Time-interleaved analog-to-digital converter sub-channel mismatch calibration method, time-interleaved analog-to-digital converter and SERDES system

By performing offset, gain and timing mismatch calibration in the subchannel of the time interleaved analog-to-digital converter, circuit design is simplified, power consumption is reduced, and calibration efficiency is improved, solving the complex circuit and high power consumption in the prior art The problem.

CN118868940BActive Publication Date: 2024-12-06CHENGDU NOVACORE MICROELECTRONICS TECHNOLOGY CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202411368256.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-29
Publication Date
2024-12-06
Estimated Expiration
2044-09-29

AI Technical Summary

Technical Problem

In the prior art, the circuit design of time interleaved analog-to-digital converters is complex and has high power consumption, so it is impossible to effectively avoid mismatch between sub-channels, resulting in spectrum spurious components, limiting equipment performance.

Method used

By accessing injection signals into each subchannel of the time interleaving analog-to-digital converter, offset, gain and timing mismatch calibration is performed separately, the mismatch value is calculated using the signal difference between the reference subchannel and the non-reference subchannel, and the calibration is determined iteratively group, avoiding the introduction of additional reference subchannels, simplifying circuit design and reducing power consumption.

Benefits of technology

The circuit design is simplified and power consumption is reduced, while the timing calibration rate is improved and the foreground calibration efficiency of time interleaved analog-to-digital converters is improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118868940B_ABST
    Figure CN118868940B_ABST
Patent Text Reader

Abstract

The present invention provides a time-interleaved analog-to-digital converter subchannel mismatch calibration method, a time-interleaved analog-to-digital converter and a SERDES system. The method comprises: after each subchannel of the time-interleaved analog-to-digital converter is respectively connected to a first injection signal, an offset mismatch value is determined according to the first signal output by each subchannel, thereby performing offset mismatch calibration on each subchannel; after each subchannel is respectively connected to a second injection signal, a gain mismatch value of each non-reference subchannel is determined according to the second signal output by the subchannel, thereby performing gain mismatch calibration on each non-reference subchannel; after each subchannel is respectively connected to a third injection signal, a first reference subchannel, a second reference subchannel and a subchannel to be time-series calibrated are iteratively determined, and, according to the third signal output by the above subchannel, a timing mismatch calibration is performed on the subchannel to be time-series calibrated. The present invention avoids setting an additional reference subchannel by determining a reference subchannel in each subchannel, and has a simple circuit design and low power consumption.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of mismatch calibration, and in particular to a time-interleaved analog-to-digital converter sub-channel mismatch calibration method, a time-interleaved analog-to-digital converter and a SERDES system. Background Art

[0002] Serializer and Deserializer (SERDES) is a common serial communication technology. SERDES system usually includes a transmitter for serializing parallel signals and a receiver for parallelizing serial signals, which is used to achieve high-speed signal transmission between electronic devices while greatly reducing the number of connecting wires. SERDES system is widely used in computer networks, communication systems, storage devices and graphic displays.

[0003] As the demand for the transmission rate of the SERDES system increases, the insertion loss at the Nyquist frequency due to electrical characteristics is also increasing. The traditional analog system architecture limits the application of advanced equalizers, so a high-speed SERDES system architecture is introduced. The high-speed SERDES system architecture usually adopts a digital reception solution based on a digital signal processor, that is, a multi-channel time interleaving solution is used to achieve a balance between low power consumption and high sampling rate. Specifically, the time-interleaved analog-to-digital converter samples at one-Mth of the system sampling rate through M sub-channels, and the sampling results are combined in a time-interleaved manner, so as to achieve the effect of a single-channel analog-to-digital converter sampling at the system sampling rate. However, in the actual circuit design of the time-interleaved analog-to-digital converter, due to the different line connections of each sub-channel, there will be a mismatch effect between the sub-channels. This mismatch effect may cause spectral spurious components, which greatly limits the performance indicators of the time-interleaved analog-to-digital converter. Since the mismatch effect cannot be completely avoided by optimizing the design, it is necessary to calibrate the mismatch effect between channels before the time-interleaved analog-to-digital converter actually works. Among them, the mismatch effect includes offset mismatch, gain mismatch and timing mismatch.

