A second-order hybrid-residual feedback noise-shaping SAR analog-to-digital converter and a control method thereof

By combining a CDAC capacitor array, a SAR logic module, and an active-passive hybrid integrator, the noise shaping problem of the SAR analog-to-digital converter under the low-order quantization core is solved, achieving high-precision second-order noise shaping and effective resolution improvement, while reducing power consumption and loop delay.

CN123122540APending Publication Date: 2026-11-17INST OF MICROELECTRONICS CHINESE ACAD OF SCI LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611337587.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-31
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing SAR analog-to-digital converters, with their low-bit quantization cores, struggle to perform low-power, high-precision second-order cross-cycle noise shaping on the final residual after conversion without excessively increasing the CDAC scale, operational amplifier power consumption, and analog summing circuit complexity. Furthermore, they suffer from issues related to residual sampling timing, multi-input comparison summation, noise shaping mode switching, digital output word length matching, and analog/digital noise isolation.

Method used

The system employs a CDAC capacitor array, a SAR logic module, a residual sampling phase generation module, an active-passive hybrid second-order noise shaping integrator, and a multi-input Double-Tail dynamic comparator. After successive approximation conversion, the residual error of the CDAC top plate is sampled, and cross-cycle feedback is formed by passive switched capacitor integration and active-passive hybrid integration. Combined with non-overlapping phase timing and digital low-pass decimation, second-order noise shaping is achieved.

Benefits of technology

Without increasing the number of CDAC bits and the power consumption of the op-amp, the effective resolution within the target bandwidth is improved, quantization noise is reduced, loop delay and digital noise are reduced, and low-power, high-precision noise shaping is achieved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123122540A_ABST
    Figure CN123122540A_ABST
Patent Text Reader

Abstract

This invention discloses a second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter and its control method, belonging to the field of analog-to-digital converter technology. It includes a CDAC capacitor array, a SAR logic module, a residual sampling phase generation module, an active / passive hybrid second-order noise-shaping integrator, and a multi-input double-tail dynamic comparator. After successive approximation conversion, the conversion residual of the CDAC top plate is sampled, and after a first-stage passive integrator and a second-stage active / passive hybrid integrator, a two-stage integration state is formed. This state is then fed back to the SAR logic module along with the CDAC residual through a weighted comparison in the dynamic comparator. A CIC3 / sinc³ digital low-pass decimation backend and analog / digital power domain partitioning can also be configured. This invention improves the effective resolution within the target bandwidth while using a low-order SAR quantization core, and reduces the implementation pressure caused by the high-order CDAC, additional analog adders, and high-gain closed-loop operational amplifiers.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of integrated circuit and analog-to-digital conversion technology, specifically relating to a second-order hybrid residual feedback noise shaping SAR analog-to-digital converter and its control method. Background Technology

[0002] Analog-to-digital converters (ADCs) are key modules in sensor interfaces, precision measurement, audio acquisition, bioelectrical signal processing, and industrial monitoring systems. They are used to convert continuous-time, continuous-amplitude analog signals into discrete codewords that can be processed by digital systems. With the development of low-power IoT, wearable medical devices, and mid-to-low frequency high-precision sensor interfaces, ADCs need to achieve high in-band dynamic range under limited power consumption, limited chip area, and low supply voltage conditions.

[0003] Successive approximation register (SARADC) analog-to-digital converters typically utilize capacitive digital-to-analog converters (CDACs) and comparators to search for the input voltage bit by bit. They are characterized by simple structure, low dynamic power consumption, digital friendliness, and ease of scaling with process technology, making them suitable for medium- to low-speed data conversion. However, as the target resolution increases further, directly increasing the number of SAR quantization bits increases the total capacitance and number of capacitors in the CDAC. This places higher demands on capacitor matching, layout symmetry, parasitic parameters, reference voltage driving capability, and settling time, leading to increased area, driving power consumption, and implementation difficulty. Simultaneously, traditional SAR quantization noise is mainly distributed within the Nyquist bandwidth, making it difficult to reduce quantization noise within the target bandwidth through its own structure.

[0004] Noise-shaping SAR ADCs combine SAR conversion with oversampling and cross-cycle residual feedback. After each successive approximation conversion, the CDAC top-plate conversion residual is retained, and this residual is integrated, stored, and fed back. This shifts the quantization error out of band after passing through a high-pass noise transfer function, thereby improving the effective in-band resolution without directly increasing the SAR quantizer bit depth. Existing noise shaping schemes typically include passive noise shaping, active noise shaping, and combinations of both. Pure passive switched-capacitor integrators have low quiescent power consumption, but charge sharing leads to limited gain, integrator poles deviate from ideal positions, and low-frequency noise suppression is limited. Pure active schemes can improve integration accuracy, but require amplifiers with higher gain and bandwidth, increasing quiescent power consumption and introducing stability and design challenges at low supply voltages.

[0005] Existing publicly available technologies have proposed a hybrid active-passive second-order noise shaping structure with a first-stage passive integrator, a second-stage low-gain amplification, and positive feedback. The CDAC residual, first-stage integration result, and second-stage integration result are then introduced into a multi-input comparator. Therefore, simply using "hybrid active-passive integrator + multi-input comparator" as an improvement is insufficient to provide a sufficient distinction. On the other hand, to improve the noise shaping order and feedback accuracy, residual sampling, first-stage integration, and second-stage integration need to be strictly coordinated with the SAR successive approximation process. If the residual sampling time falls within the CDAC switching phase, bottom-plate switching glitches and the still-unstable top-plate voltage will be introduced into the integration state and fed back across cycles to subsequent comparison processes. Therefore, ensuring that the residual sampling corresponds only to the final residual after SAR conversion and updating the integration state through non-overlapping phases is a key timing issue in practical circuit implementation.

[0006] Furthermore, second-order noise shaping typically requires analog summation of the current CDAC residual with the first and second-order integral states. Adding an analog adder or closed-loop summing amplifier would increase input noise, power consumption, and feedback loop delay; using a multi-input comparator directly would require controlling the thermal noise, kickback noise, and weighting bias introduced by multiple input pairs. Therefore, the weighted summation of multiple differential components needs to be performed within the comparator itself, and the interference of the comparator regeneration process on the CDAC top plate and integrator-hold node needs to be reduced.

