levels. The three eye heights show an inverse non-uniform distribution from high to low optical power, indicating that the compensation intensity exceeds the target value.

[0115] Figure 7 The diagram illustrates the output optical eye diagram of a micro-ring modulator according to an embodiment of this application with appropriate compensation.

[0116] This figure shows the output optical eye diagram of a 64 Gb / s PAM4 optical circuit with appropriate compensation parameters for the nonlinear drive circuit. The horizontal axis represents Time (ps), and the vertical axis represents Optical Power (µW). Each horizontal trajectory corresponds to four optical power levels. The three eye heights are basically balanced, indicating that the electrical predistortion and the micro-ring static nonlinearity cancel each other out.

[0117] To gain a clearer understanding of the compensation effect of this circuit structure on the nonlinearity of the silicon photonic modulator, a micro-ring modulator will be used as an example to conduct mathematical derivation and simulation analysis of the mechanism of micro-ring nonlinearity and the nonlinear driving circuit.

[0118] First, the relationship between the output optical power of the microring resonator and the input wavelength can be described by a Lorentz-type transmission function, that is:

[0119]

[0120] in, The output optical power of the micro-ring modulator is equal to the input optical wavelength. With modulation voltage The function, This represents the incident light power input to the micro-ring modulator. The internal loss rate of the microring is... The total attenuation rate of the microrings. This represents the difference between the angular frequency of the input light and the angular frequency of the microring resonance, specifically:

[0121]

[0122] in, The speed of light in a vacuum. The resonant center wavelength of the microring is . For the input light wavelength, λ is the resonant center wavelength of the microring.

[0123] Figure 8 The diagram schematically illustrates the output power spectrum of the microring under different bias voltages according to an embodiment of this application. The horizontal axis represents wavelength (µm), and the vertical axis represents optical power (µW), including 0V, 1V, and 2V. The diagram shows the curves of optical power versus wavelength under different modulation voltages. It can be seen that at the center frequency of 1310nm, there is a significant non-uniform power distribution across the three levels, indicating static nonlinearity of the microring.

[0124] Therefore, this application introduces a nonlinear driving unit to compensate for the nonlinearity of the microring.

[0125] Figure 9 A schematic diagram of a nonlinear NMOS driving unit according to an embodiment of this application is shown. Vin represents the input voltage signal, Vbias represents the DC bias voltage, and MN1 and MN2 are both NMOS transistors, illustrating the nonlinear NMOS driving unit. MN2 operates in the saturation region to maintain a high output impedance, suppress the Miller effect, and provide stable current transmission. MN1 receives the signal and amplifies the input signal current. Its operating state changes with the input voltage swing to construct a Class B nonlinear amplification. It only conducts when the input voltage is greater than the threshold voltage Vth, exhibiting linear amplification. When the voltage is lower than Vth, MN1 operates in the subthreshold region, resulting in nonlinear signal amplification, thus constructing the nonlinear unit. The nonlinear PMOS driving unit operates on the same principle. Therefore, by making the nonlinear unit into a configurable unit module and biasing it under different Class B operating states, different nonlinear driving waveforms can be obtained by configuring the weights.

[0126] Figure 10 A schematic diagram of a non-linear drive operating mode according to an embodiment of this application is shown.

[0127] This diagram illustrates the overall architecture of the nonlinear compensation drive circuit. At the center is an adder (symbol "⊕"), to the left is an NMOS nonlinear unit (NLNU), to the right is a PMOS nonlinear unit (NLPU), and directly below is a linear drive unit. The outputs of these three units are pointed to by arrows at the central adder, indicating that the signals are summed and superimposed here to obtain a superimposed current signal. The arrow pointing upwards from the adder indicates that the superimposed current signal is output from this terminal; for example, it could be output to a silicon photonics modulator to compensate for the nonlinear distortion of the modulator.

[0128] Those skilled in the art will understand that the features described in the various embodiments of this application can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in this application. In particular, the features described in the various embodiments of this application can be combined and / or combined in various ways without departing from the spirit and teachings of this application. All such combinations and / or combinations fall within the scope of this application.

[0129] The embodiments of this application have been described above. However, these embodiments are merely illustrative and not intended to limit the scope of this application. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Without departing from the scope of this application, those skilled in the art can make various substitutions and modifications, all of which should fall within the scope of this application.

Claims

1. A nonlinear compensation driving circuit for driving a silicon optical modulator, characterized by, include: The transconductance amplifier unit is used to amplify the input voltage signal and convert it into a linear current signal; A linear amplifier unit, connected to the output terminal of the transconductance amplifier unit, is used to linearly amplify the linear current signal to obtain an amplified linear current signal. A nonlinear amplification unit, connected to the output terminal of the transconductance amplification unit, is used to perform nonlinear amplification based on the linear current signal to generate a nonlinear compensation current signal that has a nonlinear relationship with the amplitude of the input voltage signal. A linear termination circuit is connected to the output terminals of the linear amplification unit and the nonlinear amplification unit, respectively. It is used to superimpose the amplified linear current signal with the nonlinear compensation current signal to obtain a superimposed current signal, convert the superimposed current signal into a nonlinear voltage signal and output it to the silicon photonic modulator to compensate for the nonlinear distortion of the silicon photonic modulator.