[0004] In the prior art, at least one additional reference sub-channel whose sampling rate is mutually prime with each sub-channel is usually introduced during the mismatch estimation of the timing mismatch, which will cause complex circuit design and high power consumption. Summary of the invention

[0005] The purpose of the present invention is to provide a time-interleaved analog-to-digital converter sub-channel mismatch calibration method, a time-interleaved analog-to-digital converter and a SERDES system to address the deficiencies in the above-mentioned prior art, so as to solve the problems of complex circuit design and high power consumption in the prior art.

[0006] To achieve the above purpose, the technical solution adopted by the embodiment of the present invention is as follows:

[0007] In a first aspect, an embodiment of the present invention provides a time-interleaved analog-to-digital converter sub-channel mismatch calibration method, the method comprising:

[0008] After each sub-channel of the time-interleaved analog-to-digital converter is respectively connected to a first injection signal, an offset mismatch value of each sub-channel is determined according to the first signal output by each sub-channel, and offset mismatch calibration is performed on each sub-channel according to the offset mismatch value of each sub-channel, wherein the first injection signal is a 0 volt DC signal;

[0009] After each sub-channel completes the offset mismatch calibration and is respectively connected to the second injection signal, the gain mismatch value of each non-reference sub-channel is determined according to the second signal output by the reference sub-channel and the non-reference sub-channel, and the gain mismatch calibration is performed on each non-reference sub-channel according to the gain mismatch value of each non-reference sub-channel, wherein the second injection signal is a DC signal;

[0010] After each non-reference sub-channel completes the gain offset mismatch calibration and each sub-channel is respectively connected to the third injection signal, the first reference sub-channel, the second reference sub-channel and the sub-channel to be timing calibrated are iteratively determined, and, according to the third signal output by the first reference sub-channel, the third signal output by the second reference sub-channel and the third signal output by the sub-channel to be timing calibrated, the timing mismatch calibration is performed on the sub-channel to be timing calibrated.

[0011] Optionally, determining the offset mismatch value of each sub-channel according to the first signal output by each sub-channel includes:

[0012] The expected value of each of the first signals is used as the offset mismatch value of each sub-channel.

[0013] Optionally, determining the gain mismatch value of each non-reference sub-channel according to the second signal output by the reference sub-channel and the non-reference sub-channel includes:

[0014] Determining a reference variance of a second signal output by the reference subchannel;

[0015] Determine a plurality of first variances of the second signal output by each non-reference subchannel;

[0016] The differences between the reference variance and each first variance are calculated respectively, and each difference is used as a gain mismatch value corresponding to each non-reference sub-channel.

[0017] Optionally, the iterative determination of the first reference sub-channel, the second reference sub-channel, and the sub-channel to be timing calibrated, and the timing mismatch calibration of the sub-channel to be timing calibrated according to the third signal output by the first reference sub-channel, the third signal output by the second reference sub-channel, and the third signal output by the sub-channel to be timing calibrated, includes:

[0018] In each iteration, a plurality of calibration groups are determined based on the binary method, each of the calibration groups respectively comprising: a first reference sub-channel, a second reference sub-channel and a sub-channel to be time-calibrated, and the first reference sub-channel and the second reference sub-channel have the same timing mismatch;

[0019] In each calibration group, the timing mismatch value of the sub-channel to be calibrated is determined according to the third signal output by the first reference sub-channel, the third signal output by the second reference sub-channel and the third signal output by the sub-channel to be calibrated, and the timing mismatch calibration is performed on the sub-channel to be calibrated according to the timing mismatch value of the sub-channel to be calibrated.

[0020] Optionally, the determining a plurality of calibration groups based on a dichotomy method comprises:

[0021] Each sub-channel to be time-series calibrated in the previous iteration is used as a reference calibration sub-channel;

[0022] According to each benchmark calibration subchannel, the first reference subchannel, the second reference subchannel and the subchannel to be time-series calibrated in the multiple calibration groups corresponding to the benchmark calibration subchannel are determined, wherein the first reference subchannel in each calibration group is the benchmark calibration subchannel.