[0007] Noise shaping also introduces problems with digital output representation. The number of bits in the raw codeword of the internal SAR quantizer is not equivalent to the number of effective bits in the band after oversampling and noise shaping. If only the low-bit raw code is retained for spectrum analysis, it is difficult to directly generate a low-speed digital output that matches the final effective resolution. Therefore, a digital low-pass filter and decimation backend are needed to filter, accumulate, and downsample multiple oversampled codewords. At the same time, the NS-SAR ADC contains a highly sensitive CDAC top board, an integral-hold node, and large-swing SAR logic, clock, and digital output signals. Digital switching noise can easily enter the analog residual path through power supply, ground, and parasitic coupling. Therefore, it is also necessary to isolate the analog modules, digital modules, and sensitive nodes in the layout and power domain.

[0008] Therefore, the difficulty of existing technologies lies not in achieving residual integration alone or increasing the noise shaping order, but in establishing a consistent and coordinated relationship among the low-bit SAR quantization core, low-power integration loop, final residual sampling timing, multi-input comparison summation, noise shaping switch control, digital decimation output, and mixed signal layout. This is to avoid high-bit CDAC and high-gain closed-loop operational amplifiers while taking into account in-band noise suppression, loop delay, output word length expression, and digital noise isolation. Summary of the Invention

[0009] The technical problem to be solved by this invention is to address the shortcomings of the prior art by providing a second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter and its control method. This invention addresses the difficulty of performing low-power, high-precision second-order cross-cycle noise shaping on the final residual after conversion in existing SAR analog-to-digital converters using low-bit quantization cores without excessively increasing the CDAC scale, operational amplifier power consumption, and analog summing circuit complexity. It also considers residual feedback error compensation, low-delay weighted summation of multi-channel integration states, noise shaping mode switching, digital output word length matching, and analog / digital noise isolation, which leads to limited effective resolution within the target bandwidth.

[0010] The present invention adopts the following technical solution: A second-order hybrid residual feedback noise shaping SAR analog-to-digital converter includes a CDAC capacitor array, a SAR logic module, a residual sampling phase generation module, an active and passive hybrid second-order noise shaping integrator, and a multi-input Double-Tail dynamic comparator. The SAR logic module is connected to the CDAC capacitor array and is used to control the CDAC capacitor array to perform successive approximation conversion; The residual sampling phase generation module is used to generate the residual sampling phase, the first integral phase, and the second integral phase after the successive approximation transformation is completed. The active-passive hybrid second-order noise shaping integrator includes a first-stage passive switched-capacitor integrator and a second-stage active-passive hybrid integrator. The first-stage passive switched-capacitor integrator samples the conversion residual of the top plate of the CDAC capacitor array during the residual sampling phase and forms a first-stage integration output during the first integration phase. The second-stage active-passive hybrid integrator includes a low-gain OTA, a positive feedback branch, and a second-stage switched-capacitor integrator, and forms a second-stage integration output based on the first-stage integration output. The multi-input double-Tail dynamic comparator receives the residual of the CDAC capacitor array, the first-level integral output, and the second-level integral output, respectively, and performs a weighted summation. The comparison output of the multi-input double-Tail dynamic comparator is connected to the SAR logic module.

[0011] Furthermore, it also includes a DAC baseboard driver module. The DAC baseboard driver unit in the DAC baseboard driver module receives FS, BSVG, PI and NI control signals. Each baseboard of the CDAC capacitor array is controlled by the PI signal and the NI signal, switching between VREF, VSS and VCM. The PI signal and the NI signal are mutually exclusive.

[0012] Furthermore, the first-stage passive switched capacitor integrator performs SAR conversion residual sampling in the residual sampling phase, and forms a first-stage integral output INT1P and INT1N through charge sharing in the first integral phase. The integral coefficient of the first-stage passive switched capacitor integrator is determined by the ratio of the sampling capacitor to the integrating capacitor.

[0013] Furthermore, the low-gain OTA of the second-stage active-passive hybrid integrator receives the first-stage integrator output and provides forward gain, and the positive feedback branch forms the positive feedback coefficient.

[0014] Furthermore, the low-gain OTA adopts a resistive-load NMOS differential common-source structure, wherein the VIP and VIN input pairs receive the first-stage integration output, and the VBP and VBN input pairs receive the second-stage integration state; the second-stage active-passive hybrid integrator also includes a switched-capacitor common-mode feedback circuit, which samples the output common mode of VOP and VON through the PHIR and PHSED two-phase switched-capacitor networks and compares it with VCM to generate a common-mode feedback control voltage.

[0015] Furthermore, the multi-input Double-Tail dynamic comparator includes three sets of differential inputs, wherein VINP and VINN are connected to the residuals of the CDAC capacitor array, VINT1P and VINT1N are connected to the first-stage integration output, and VINT2P and VINT2N are connected to the second-stage integration output. The equivalent inputs of the multi-input Double-Tail dynamic comparator satisfy the following:

[0016] in, This represents the equivalent input of the multi-input Double-Tail dynamic comparator; and These represent a set of differential inputs corresponding to the CDAC residuals; and These represent a set of differential inputs corresponding to the first-level integral output; and These represent a set of differential inputs corresponding to the second-order integral output; , and The size, transconductance, and dynamic node load of the corresponding input pair are determined; the multi-input Double-Tail dynamic comparator adopts a double-tailed dynamic structure with the input stage isolated from the regenerative latch stage.

[0017] Furthermore, the residual sampling phase generation module responds to the RDY signal to generate the PHRES residual sampling phase, and generates PH1 and PH2 non-overlapping integral phases as the first integral phase and the second integral phase. The PHRES residual sampling phase is used to sample the conversion residual difference of the top plate of the CDAC capacitor array after the successive approximation conversion is completed, and the PH1 and PH2 non-overlapping integral phases are used to update the integration state.

[0018] Furthermore, the SAR analog-to-digital converter has an NS-ON / NS-OFF switchable verification structure; when noise shaping is off, the inputs VINT1P, VINT1N, VINT2P and VINT2N of the multi-input Double-Tail dynamic comparator are clamped to the VCM, and the residual sampling phase, integral phase and the low-gain OTA are disabled.

[0019] Furthermore, this also includes the Memory output register and CIC3 / sinc 3 The digital low-pass extraction backend includes a 10-bit CDAC capacitor array. The memory output register outputs a 10-bit raw bitstream, which is then processed by the CIC3 / sinc... 3 The digital low-pass extraction backend performs digital low-pass and extraction processing to form an 18-bit low-speed output code.