2. The non-linear compensation driving circuit according to claim 1, wherein, The transconductance amplification unit includes: A first PMOS transistor and a second PMOS transistor, wherein the gate of the first PMOS transistor receives a first differential input positive signal, and the gate of the second PMOS transistor receives a first differential input negative signal; the source of the first PMOS transistor and the source of the second PMOS transistor are connected together and connected to a first current source. A first NMOS transistor and a second NMOS transistor, wherein the gate of the first NMOS transistor receives a second differential input positive signal, and the gate of the second NMOS transistor receives a second differential input negative signal; the source of the first NMOS transistor and the source of the second NMOS transistor are connected together and connected to a second current source. Wherein, the first differential input positive signal and the first differential input negative signal constitute the first differential input signal, the second differential input positive signal and the second differential input negative signal constitute the second differential input signal, and the voltage amplitude of the first differential input signal is greater than the voltage amplitude of the second differential input signal.

3. The non-linear compensation driving circuit according to claim 2, wherein The transconductance amplification unit further includes: The bias circuit provides bias voltage and bias current to the first PMOS transistor, the second PMOS transistor, the first NMOS transistor, and the second NMOS transistor, so that the first PMOS transistor, the second PMOS transistor, the first NMOS transistor, and the second NMOS transistor all operate in the saturation region, so as to amplify the input voltage signal and convert it into a linear current signal through transconductance gain.

4. The nonlinear compensation drive circuit according to claim 2, characterized in that, The output terminal of the transconductance amplifier unit includes a first PMOS output node, a second PMOS output node, a first NMOS output node, and a second NMOS output node; The first PMOS output node and the second PMOS output node are respectively formed by the drains of the first PMOS transistor and the second PMOS transistor, and are respectively connected to the corresponding input terminals of the linear amplification unit and the nonlinear amplification unit; The first NMOS output node and the second NMOS output node are respectively formed by the drains of the first NMOS transistor and the second NMOS transistor, and are respectively connected to the corresponding input terminals of the linear amplification unit and the nonlinear amplification unit; The first PMOS output node, the second PMOS output node, the first NMOS output node, and the second NMOS output node are electrically isolated from each other, and the input impedance presented by the linear amplification unit is less than the input impedance presented by the nonlinear amplification unit at the corresponding output node.

5. The nonlinear compensation drive circuit according to claim 4, characterized in that, The linear amplification unit includes: The third NMOS transistor and the fourth NMOS transistor have their gates connected to the NMOS common bias voltage terminal, their sources connected to the first NMOS output node and the second NMOS output node, respectively, and their drains connected to the first differential output node and the second differential output node, respectively. The third PMOS transistor and the fourth PMOS transistor have their gates connected to the PMOS common bias voltage terminal, their sources connected to the first PMOS output node and the second PMOS output node, respectively, and their drains connected to the first differential output node and the second differential output node, respectively. The drains of the third PMOS transistor and the third NMOS transistor are connected together to the first differential output node, and the drains of the fourth PMOS transistor and the fourth NMOS transistor are connected together to the second differential output node; the third NMOS transistor, the fourth NMOS transistor, the third PMOS transistor, and the fourth PMOS transistor are all configured to operate in the saturation region within the signal swing range to transmit the linear current signal.

6. The nonlinear compensation drive circuit according to claim 5, characterized in that, The sources of the third NMOS transistor, the fourth NMOS transistor, the third PMOS transistor, and the fourth PMOS transistor respectively constitute the first NMOS low-resistance input node, the second NMOS low-resistance input node, the first PMOS low-resistance input node, and the second PMOS low-resistance input node. The drains of the third PMOS transistor and the third NMOS transistor together form the first differential output node, and the drains of the fourth PMOS transistor and the fourth NMOS transistor together form the second differential output node. The first differential output node and the second differential output node are connected to the linear termination circuit. The input impedance of each of the low-impedance input nodes is less than the output impedance of the first differential output node and the second differential output node.

7. The nonlinear compensation drive circuit according to claim 1, characterized in that, The nonlinear amplification unit includes multiple NMOS nonlinear units and multiple PMOS nonlinear units; Each of the NMOS nonlinear units and each of the PMOS nonlinear units includes an input transistor and a bias transistor connected in series; The gate of the input transistor of the PMOS nonlinear unit is connected to the corresponding PMOS output node, and the gate of the input transistor of the NMOS nonlinear unit is connected to the corresponding NMOS output node. The source of each input transistor receives an independently adjustable control voltage, and the drain is connected to the source of the corresponding bias transistor. The gate of each bias transistor receives an independent bias voltage, and the drain is connected to the linear termination circuit. Each of the input transistors is configured to operate in the transition range from the subthreshold region, the weak inversion region to the strong inversion region, so as to generate a nonlinear compensation current signal when the input amplitude exceeds the equivalent conduction start point; each of the bias transistors is configured to operate in the saturation region within a predetermined output swing range; The independently adjustable control voltage is used to adjust the equivalent turn-on start point without changing the intrinsic threshold voltage of the input transistor.

8. The nonlinear compensation drive circuit according to claim 7, characterized in that, The nonlinear amplification unit also includes multiple adjustable control ports; The adjustable control port is connected to the source of each input transistor in a one-to-one correspondence. The adjustable control port is used to provide an independently adjustable control voltage to the corresponding input transistor to independently adjust the current weight of each nonlinear unit and adapt to the compensation requirements of the silicon photonic modulator under different distortion levels.

9. The nonlinear compensation drive circuit according to claim 1, characterized in that, The linear termination circuit includes a resistive load, the first end of which is connected to the output terminal of the linear amplifier unit and the output terminal of the nonlinear amplifier unit, respectively, and the second end of which is connected to a reference voltage terminal. The resistive load is configured to perform ohmic conversion on the superimposed current signal flowing through it, so as to convert the superimposed current signal in current form into the nonlinear voltage signal in voltage form, which can be used as a voltage signal to directly drive the silicon photonic modulator.

10. The nonlinear compensation drive circuit according to claim 1, characterized in that, The linear termination circuit includes a passive polysilicon resistive load or an active transimpedance converter circuit with a low input impedance current receiver.