[0023] Optionally, the determining, according to each reference calibration sub-channel, a first reference sub-channel, a second reference sub-channel, and a sub-channel to be time-series calibrated in a plurality of calibration groups corresponding to the reference calibration sub-channel comprises:

[0024] Using the reference calibration subchannel as the first reference subchannel in a first calibration group and the first reference subchannel in a second calibration group corresponding to the reference calibration subchannel;

[0025] Two subchannels satisfying preset conditions are used as the second reference subchannel in the first calibration group and the second reference subchannel in the second calibration group, respectively, wherein the preset conditions include: the channel interval between the subchannel and the reference calibration subchannel is half of the channel interval used in the previous iteration;

[0026] The subchannel located between the first reference subchannel and the second reference subchannel in the first calibration group is used as the subchannel to be timing calibrated in the first calibration group, and the subchannel located between the first reference subchannel and the second reference subchannel in the second calibration group is used as the subchannel to be timing calibrated in the second calibration group.

[0027] Optionally, determining a timing mismatch value of the sub-channel to be timing calibrated according to a third signal output by the first reference sub-channel, a third signal output by the second reference sub-channel, and a third signal output by the sub-channel to be timing calibrated includes:

[0028] Determining a first expected value according to the third signal of the sub-channel to be time-calibrated and the third signal of the first reference sub-channel;

[0029] Determining a second expected value according to the third signal of the sub-channel to be time-calibrated and the third signal of the second reference sub-channel;

[0030] According to the first expected value, the second expected value and the preset constraint condition, the timing mismatch value of the sub-channel to be timing calibrated is determined, the preset constraint condition indicates that the frequency of the third injection signal is equal to the ratio of the frequency of the reference clock signal to the preset value, and the lowest common divisor of the preset value and the number of sub-channels of the time-interleaved analog-to-digital converter is less than or equal to 2.

[0031] Optionally, performing timing mismatch calibration on the sub-channel to be timing calibrated according to the timing mismatch value of the sub-channel to be timing calibrated includes:

[0032] An alpha filter process is performed on the timing mismatch value of the sub-channel to be timing-calibrated to obtain a timing mismatch calibration value, and timing calibration is performed on the sub-channel to be timing-calibrated based on the timing mismatch calibration value.

[0033] The beneficial effect of the present invention is that: in the present embodiment, a reference sub-channel or a first reference sub-channel and a second reference sub-channel are determined in each sub-channel during the gain mismatch calibration and the timing mismatch calibration process, so that there is no need for channels other than the original sub-channels of the time-interleaved analog-to-digital converter to be used as reference sub-channels, so that the circuit design is simple, the occupied area is small, and the power consumption is low. In addition, on the basis of completing the offset mismatch calibration and the gain mismatch calibration in each sub-channel, the first reference sub-channel, the second reference sub-channel and the sub-channel to be calibrated with timing are iteratively determined in each sub-channel. Since in the process of iteratively determining the first reference sub-channel, the second reference sub-channel and the sub-channel to be calibrated with timing, except for the first determination process, the remaining determination processes can simultaneously determine multiple sub-channels to be calibrated with timing, that is, the calibration of multiple sub-channels to be calibrated with timing is completed at the same time, thereby improving the timing calibration rate, and thus improving the efficiency of the foreground calibration of the time-interleaved analog-to-digital converter. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0035] Figure 1 It is a schematic diagram of the architecture of a SERDES system in the prior art;

[0036] Figure 2 It is a schematic diagram of the structural principle of a time-interleaved analog-to-digital converter;

[0037] Figure 3 It is a schematic diagram of the principle of a time-interleaved analog-to-digital converter;

[0038] Figure 4 It is a calibration schematic diagram of a mismatch effect provided by the present invention;

[0039] Figure 5 It is a schematic diagram of a mismatch calibration structure of a time-interleaved analog-to-digital converter provided by the present invention;

[0040] Figure 6 A time-interleaved analog-to-digital converter sub-channel mismatch calibration method provided by the present invention;

[0041] Figure 7 It is a schematic diagram of an implementation architecture of an offset mismatch calibration provided by the present invention;

[0042] Figure 8 It is a schematic diagram of a gain mismatch calibration structure provided by the present invention;

[0043] Fig. 9 A schematic flow chart of a method for determining a gain mismatch value of each non-reference sub-channel provided by the present invention;

[0044] Fig.10 It is a flow chart of a method for determining a first reference sub-channel, a second reference sub-channel and a sub-channel to be time-series calibrated in a plurality of calibration groups corresponding to a reference calibration sub-channel, proposed by the present invention;

[0045] Fig.11 It is a flow chart of a method for determining a timing adaptation value of a sub-channel to be timing calibrated provided by the present invention;