[0020] Furthermore, it also includes a DAC baseboard driver module, a bootstrap switch module, a common-mode feedback circuit, and a digital output buffer; the CDAC capacitor array, DAC baseboard driver module, bootstrap switch module, active / passive hybrid second-order noise shaping integrator, low-gain OTA, common-mode feedback circuit, and the comparator core of the multi-input Double-Tail dynamic comparator adopt analog power supply AVDD and analog ground AVSS power domains; the SAR logic module, memory output register, residual sampling phase generation module body, digital output buffer, and CIC3 / sinc 3 The digital low-pass decimation backend uses a digital power supply DVDD and a digital ground DVSS power domain; the transmission gate TG analog switch, which is used to select and connect the common-mode voltage VCM, the DAC positive terminal output DACP, the DAC negative terminal output DACN, the first-stage integration node INT1, and the second-stage integration node INT2, is powered by an analog power supply AVDD and an analog ground AVSS; and the ordinary clock logic gate is powered by a digital power supply DVDD and a digital ground DVSS.

[0021] Another technical solution of the present invention is a control method for a second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter, comprising the following steps: The differential input signal is sampled by a CDAC capacitor array, and the CDAC capacitor array is controlled by a SAR logic module to perform successive approximation conversion. After the successive approximation conversion is completed, the conversion residual difference of the top plate of the CDAC capacitor array is sampled by the residual sampling phase. The conversion residual is charged by a first-stage passive switched capacitor integrator to form a first-stage integral output; A second-stage integral output is formed from the first-stage integral output through a low-gain OTA, a positive feedback branch, and a second-stage switched capacitor integrator. The residuals of the CDAC capacitor array, the first-level integral output, and the second-level integral output are input into a multi-input Double-Tail dynamic comparator for weighted summation and comparison, and the comparison result is fed back to the SAR logic module.

[0022] Furthermore, when noise shaping is off, the differential inputs corresponding to the first-level integral output and the second-level integral output in the multi-input Double-Tail dynamic comparator are clamped to the VCM, and the residual sampling phase, integral phase, and low-gain OTA are disabled.

[0023] Compared with the prior art, the present invention has at least the following beneficial effects: A second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter (ADC) comprises a CDAC capacitor array, a SAR logic module, a residual sampling phase generation module, an active / passive hybrid second-order noise-shaping integrator, and a multi-input double-tail dynamic comparator forming the main conversion and feedback chain. After successive approximation conversion, the residual error at the top of the CDAC is sampled, and cross-cycle feedback is formed through a first-stage passive integrator and a second-stage active / passive hybrid integrator. The purpose of this design is to reduce quantization noise within the target bandwidth by utilizing a low-bit SAR quantization core in conjunction with second-order residual noise shaping, achieving higher in-band effective resolution without directly increasing the CDAC bit depth, while simultaneously alleviating the implementation burden on the high-bit CDAC in terms of total capacitance, matching, reference drive, and layout area.

[0024] Furthermore, a DAC baseboard driver module is configured, and the DACDrive unit receives FS, BSVG, PI, and NI control signals. Mutually exclusive PI and NI signals control the switching of each CDAC baseboard state between VREF, VSS, and VCM. This completes the controlled switching of the CDAC baseboard state during the SAR successive approximation process. The mutual exclusion of PI and NI avoids short circuits at the reference terminal and uncertainty in the baseboard state caused by simultaneous activation of two control signals, ensuring that the CDAC baseboard switching state is consistent with the SAR logic output, providing a definite circuit state for the formation and maintenance of subsequent conversion residuals.

[0025] Furthermore, a first-stage passive switched-capacitor integrator is employed. The SAR conversion residual is sampled at the PHRES phase, and at the PH1 phase, charge sharing forms the first-stage integration outputs INT1P and INT1N. The passive switched-capacitor method utilizes charge redistribution to complete the residual integration. Simultaneously, the integration coefficient is determined by the ratio of the sampling capacitor to the integrating capacitor, ensuring that the current cycle's conversion residual is preserved as a first-stage integration state at a defined ratio. This provides the first-order state quantity for subsequent second-order cross-cycle noise shaping.

[0026] Furthermore, a low-gain over-the-air (OTA) circuit and a positive feedback branch are incorporated into the second-stage active-passive hybrid integrator. The low-gain OTA circuit provides forward gain to the first-stage integrator output, while the positive feedback branch introduces positive feedback coefficients into the second-stage integrator state. This compensates for the finite gain and phase error of the purely passive integrator, making the second-stage integrator poles closer to the desired discrete-time integration characteristics. This avoids the static power consumption and stability design pressures associated with traditional high-gain closed-loop operational amplifiers, forming a second-stage integrator state suitable for second-order residual feedback.

[0027] Furthermore, the low-gain OTA employs a resistive-load NMOS differential common-source structure. The VIP / VIN input pair receives the first-stage integrator output, and the VBP / VBN input pair receives the second-stage feedback state. A switched-capacitor CMFB circuit samples and adjusts the VOP / VON output common-mode. This establishes a corresponding gain and positive feedback relationship between the forward integral and the second-stage feedback state in the low-gain OTA. Common-mode feedback maintains the output common-mode level, ensuring a stable operating level for the second-stage active-passive hybrid integrator, providing the necessary conditions for maintaining the second-stage integration state and subsequent comparator feedback.

[0028] Furthermore, a multi-input double-tail dynamic comparator with three sets of differential inputs is employed. The CDAC residual, first-stage integral output, and second-stage integral output are directly introduced into the comparator input stage, and corresponding weighting coefficients are formed based on the size, transconductance, and dynamic node load of each input. This allows the three analog state variables to be weighted and summed at the comparator input stage, eliminating the need for a separate analog adder or closed-loop summing amplifier, thus reducing noise, power consumption, and feedback path delay caused by additional analog circuitry. The double-tail structure further isolates the input stage and the regeneration latch stage, reducing the reverse disturbance of the regeneration process to the CDAC top plate and the integral-hold node.

[0029] Furthermore, the RDY signal triggers the PHRES residual sampling phase, generating non-overlapping PH1 and PH2 integral phases. PHRES is used to sample the CDAC top plate residual after successive approximation conversions, while PH1 and PH2 are used to update the integral state. This timing arrangement separates the residual sampling action from the SAR bit-by-bit switching process, ensuring that the noise shaping module obtains the final residual after conversion, rather than the transient residual generated during intermediate code switching. This gives the residual signal fed into the integration loop a definite sampling time point, facilitating the updating of the second-order integral state according to a predetermined phase sequence.

[0030] Furthermore, an NS-ON / NS-OFF switchable verification structure is configured. When noise shaping is off, the inputs VINT1P, VINT1N, VINT2P, and VINT2N of the comparator are clamped to the VCM, and the residual sampling phase, integral phase, and low-gain OTA enable are disabled. This ensures that the integral feedback quantity no longer participates in the comparator input in the NS-OFF state, allowing the same converter to switch between noise-shaping and normal SAR states. This facilitates comparison of conversion results before and after noise shaping under the same hardware and operating conditions, and prevents the integral branch from continuing to operate in the off state.