[0046] Fig.12 It is a schematic diagram of an implementation architecture of a timing mismatch calibration provided by the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution in the embodiment of the present invention will be clearly and completely described in conjunction with the drawings in the embodiment of the present invention. It should be understood that the drawings in the present invention only serve the purpose of illustration and description and are not used to limit the scope of protection of the present invention. In addition, it should be understood that the schematic drawings are not drawn in real proportion. The flowcharts used in the present invention illustrate the operations implemented according to some embodiments of the present invention. It should be understood that the operations of the flowchart can be implemented out of sequence, and the steps without logical context can be reversed in order or implemented simultaneously. In addition, under the guidance of the content of the present invention, those skilled in the art can add one or more other operations to the flowchart, or remove one or more operations from the flowchart.

[0048] In addition, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present invention.

[0049] It should be noted that the term “comprising” will be used in the embodiments of the present invention to indicate the existence of the features declared thereafter, but does not exclude the addition of other features.

[0050] SERDES is a common serial communication technology. SERDES systems usually include a transmitter for serializing parallel signals and a receiver for parallelizing serial signals, which is used to achieve high-speed signal transmission between electronic devices while greatly reducing the number of connecting lines. SERDES systems are widely used in computer networks, communication systems, storage devices, graphic displays and other fields.

[0051] As the demand for SERDES system transmission rate becomes higher and higher, for example, the Ethernet protocol stipulates that the single-channel bit rate can reach 112 Gbps. The insertion loss at the Nyquist frequency caused by electrical characteristics is also increasing. The traditional analog system architecture limits the application of advanced equalizers. Therefore, a high-speed SERDES system architecture is introduced. The high-speed SERDES system architecture usually adopts a digital reception method based on a digital signal processor, that is, a multi-channel time interleaving method is used to achieve a balance between low power consumption and high sampling rate.

[0052] Specifically, Figure 1 Schematic diagram of the architecture of the SERDES system in the prior art. Figure 1As shown, the SERDES system in the prior art includes a transmitter, a channel equalizer, an automatic gain controller, an analog-to-digital converter, a digital signal processor, and a deserializer. Compared with the traditional technology, the prior art is not limited to the channel equalizer and the decision feedback equalizer, but after converting the analog signal into a digital signal through a time-interleaved analog-to-digital converter, a more complex equalizer with better equalization effect can be used in the digital domain to perform equalization processing on the digital signal, thereby reducing insertion loss and improving performance.

[0053] In the prior art, since the high-speed signal transmission requirement of the SERDES system needs to be met, the high-speed signal transmission requirement takes a 112-gigabit-per-second SERDES system with four-level pulse amplitude modulation as an example. According to the Nyquist sampling theorem, the system sampling rate of the analog-to-digital converter must reach at least 56 gigahertz per second. Therefore, a time-interleaved analog-to-digital converter is used as an analog-to-digital conversion processing device to increase the sampling rate. Figure 1 As shown, there are multiple subchannels, such as subchannel 0 to subchannel M-1, with a total of M subchannels.

[0054] Furthermore, the time-interleaved analog-to-digital converter samples at one-Mth of the system sampling rate through M sub-channels, and the sampling results are combined in a time-interleaved manner, so as to achieve the effect of a single-channel analog-to-digital converter sampling at the system sampling rate, thereby improving the overall sampling rate. Because the sampling rate of each sub-channel is reduced to one-Mth of the system sampling rate, it can meet lower power consumption on the basis of easy circuit design. However, in the actual circuit design of the time-interleaved analog-to-digital converter, due to the different line connections of each sub-channel, mismatch effects will appear between the sub-channels. This mismatch effect may cause spectral spurious components, greatly limiting the performance indicators of the time-interleaved analog-to-digital converter. Since the mismatch effect cannot be completely avoided by optimizing the design, it is necessary to calibrate the mismatch effect between channels before the time-interleaved analog-to-digital converter actually works. Among them, the mismatch effect includes offset mismatch, gain mismatch and timing mismatch.

[0055] In the prior art, at least one additional reference sub-channel whose sampling rate is mutually prime with each sub-channel is usually introduced during the mismatch estimation of the timing mismatch, which will cause complex circuit design and high power consumption.

[0056] Based on the above problems, the present invention proposes a time-interleaved analog-to-digital converter sub-channel mismatch calibration method, which performs offset mismatch calibration on each sub-channel according to the first signal output by each sub-channel after each sub-channel of the time-interleaved analog-to-digital converter is respectively connected to the first injection signal. After the calibration is completed, the second injection signal is injected into each sub-channel respectively, and the gain mismatch calibration is performed on each sub-channel according to the second signal output by the reference sub-channel and the non-reference sub-channel. After the calibration is completed, the third injection signal is injected into each sub-channel respectively, and the first reference sub-channel, the second reference sub-channel and the sub-channel to be time-series calibrated are iteratively determined, and then the timing mismatch calibration is performed on the sub-channel to be time-series calibrated according to the third signal output by the first reference sub-channel, the second reference sub-channel and the sub-channel to be time-series calibrated. The present invention implements calibration by iteratively determining the first reference sub-channel, the second reference sub-channel and the sub-channel to be time-series calibrated in each sub-channel, avoiding the setting of additional reference sub-channels, and the circuit design is simple and the power consumption is low.

[0057] Before introducing the time-interleaved analog-to-digital converter sub-channel mismatch calibration method in detail, the implementation principle of the time-interleaved analog-to-digital converter is first explained.

[0058] After receiving the analog signal, each sub-channel will work alternately according to the uniform clock phase. The output signal of each sub-channel is synthesized in the working order using a multiplexer to achieve the function that the sampling rate meets the system sampling rate.

[0059] Figure 2 It is a schematic diagram of the structural principle of a time-interleaved analog-to-digital converter. Figure 3 This is a schematic diagram of the principle of a time-interleaved analog-to-digital converter. Figure 2 As shown, taking the time interleaved analog-to-digital converter including 4 sub-channels as an example, sub-channel 0 receives clock signal 0, sub-channel 1 receives clock signal 1, sub-channel 2 receives clock signal 2, and sub-channel 3 receives clock signal 3. Figure 3 As shown, the frequency of each clock signal is one quarter of the frequency of the reference clock signal, and has a phase offset of a pulse period of the reference clock signal in turn, so that each sub-channel processes a narrower frequency band according to each clock signal, samples the analog signal, and then uses a multiplexer to synthesize it in the working order.

[0060] During the operation of a time-interleaved analog-to-digital converter, mismatch effects such as offset mismatch, gain mismatch and timing mismatch may occur, so a combination of foreground calibration and background calibration is usually used to implement mismatch calibration. The present invention is directed to the foreground calibration of a time-interleaved analog-to-digital converter. Foreground calibration is to calibrate the mismatch of the time-interleaved analog-to-digital converter using an injection signal before the time-interleaved analog-to-digital converter starts to work, thereby improving linearity.

[0061] Next, based on the understanding of the structure and principle of the time-interleaved analog-to-digital converter, Figure 4 FIG. 1 is a calibration diagram of a mismatch effect provided by the present invention. Figure 4 As shown, taking the time-interleaved analog-to-digital converter containing 4 sub-channels as an example, after the analog signal is connected, each sub-channel samples the connected signal according to each clock signal, and all are based on the offset mismatch calibration value , gain mismatch calibration value and timing mismatch calibration values After calibration, the analog signal is quantized and converted into a digital signal and then output through a multiplexer, thereby achieving the mismatch calibration process.