[0031] Furthermore, configure the Memory output register and CIC3 / sinc. 3 The digital low-pass decimation backend processes the 10-bit SAR raw code stream through digital low-pass, accumulation, and decimation to form an 18-bit low-speed output code. This setup aims to match the digital output format with the effective resolution within the target bandwidth after noise shaping. The 18-bit output is obtained by filtering and accumulating multiple oversampled codes, utilizing the spectral characteristics where out-of-band quantization noise is boosted and in-band noise is suppressed to form a corresponding low-speed, high-word-length digital output.

[0032] Furthermore, power domains are partitioned for analog and digital modules. The CDAC, DAC baseboard driver module, Bootstrap, active / passive hybrid integrator, low-gain OTA, CMFB, and comparator core utilize AVDD / AVSS, while the SAR logic, memory, residual sampling phase generation module body, DOUT buffer, and CIC3 digital backend utilize DVDD / DVSS. The TG switch directly connected to sensitive analog nodes uses an analog power domain. Power and layout isolation is achieved between the large-swing digital clock, SAR code stream, and switching disturbances generated by output flipping and the CDAC top board, comparator inputs, and INT1 / INT2 holding nodes, reducing digital noise coupling to the residual feedback link along power supply and interconnect paths.

[0033] A control method for a second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter (ADC) is disclosed. The control process sequentially executes differential input sampling, SAR successive approximation conversion, final residual sampling, first-level passive integration, second-level active-passive hybrid integration, and three-way weighted comparison, returning the comparison result to the SAR logic module. This method sequentially connects single-cycle SAR conversion with cross-cycle residual feedback, transforming the quantization residual of the current cycle into the integral feedback state of subsequent cycles. Thus, while maintaining the SAR successive approximation conversion method, a second-order noise shaping process is introduced, enabling the coordinated operation of the low-order quantization core and the cross-cycle noise shaping loop.

[0034] Furthermore, when noise shaping is off, the differential inputs corresponding to the first-stage and second-stage integral outputs of the multi-input Double-Tail dynamic comparator are clamped to the VCM, while the residual sampling phase, integral phase, and low-gain OTA are simultaneously disabled. The noise shaping feedback path is cut off from both the comparator input and the integral loop, ensuring that the comparison process in the off state is primarily determined by the CDAC residual, and preventing further updates from the residual sampling and integral branch, thus providing a clear control state for switching between NS-ON and NS-OFF operating modes.

[0035] In summary, this invention achieves in-band noise shaping based on the low-position SAR quantization core through residual sampling, second-order active-passive hybrid integration, and multi-input dynamic comparison feedback. Combined with non-overlapping timing, digital decimation, and analog-digital power domain isolation, it balances effective resolution, power consumption, loop delay, and layout complexity.

[0036] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the following description of the relative embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 This is a block diagram of the second-order noise-shaping SAR ADC system of the present invention; Figure 2 This is a schematic diagram illustrating the timing relationship between CLK, FS, and noise-shaping phase in this invention; Figure 3 This is a schematic diagram of the CDAC baseboard driver and 10-bit output register structure of the present invention; Figure 4 This is a schematic diagram of the principle of the second-order active-passive hybrid noise shaping integrator of the present invention; Figure 5 This is a schematic diagram of the low-gain OTA and switched-capacitor CMFB circuit of the present invention; Figure 6 This is a schematic diagram of the multi-input Double-Tail dynamic comparator structure of the present invention; Figure 7 This is a schematic diagram of the 10-bit raw code to 18-bit digital extraction output interface of the present invention; Figure 8 This is a schematic diagram of the layout planning and module partitioning of the NS-SAR ADC of the present invention; Figure 9 This is a schematic diagram of the pre-FFT spectrum and performance verification results of an embodiment of the present invention. Detailed Implementation

[0039] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "one side," "one end," and "one side," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0041] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0042] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0043] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0044] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0045] The accompanying drawings illustrate various structural schematic diagrams according to embodiments disclosed in this invention. These drawings are not to scale, and some details have been enlarged for clarity, and some details may have been omitted. The shapes of the various regions and layers shown in the drawings, as well as their relative sizes and positional relationships, are merely exemplary and may deviate from reality due to manufacturing tolerances or technical limitations. Furthermore, those skilled in the art can design regions / layers with different shapes, sizes, and relative positions as needed.

[0046] This invention provides a second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter and its control method. Using a 10-bit SAR quantizer as the core, after SAR conversion, RDY triggers PHRES to sample the final conversion residual of the CDAC top board, and the second-order integral state is updated sequentially using the non-overlapping phases of PH1 and PH2. The first stage uses passive switched-capacitor integration to reduce static power consumption. The second stage uses low-gain OTA, a positive feedback branch, second-stage switched-capacitor integration, and switched-capacitor common-mode feedback to compensate for the gain and phase error of the passive integrator. The CDAC residual, INT1, and INT2 are directly connected to a multi-input double-tail dynamic comparator, and weighted summation is performed at the comparator input stage to eliminate the need for an additional analog adder. NS_EN clamps the integral input to VCM and disables residual sampling, integral phase, and OTA when noise shaping is off, achieving switching between normal SAR / noise shaping operating states within the same circuit. Simultaneously, the high-speed 10-bit raw data is fed into the CIC3 / sinc3 backend. The code is filtered and decimated to form an 18-bit low-speed output. The layout is divided into AVDD / AVSS and DVDD / DVSS, and the sensitive traces of DACP / DACN and INT1 / INT2 are shortened. In this way, the effective resolution within the target bandwidth is improved without directly expanding the high-bit CDAC and without introducing a high-gain closed-loop summing op-amp, and the impact of residual sampling glitches, loop delay and digital switching noise on the analog residual path is reduced.

[0047] Please see Figure 1 This invention discloses a second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter, comprising a differential input sampling module, a CDAC capacitor array, a DAC backplane driver module, a SAR logic module, a memory output register, a multi-input Double-Tail dynamic comparator, an active-passive hybrid second-order noise-shaping integrator, a Dynamic_NolapClocks residual sampling phase generation module, and an optional CIC3 digital decimation backend. The CDAC capacitor array, SAR logic module, active-passive hybrid second-order noise-shaping integrator, and multi-input Double-Tail dynamic comparator constitute the main signal chain for SAR conversion and residual noise shaping.