[0062] Specifically, to clearly explain how to perform mismatch calibration on a time-interleaved analog-to-digital converter, Figure 5 FIG. 1 is a schematic diagram of a mismatch calibration structure of a time-interleaved analog-to-digital converter provided by the present invention. Figure 5 As shown, after the time-interleaved analog-to-digital converter is connected to the injection signal, each sub-channel receives the signal output by the demultiplexer and samples it based on the clock signal with phase offset. The mismatch calibration process is divided into offset mismatch calibration, gain mismatch calibration and timing mismatch calibration. Figure 5 In the figure, the mismatch calibration process is illustrated by calibrating three subchannels as an example, where the timing mismatch of subchannel 0 and subchannel 2k is consistent, that is, subchannel 0 and subchannel 2k are sampled based on the same clock signal, and subchannel k is a subchannel whose sampling point is between the sampling point of subchannel 0 and the sampling point of subchannel 2k. In the mismatch calibration process: first determine the mismatch value according to the signal output by each subchannel, and determine the mismatch calibration value according to the mismatch value , and , thereby calibrating each sub-channel. After the offset mismatch calibration, the gain mismatch calibration process is performed: adjust the injection signal and determine the reference sub-channel, Figure 5 Taking subchannel 0 as the reference subchannel example, the signals output by each subchannel are received, the gain mismatch values ​​of each non-reference subchannel are determined according to the signals output by the reference subchannel and the signals output by the non-reference subchannel, and the gain mismatch calibration value is determined according to the gain mismatch values. and After completing the offset mismatch calibration and gain mismatch calibration, the timing mismatch calibration process is performed: the injection signal is adjusted and two reference sub-channels are determined. Figure 5 Taking subchannel 0 and subchannel 2k as reference subchannel examples, the signals output by each subchannel are received, the timing mismatch value of the non-reference subchannel is determined according to the signals output by the two reference subchannels, and the timing mismatch calibration value is determined according to the timing mismatch value. Thus, the non-reference sub-channel is calibrated. From this point on, the sub-channel foreground mismatch calibration of the time-interleaved analog-to-digital converters of the three sub-channels is completed.

[0063] The present invention is based on the above-mentioned method of obtaining the offset mismatch calibration value on the basis of simple circuit design and low power consumption. , gain mismatch calibration value and timing mismatch calibration values For detailed methods, refer to Figure 6 . Figure 6 The present invention provides a method for calibrating the mismatch of a time-interleaved analog-to-digital converter subchannel. Figure 6 The sub-channel mismatch calibration method of time-interleaved analog-to-digital converter is introduced.

[0064] Step 601: After each sub-channel of the time-interleaved analog-to-digital converter is respectively connected to the first injection signal, the offset mismatch value of each sub-channel is determined according to the first signal output by each sub-channel, and the offset mismatch calibration is performed on each sub-channel according to the offset mismatch value of each sub-channel. The first injection signal is a DC signal of 0 volts.

[0065] Specifically, after injecting a 0V DC injection signal into each sub-channel, the first signal output by each sub-channel is collected, and the offset mismatch value of each sub-channel is determined according to the first signal output by each sub-channel. As an optional implementation, the statistical average value of the digital signal output by each sub-channel can be used as the offset mismatch value of the current sub-channel. Exemplarily, the offset mismatch value of the sub-channel can be expressed by the following formula (1):

[0066] (1)

[0067] in, is the offset mismatch value of the sub-channel, N is the total number of sampling times, It is the digital signal output by the nth sampling point of the kth sub-channel in the time-interleaved analog-to-digital converter sub-channel. Indicates that the sub-channel input signal is a 0V DC signal.

[0068] If the time-interleaved analog-to-digital converter includes M sub-channels, the equivalent system sampling interval of the time-interleaved analog-to-digital converter is , then the output interval of the offset mismatch value is .

[0069] As an optional implementation, after the offset mismatch value of each sub-channel is calculated, the offset mismatch calibration value can be calculated according to the offset mismatch value of each channel by using a filtering method, and then the offset mismatch calibration value is fed back to the analog compensation circuit for offset mismatch compensation. In actual calculations, an offset mismatch calibration estimate is usually calculated to replace the offset mismatch calibration value. For example, the offset mismatch calibration estimate can be calculated based on alpha filtering by the following formula (2):

[0070] (2)

[0071] in, is the current offset mismatch calibration estimate, is the offset mismatch calibration estimate corresponding to the previous output interval, is the preset offset loop parameter. It is worth noting that when sampling and calibration are performed for the first time, the current offset mismatch value calculation formula The value of is 0.

[0072] To further describe the offset mismatch calibration process, Figure 7 It is a schematic diagram of an implementation architecture of an offset mismatch calibration provided by the present invention. Figure 7 In the example, a first injection signal of 0V DC is first injected into the input terminal, and then a clock signal with a frequency of one-Mth of the reference clock signal is input into the demultiplexer, so that each subchannel is sampled based on the clock signals with different phases. Taking subchannel k as an example, the first signal outputted by subchannel k for the Nth time is The first signal obtained from the previous sampling points is accumulated and summed, and then the sum is divided by the total number of sampling times N to obtain the offset mismatch value Then the offset mismatch value With preset offset loop parameters The offset mismatch calibration estimate corresponding to the previous output interval is obtained by multiplying The sum of the offset and mismatch calibration estimates is used as the current offset mismatch calibration estimate. . Then the current offset mismatch calibration estimate is The feedback is fed back to the analog compensation circuit to perform offset mismatch compensation on the current sub-channel.