[0048] The input signal is differentially sampled by the positive and negative CDAC terminals. The SAR logic module controls the CDAC capacitor array to perform successive approximation conversion. The comparison output of the multi-input Double-Tail dynamic comparator is fed back to the SAR logic module to determine the switching state of the CDAC capacitor array during successive approximation. After the successive approximation conversion of the current sampling period is completed, the top plate of the CDAC retains the conversion residual. The residual sampling phase generation module generates the PHRES residual sampling phase, samples the conversion residual, and forms a first-level integration state and a second-level integration state through an active-passive hybrid second-order noise shaping integrator. The first-level integration output and the second-level integration output are input together with the current CDAC residual in subsequent periods to the multi-input Double-Tail dynamic comparator. Weighted summation and comparison are performed at the comparator input stage, thus forming a residual feedback path across sampling periods.

[0049] Please see Figure 2 In this embodiment, the ADC sampling rate is 320kS / s, the signal bandwidth is 10kHz, the OSR is 16, the SAR master clock is about 4.8MHz, and each sampling period contains 15 CLK periods. Among them, the FS high level is used for input sampling for about 3 CLK periods, and the SAR conversion, residual preservation and noise shaping phase generation are completed during the FS low level period.

[0050] After receiving the RDY signal, the residual sampling phase generation module generates the PHRES residual sampling phase and produces non-overlapping PH1 and PH2 integral phases. PH1 corresponds to the first integral phase, and PH2 corresponds to the second integral phase. PHRES is located after the successive approximation conversion is completed, ensuring the final conversion residual of the first-stage passive switched-capacitor integrator at the top of the PHRES phase sampling CDAC. Subsequently, PH1 and PH2 control the updating of the first-stage and second-stage integral states. Thus, the residual sampling action is time-separated from the intermediate CDAC switching process in the SAR successive approximation process.

[0051] Please see Figure 3 The CDAC baseboard driver and 10-bit output register structure specifically include: The CDAC employs a 10-bit binary weighted capacitor array, containing equally weighted capacitors of 512C, 256C, 128C, 64C, 32C, 16C, 8C, 4C, 2C, and 1C, and includes dummy capacitors to balance the total capacitance and improve edge matching. During the sampling phase, the CDAC baseboard connects to the input signal via a bootstrap sampling path; during the conversion phase, each baseboard switches between VREF, VSS, and VCM according to control signals.

[0052] The DAC baseboard driver module includes a DACDrive unit, which receives FS, BSVG, PI, and NI control signals, where the PI and NI signals are mutually exclusive. The PI and NI signals are used to control the corresponding CDAC baseboard switching between VREF, VSS, and VCM. The P / N control lines are generated by SAR logic, where the P terminal can be used as the final latch code for memory, and the N terminal is used for CDAC differential baseboard driving. The single-bit DACDrive unit receives control signals such as FS, BSVG, PI, and NI. PI and NI must be mutually exclusive; they cannot be active simultaneously to avoid short circuits at the reference terminal and uncertainties in the backplane state. The P / N control lines are generated by SAR logic, where the P terminal can also serve as the final memory latch code, and the N terminal is mainly used for CDAC differential backplane driving and is not necessarily used as a user output.

[0053] Please see Figure 4 The active-passive hybrid second-order noise shaping integrator includes a first-stage passive switched capacitor integrator and a second-stage active-passive hybrid integrator.

[0054] The first stage employs a passive switched-capacitor integrator, sampling the SAR conversion residual at the PHRES phase and forming a first-stage integration output INT1P / INT1N through charge sharing at the PH1 phase; the first-stage passive integrator is represented as:

[0055] in, The ratio of the sampling capacitor to the integrating capacitor is determined. For complex variables in the Z-transform domain, The unit delay is one sampling period.

[0056] The second stage adopts a hybrid active and passive structure, including a low-gain OTA, a positive feedback branch, and a second-stage switched-capacitor integrator.

[0057] Low-gain OTA provides forward gain to the first-level integral output. The positive feedback branch introduces a second-order integral state and forms Coefficients that make the equivalent transfer function:

[0058] in, This is the equivalent transfer function of the second-stage active-passive hybrid integrator. This is the integral coefficient of the second-stage switched capacitor integrator.

[0059] when When the value approaches 1, the error term in the denominator decreases, and the poles of the integrator are closer to the ideal discrete-time integrator, thereby improving the second-order noise shaping effect.

[0060] Please see Figure 5 The low-gain OTA employs a resistive-loaded NMOS differential common-source structure. The VIP and VIN input pairs receive the first-stage integral output and provide forward gain. The VBP and VBN input pairs receive the secondary feedback state and form a positive feedback coefficient. Since the low-gain OTA processes the integral state corresponding to the SAR conversion residual, it adopts a low-gain structure instead of a traditional high-gain closed-loop operational amplifier.

[0061] The second-stage active-passive hybrid integrator also includes a switched-capacitor common-mode feedback circuit, CMFB_Circuit. CMFB_Circuit samples the output common-mode values ​​of VOP and VON through the two-phase switched-capacitor networks PHIR and PHSED, compares them with VCM to generate a CMFB control voltage, and adjusts the pull-down branch current using this control voltage. Keep it near VCM.

[0062] Please see Figure 6 The multi-input double-tail dynamic comparator specifically includes three sets of differential inputs: VINP / VINN connected to the CDAC residual, VINT1P / VINT1N connected to the first-level integral output, and VINT2P / VINT2N connected to the second-level integral output.

[0063] The equivalent input of the comparator is represented as:

[0064] in, This represents the equivalent input of the multi-input Double-Tail dynamic comparator; and These represent a set of differential inputs corresponding to the CDAC residuals; and These represent a set of differential inputs corresponding to the first-level integral output; and These represent a set of differential inputs corresponding to the second-order integral output; , and It is determined by the size, transconductance, and dynamic node load of the corresponding input pair.

[0065] The multi-input double-tail dynamic comparator directly introduces the current CDAC residual, the first-order integration state, and the second-order integration state into the comparator input stage, where a weighted summation is performed. The comparator employs a double-tailed dynamic structure, with the input stage isolated from the regenerative latch stage. The comparison result is output to the SAR logic module, enabling the integration state to participate in the successive approximation judgment of subsequent cycles, forming a second-order residual feedback loop.

[0066] To support switching between noise shaping and normal SAR modes, this embodiment features an NS-ON / NS-OFF switchable verification structure.

[0067] In NS-ON state, the PHRES residual sampling phase, PH1 / PH2 integral phase, and low-gain OTA operate according to the noise shaping timing sequence, and VINT1P / VINT1N and VINT2P / VINT2N input the first-level integral state and the second-level integral state to the multi-input Double-Tail dynamic comparator, respectively.

[0068] In the NS-OFF state, i.e. when noise shaping is off, the VINT1P, VINT1N, VINT2P and VINT2N inputs of the multi-input Double-Tail dynamic comparator are clamped to the VCM. At the same time, the residual sampling phase, integral phase and low-gain OTA are disabled, so that the integral feedback quantity no longer participates in the comparator input.

[0069] Please see Figure 7 The CDAC uses a 10-bit quantization core, and the memory output register outputs 10-bit raw code. This raw code stream is input to CIC3 / sinc. 3 The digital low-pass extraction backend, after digital low-pass and extraction processing, forms an 18-bit low-speed output code.

[0070] Under one specific setup, the memory outputs 10-bit raw code at a sampling rate of 320 kS / s. This raw bitstream is then fed into CIC3 / sinc. 3 The digital low-pass decimation back-end can be set to a decimation factor of 8 or 16. When DECIM=8, the output rate is 40kS / s, and the 10kHz bandwidth has a sufficient transition band; when DECIM=16, the output rate is 20kS / s, which corresponds to OSR=16, but 10kHz is close to the Nyquist boundary, so attention should be paid to the design of the filter transition band.

[0071] The CIC3 backend uses a Verilog-A behavioral model for pre-simulation verification, and can also be synthesized using RTL and standard cells in subsequent chip implementations. Its internal word length is designed to be 22~24 bits, and the final output is rounded or truncated to 18 bits, while providing DOUT18<18:1>, VALID18 and VCODE monitoring signals.

[0072] Please see Figure 8 The layout planning and module partitioning of the NS-SAR ADC in this invention are specifically as follows: The positive and negative DAC capacitor arrays are arranged symmetrically. The DACP / DACN on the top of the CDAC should be short and symmetrical, and kept away from large-swing digital lines such as CLK, P / N, and DOUT. Low-gain OTA, CMFB, and Active-Passive integrators should be placed close to the comparator and integrating capacitor array to shorten the INT1 / INT2 holding node traces. SAR logic, memory, and CIC3 digital back-ends can be placed in the digital area, and digital switching noise coupling can be reduced through power ground partitioning.

[0073] The preferred power domain partitioning is as follows: CDAC, DACDrive, Bootstrap, ActivePassive_SCint, LowGainAMP, CMFB, and the comparator core use analog power supply AVDD and analog ground AVSS; SARLogic, Memory, Dynamic_NolapClocks main body, digital output buffer (DOUT buffer), and CIC3 digital back-end use digital power supply DVDD and digital ground DVSS; analog switches for the transmission gates (TG) directly connected to VCM, DAC positive output DACP and DAC negative output DACN, and INT1 / INT2 use AVDD / AVSS, while ordinary clock logic gates use digital power supply DVDD and digital ground DVSS. AVSS and DVSS are connected at a controlled single point on the chip periphery or test platform.

[0074] When the second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter of the present invention is working, it first samples the differential input signal through the CDAC capacitor array, and the SAR logic module controls the CDAC capacitor array to perform successive approximation conversion. After the successive approximation conversion is completed, the residual sampling phase generation module uses the PHRES residual sampling phase to sample the conversion residual difference retained by the CDAC top plate; The first-stage passive switched capacitor integrator forms a first-stage integral output through charge sharing in phase PH1. The second-stage active-passive hybrid integrator utilizes a low-gain OTA, a positive feedback branch, and a second-stage switched capacitor integrator to form a second-stage integral output. Subsequently, the current residual of CDAC, the first-level integral output, and the second-level integral output are fed into the three differential inputs of the multi-input Double-Tail dynamic comparator. Weighted summation and comparison are performed at the comparator input stage, and the comparison results are fed back to the SAR logic module for subsequent successive approximation conversion and cross-cycle residual feedback.

[0075] When noise shaping is off, the differential inputs of the corresponding first-stage and second-stage integral outputs in the comparator are clamped to the VCM, and the residual sampling phase, integral phase, and low-gain OTA are disabled.

[0076] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0077] Please see Figure 9 Based on the aforementioned embodiments employing a 10-bit SAR quantizer, a second-order active-passive hybrid residual feedback structure, and oversampling processing, frequency domain analysis is performed on the analog-to-digital converter output code. Figure 9 As can be seen from the FFT spectrum, the fundamental spectrum corresponding to the input signal is significantly higher than the noise floor. The noise remains at a low level in the low-frequency target signal band. Furthermore, as the frequency increases, the noise spectrum energy generally shows an upward trend, indicating that the quantization noise after residual feedback is transferred to outside the target bandwidth, which corresponds to the spectral characteristics of second-order noise shaping.

[0078] Figure 9 In the simulation results shown, the effective number of bits (ENOB) is 14.965736 bits, the signal-to-noise ratio (SINAD) is 91.863619 dB, the signal-to-noise ratio (SNR) is 91.863619 dB, the spurious-free dynamic range (SFDR) is 96.944179 dBc, and the simulation results interface shows a THD of 0%. Here, ENOB represents the equivalent resolution of the analog-to-digital converter after considering noise and distortion; SINAD represents the ratio of signal power to the sum of noise and distortion power; SNR represents the ratio of signal power to noise power; and SFDR represents the amplitude difference between the fundamental signal and the largest spurious component in the spectrum.

[0079] The above simulation results show that, under the condition of using a 10-bit SAR quantization core, by sampling the residual of the CDAC top plate after conversion, and feeding it back to the multi-input dynamic comparator across cycles after a first-stage passive integration and a second-stage active-passive hybrid integration, it is possible to form in-band noise shaping characteristics corresponding to the design target, so that this embodiment can obtain an effective resolution of close to 15 bits.

[0080] After the SAR successive approximation conversion is completed in the current sampling period, the PHRES residual sampling phase samples the conversion residual of the CDAC top plate. The first-stage passive switched capacitor integrator forms the first-stage integration state, and the second-stage active-passive hybrid integrator further forms the second-stage integration state. In subsequent periods, the current CDAC residual, the first-stage integration output, and the second-stage integration output are input to the multi-input Double-Tail dynamic comparator and weighted summation is performed at the comparator input stage. Figure 9 The spectral distribution, in which noise is suppressed in the low-to-mid frequency region and gradually increases in the high-frequency region, is the manifestation of the above-mentioned second-order cross-period residual feedback process in the frequency domain.