[0073] Step 602: After each sub-channel completes the offset mismatch calibration and is respectively connected to the second injection signal, the gain mismatch value of each non-reference sub-channel is determined according to the second signal output by the reference sub-channel and the non-reference sub-channel, and the gain mismatch calibration is performed on each non-reference sub-channel according to the gain mismatch value of each non-reference sub-channel. The second injection signal is a DC signal.

[0074] Specifically, after each sub-channel completes the offset mismatch calibration, a second injection signal is respectively connected to each sub-channel. The second injection signal is a DC signal. As an optional implementation, the second injection signal can be a DC signal of a half-scale of a time-interleaved analog-to-digital converter.

[0075] After the second injection signal is connected, any subchannel is determined as a reference subchannel, and the subchannels other than the reference subchannel in each subchannel are determined as non-reference subchannels. Then, the second signal output by the reference subchannel and the non-reference subchannel is obtained. The gain mismatch value of each non-reference subchannel is determined according to the second signal of the reference subchannel and the second signal of the non-reference subchannel.

[0076] As an optional implementation, the absolute error method can be used to calculate the gain mismatch value of each sub-channel. Specifically, the cumulative sum of the difference between the absolute value of the second signal output by the non-reference sub-channel and the absolute value of the second signal output by the reference sub-channel is calculated, and the gain mismatch value of each sub-channel is obtained by dividing the above cumulative sum by the total number of sampling times. For example, taking the reference sub-channel as sub-channel 0 as an example, the calculation method of the gain mismatch value can be expressed by formula (3):

[0077] (3)

[0078] in, is the gain mismatch value of the non-reference sub-channel, N is the total number of sampling times, is a digital signal outputted by the nth sampling point of the kth sub-channel in the sub-channels of the time-interleaved analog-to-digital converter, wherein the kth sub-channel is a non-reference sub-channel, It is the nth sampling point of the digital signal output by the reference sub-channel in the sub-channel of the time-interleaved analog-to-digital converter. It indicates that the second injection signal inputted by each sub-channel is a DC signal. The digital signal outputted from the nth sampling point of the kth sub-channel in the sub-channels of the time-interleaved analog-to-digital converter is a signal that has been calibrated for offset mismatch.

[0079] If the time-interleaved analog-to-digital converter includes M sub-channels, the equivalent system sampling interval of the time-interleaved analog-to-digital converter is , then the output interval of the gain mismatch value is It is worth noting that the output intervals for the gain mismatch values ​​are the same as the output intervals for the offset mismatch values.

[0080] As an optional implementation, after the gain mismatch values ​​of each non-reference sub-channel are calculated, a gain mismatch calibration value can be calculated based on the gain mismatch values ​​of each reference sub-channel using a filtering method, and then the gain mismatch calibration value is fed back to the analog compensation circuit for gain mismatch compensation. In actual calculations, a gain mismatch calibration estimate is usually calculated to replace the gain mismatch calibration value. For example, the gain mismatch calibration estimate can be calculated based on alpha filtering using the following formula (4):

[0081] (4)

[0082] in, is the current gain mismatch calibration estimate, is the gain mismatch calibration estimate corresponding to the previous output interval, is the preset gain loop parameter. It is worth noting that when sampling and calibration are performed for the first time, the current gain mismatch value calculation formula The value of is 0.

[0083] To further describe the gain mismatch process, Figure 8 Schematic diagram of a gain mismatch calibration structure provided by the present invention. Figure 8 In the process, a DC second injection signal is first injected into the input end, and then a clock signal with a frequency of one-Mth of the reference clock signal is input into the demultiplexer, so that each subchannel is sampled based on each clock signal with a different phase. Taking subchannel 0 as the reference subchannel and calibrating subchannel k as an example, the absolute value of the second signal output by subchannel 0 and the second signal obtained at each previous sampling point are calculated and cumulatively added, and the sum is divided by the total number of sampling times N to obtain a first quotient, and the absolute value of the second signal output by subchannel k and the second signal obtained at each previous sampling point are calculated and cumulatively added, and the sum is divided by the total number of sampling times N to obtain a second quotient, and the first quotient and the second quotient are added to obtain a gain mismatch value. , the gain mismatch value Multiply by the preset gain loop parameters Gain mismatch correction estimate corresponding to the previous output interval The sum of the gain mismatch calibration estimate is . Then the current gain mismatch calibration estimate is The feedback is sent to the analog compensation circuit to perform gain mismatch compensation on the current reference sub-channel.