[0081] Based on the above structure and control process, this embodiment uses a 10-bit SAR quantizer as the quantization core. After successive approximation conversions, the conversion residual retained by the CDAC top plate is sampled, and a second-order cross-cycle residual feedback is formed through a first-stage passive switched-capacitor integrator and a second-stage active-passive hybrid integrator. This causes the quantization noise to be shaped outside the target bandwidth via the noise transfer function. Therefore, the effective resolution within the target bandwidth can be improved without directly increasing the number of bits in the SAR quantizer, while reducing the requirements for capacitor array size, capacitor matching, reference voltage driving capability, and layout area when using a high-bit CDAC.

[0082] The second-stage active-passive hybrid integrator provides forward gain to the first-stage integrator output through a low-gain over-the-air (OTA) circuit and introduces the second-stage integrator state using a positive feedback branch. Compared to structures that only use passive integrators, this approach can compensate for the limited gain and phase error of passive integrators; at the same time, the low-gain OTA eliminates the need for traditional high-gain closed-loop operational amplifiers, balancing residual integration accuracy and circuit implementation complexity.

[0083] The multi-input double-tail dynamic comparator introduces the current CDAC residual, the first-stage integration output INT1, and the second-stage integration output INT2 into three sets of differential inputs, and performs weighted summation at the comparator input stage. This allows the three analog quantities to directly participate in the comparison process, eliminating the need for a separate analog adder before the comparator. This shortens the residual feedback path and reduces noise, power consumption, and loop delay caused by additional analog summing circuitry.

[0084] In terms of timing control, the residual sampling phase generation module uses RDY to trigger the PHRES residual sampling phase, and updates the first-level integration state and the second-level integration state through the non-overlapping PH1 and PH2 integration phases, so that PHRES is located after the SAR successive approximation conversion is completed, so that the noise shaping loop collects the final conversion residual of the CDAC top plate, rather than the intermediate switching state formed during the successive approximation process.

[0085] For digital output, this embodiment inputs the high-speed 10-bit raw bitstream output from the Memory into CIC3 / sinc. 3 The digital low-pass decimation backend forms an 18-bit low-speed output code through digital low-passing, accumulation, and decimation of multiple oversampled output codes. Therefore, this 18-bit output is not a bit extension of a single 10-bit output code, but corresponds to the effective resolution within the target bandwidth after noise shaping.

[0086] In terms of chip implementation, CDAC, DACDrive, Bootstrap, ActivePassive_SCint, LowGainAMP, CMFB, and the comparator core are placed in the AVDD / AVSS analog power domain, while SARLogic, Memory, Dynamic_NolapClocks main body, DOUT buffer, and CIC3 digital back-end are placed in the DVDD / DVSS digital power domain. The analog residual path and digital logic area are partitioned to reduce the coupling of switching noise generated by digital clock, SAR code stream, and output code flipping to the CDAC top board, comparator input, and INT1 / INT2 holding nodes.

[0087] This invention is not limited to the specific embodiments described above. In other embodiments, in addition to using a 10-bit quantizer, the SAR quantizer can also use an 8-bit, 9-bit, 11-bit, or 12-bit quantizer, and the noise shaping order and digital filter bit width can be adjusted accordingly based on the target signal bandwidth, oversampling rate (OSR), power consumption, and capacitor matching requirements.

[0088] The active-passive hybrid second-order noise shaping integrator can also be replaced by a pure passive first-order integrator structure, a pure passive second-order integrator structure, a low-gain active first-order integrator structure, a fully active second-order integrator structure, or a gain-enhanced active-passive hybrid structure. As long as it can perform cross-cycle filtering feedback on the SAR conversion residual, it can form the corresponding residual noise shaping path.

[0089] For the analog summation of residuals and integral states, this embodiment uses a multi-input Double-Tail dynamic comparator to directly receive the CDAC residuals, INT1, and INT2; in other embodiments, the combination of the above analog quantities can also be achieved by using a pre-capacitor summation, a dynamic preamplifier summation, or an independent analog summation node.

[0090] The CDAC baseboard driving method is not limited to VREF, VCM and VSS three-level switching. In other implementations, monotonic switching, split capacitor array, Vcm-based switching or other DAC switching methods can also be used.

[0091] For the digital back-end, this embodiment sets CIC3 / sinc on-chip. 3 The digital low-pass extraction backend is equipped with DOUT18 and VCODE monitoring interfaces. In other implementations, 10-bit raw code can be directly output and the digital low-pass and extraction processing can be completed by an off-chip digital processor. When the digital filtering function is integrated into the chip, it can be described using RTL and implemented through standard cell synthesis.

[0092] In terms of power domain configuration, in addition to the above-mentioned AVDD / AVSS and DVDD / DVSS power domains, DRVDD / DRVSS can also be set separately as the CDAC baseboard driver power domain to provide independent power supply for the DAC baseboard driver section.

[0093] This embodiment can be implemented using SMIC 0.18 μm CMOS technology, or it can be migrated to other CMOS process nodes. When the process node changes, the CDAC capacitor parameters, capacitor ratio, comparator device size, and clock timing can be adjusted accordingly based on the device threshold, capacitance density, MIM capacitor matching performance, and power supply voltage of the process used.

[0094] In summary, this invention provides a second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter and its control method. Based on a low-end SAR quantization core, it samples the CDAC top plate residual after successive approximation conversions. A second-order cross-cycle residual feedback is formed through a first-stage passive integrator and a second-stage active-passive hybrid integrator, shaping the quantization noise out of bandwidth. The second stage uses a low-gain OTA and a positive feedback branch to compensate for the finite gain and phase error of the passive integrator, avoiding reliance on high-gain operational amplifiers. The CDAC residual, the first-stage integrator output, and the second-stage integrator output are connected to a multi-input double-tail dynamic comparator for weighted summation, eliminating the need for an analog adder. RDY triggers PHRES, updating the integration state using the non-overlapping phases of PH1 and PH2, sampling only the residual after SAR conversion. CIC3 / sinc 3 The digital back-end extracts the 10-bit raw bitstream into an 18-bit output; the analog and digital power domains are partitioned to reduce noise coupling from digital switches, while taking into account effective resolution, power consumption, delay, and layout implementation.

[0095] The above content is only for illustrating the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solution based on the technical concept proposed in this invention shall fall within the scope of protection of the claims of this invention.

Claims

1. A second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter, characterized in that, It includes a CDAC capacitor array, a SAR logic module, a residual sampling phase generation module, an active-passive hybrid second-order noise shaping integrator, and a multi-input Double-Tail dynamic comparator; The SAR logic module is connected to the CDAC capacitor array and is used to control the CDAC capacitor array to perform successive approximation conversion; The residual sampling phase generation module is used to generate the residual sampling phase, the first integral phase, and the second integral phase after the successive approximation transformation is completed. The active-passive hybrid second-order noise shaping integrator includes a first-stage passive switched-capacitor integrator and a second-stage active-passive hybrid integrator. The first-stage passive switched-capacitor integrator samples the conversion residual of the top plate of the CDAC capacitor array during the residual sampling phase and forms a first-stage integration output during the first integration phase. The second-stage active-passive hybrid integrator includes a low-gain OTA, a positive feedback branch, and a second-stage switched-capacitor integrator, and forms a second-stage integration output based on the first-stage integration output. The multi-input double-Tail dynamic comparator receives the residual of the CDAC capacitor array, the first-level integral output, and the second-level integral output, respectively, and performs a weighted summation. The comparison output of the multi-input double-Tail dynamic comparator is connected to the SAR logic module.

2. The second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter according to claim 1, characterized in that, It also includes a DAC baseboard driver module, in which the DAC baseboard driver unit receives FS, BSVG, PI and NI control signals. Each baseboard of the CDAC capacitor array is controlled by the PI signal and the NI signal, switching between VREF, VSS and VCM, wherein the PI signal and the NI signal are mutually exclusive.

3. The second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter according to claim 1, characterized in that, The first-stage passive switched capacitor integrator performs SAR conversion residual sampling in the residual sampling phase, and forms first-stage integration outputs INT1P and INT1N through charge sharing in the first integration phase. The integration coefficient of the first-stage passive switched capacitor integrator is determined by the ratio of the sampling capacitor to the integration capacitor.

4. The second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter according to claim 1, characterized in that, The low-gain OTA of the second-stage active-passive hybrid integrator receives the first-stage integrator output and provides forward gain, and the positive feedback branch forms the positive feedback coefficient.

5. The second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter according to claim 4, characterized in that, The low-gain OTA adopts a resistive-loaded NMOS differential common-source structure, wherein the VIP and VIN input pairs receive the first-stage integration output, and the VBP and VBN input pairs receive the second-stage integration state; the second-stage active-passive hybrid integrator also includes a switched-capacitor common-mode feedback circuit, which samples the output common mode of VOP and VON through the PHIR and PHSED two-phase switched-capacitor networks and compares it with VCM to generate a common-mode feedback control voltage.

6. The second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter according to claim 1, characterized in that, The multi-input Double-Tail dynamic comparator includes three sets of differential inputs, where VINP and VINN are connected to the residuals of the CDAC capacitor array, VINT1P and VINT1N are connected to the first-stage integration output, and VINT2P and VINT2N are connected to the second-stage integration output. The equivalent inputs of the multi-input Double-Tail dynamic comparator satisfy the following: in, This represents the equivalent input of the multi-input Double-Tail dynamic comparator; and These represent a set of differential inputs corresponding to the CDAC residuals; and These represent a set of differential inputs corresponding to the first-level integral output; and These represent a set of differential inputs corresponding to the second-order integral output; , and The size, transconductance, and dynamic node load of the corresponding input pair are determined; the multi-input Double-Tail dynamic comparator adopts a double-tailed dynamic structure with the input stage isolated from the regenerative latch stage.

7. The second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter according to claim 1, characterized in that, The residual sampling phase generation module responds to the RDY signal to generate the PHRES residual sampling phase, and generates PH1 and PH2 non-overlapping integral phases as the first integral phase and the second integral phase. The PHRES residual sampling phase is used to sample the conversion residual difference of the top plate of the CDAC capacitor array after successive approximation conversion is completed, and the PH1 and PH2 non-overlapping integral phases are used to update the integration state.

8. The second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter according to claim 1, characterized in that, The SAR analog-to-digital converter has an NS-ON / NS-OFF switchable verification structure; when noise shaping is off, the inputs VINT1P, VINT1N, VINT2P and VINT2N of the multi-input Double-Tail dynamic comparator are clamped to the VCM, and the residual sampling phase, integral phase and the low-gain OTA are disabled.

9. The second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter according to claim 1, characterized in that, Memory output register and CIC3 / sinc 3 Digital low pass decimation backend, the CDAC capacitor array is a 10 bit CDAC capacitor array, the Memory output register outputs a 10 bit raw code stream, and the 10 bit raw code stream is subjected to digital low pass and decimation processing through the CIC3 / sinc 3 Digital low pass decimation backend, the CDAC capacitor array is a 10 bit CDAC capacitor array, the Memory output register outputs a 10 bit raw code stream, and the 10 bit raw code stream is subjected to digital low pass and decimation processing through the CIC3 / sinc 10. The second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter according to claim 9, characterized in that, Also includes DAC backplane drive module, bootstrap switch module, common mode feedback circuit and digital output buffer; the CDAC capacitor array, DAC backplane drive module, bootstrap switch module, active and passive hybrid second-order noise shaping integrator, low-gain OTA, common mode feedback circuit and the comparator core of the multi-input Double-Tail dynamic comparator adopt analog power supply AVDD and analog ground AVSS power supply domain, the SAR logic module, Memory output register, residual sampling phase generation module main body, digital output buffer and CIC3 / sinc 3 Digital low-pass decimation backend adopts digital power supply DVDD and digital ground DVSS power supply domain; the transmission gate TG analog switch for gating connection of common mode voltage VCM, DAC positive end output DACP, DAC negative end output DACN, first stage integration node INT1 and second stage integration node INT2 adopts analog power supply AVDD and analog ground AVSS for power supply, and ordinary clock logic gate adopts digital power supply DVDD and digital ground DVSS for power supply.

11. A control method for a second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter, characterized in that, Includes the following steps: The differential input signal is sampled by a CDAC capacitor array, and the CDAC capacitor array is controlled by a SAR logic module to perform successive approximation conversion. After the successive approximation conversion is completed, the conversion residual difference of the top plate of the CDAC capacitor array is sampled by the residual sampling phase. The conversion residual is charged by a first-stage passive switched capacitor integrator to form a first-stage integral output; A second-stage integral output is formed from the first-stage integral output through a low-gain OTA, a positive feedback branch, and a second-stage switched capacitor integrator. The residuals of the CDAC capacitor array, the first-level integral output, and the second-level integral output are input into a multi-input Double-Tail dynamic comparator for weighted summation and comparison, and the comparison result is fed back to the SAR logic module.

12. The control method for a second-order hybrid residual feedback noise-shaping SAR analog-to-digital converter according to claim 11, characterized in that, When noise shaping is off, the differential inputs corresponding to the first-level integral output and the second-level integral output in the multi-input Double-Tail dynamic comparator are clamped to the VCM, and the residual sampling phase, integral phase and low-gain OTA are disabled.