[0084] It is worth noting that the gain mismatch calibration estimate of the reference subchannel is 0.

[0085] Step 603: After each non-reference sub-channel completes gain offset mismatch calibration and each sub-channel is respectively connected to the third injection signal, iteratively determine the first reference sub-channel, the second reference sub-channel and the sub-channel to be timing calibrated, and, based on the third signal output by the first reference sub-channel, the third signal output by the second reference sub-channel and the third signal output by the sub-channel to be timing calibrated, perform timing mismatch calibration on the sub-channel to be timing calibrated.

[0086] Specifically, after each sub-channel completes mismatch calibration and each non-reference sub-channel completes gain mismatch calibration, a third injection signal is connected to each sub-channel. The third injection signal may be a triangular wave signal of a preset frequency. In order to ensure the robustness of the timing mismatch calibration of the timing calibration sub-channel according to the third signal, the ratio of the preset frequency to the frequency of the reference clock signal is the preset value to 1, and the greatest common divisor of the preset ratio and the number of sub-channels should be less than and equal to 2.

[0087] After accessing the third injection signal, the first reference sub-channel and the second reference sub-channel are iteratively determined from each sub-channel, wherein the first reference sub-channel and the second reference sub-channel are any two sub-channels with the same phase timing mismatch in each sub-channel. Specifically, the first reference sub-channel and the second reference sub-channel are first determined from each sub-channel, and then the sub-channel to be timing calibrated is determined based on the first reference sub-channel and the second reference sub-channel. Among them, the sub-channel to be timing calibrated is the sub-channel located between the first reference sub-channel and the second reference sub-channel in the first calibration group. Exemplarily, if the first reference sub-channel is sub-channel 0 and the second reference sub-channel is sub-channel 4, the sub-channel to be timing calibrated is sub-channel 2.

[0088] Then, the subchannel to be calibrated for timing mismatch is calibrated according to the third signal output by the first reference subchannel, the second reference subchannel and the subchannel to be calibrated for timing. After the calibration is completed, the first reference subchannel and the second reference subchannel of the next round are iteratively determined according to the current first reference subchannel, the second reference subchannel and the subchannel to be calibrated for timing. The first reference subchannel or the second reference subchannel of this round can be the first reference subchannel, the second reference subchannel and the subchannel to be calibrated for timing after the previous iteration. Since the first reference subchannel, the second reference subchannel and the subchannel to be calibrated for timing have been calibrated for timing mismatch in the previous iteration, the timing mismatch of the first reference subchannel and the second reference subchannel of the current round is the same. The above steps are iteratively executed until the timing mismatch calibration of all subchannels is completed.

[0089] It is worth mentioning that in the above-mentioned iterative process of determining the first reference sub-channel and the second reference sub-channel, the channel interval between the first reference sub-channel and the second reference sub-channel in the previous round is twice the channel interval between the first reference sub-channel and the second reference sub-channel in the current round. The channel interval is the phase shift of the clock signal referenced by each channel. In other words, based on the determination of the first reference sub-channel and the second reference sub-channel in the previous round, two groups of first reference sub-channels and second reference sub-channels can be determined in the current round. Therefore, if the time-interleaved analog-to-digital converter has a total of M sub-channels, it takes 10 iterations to complete the timing calibration. Secondly, compared with calibrating the timing offset of each sub-channel one by one, this method greatly improves the timing calibration rate of each sub-channel.

[0090] As an optional implementation, the mathematical modeling reference formulas of the above three mismatch errors can be shown as the following formula (5):

[0091] (5)

[0092] in, is the digital signal output by the nth sampling point of the kth sub-channel in the sub-channel of the time-interleaved analog-to-digital converter, is the kth subchannel of the time-interleaved analog-to-digital converter The analog signal is sampled at a sampling frequency, M is the total number of sub-channels, is the equivalent system sampling interval, is the offset mismatch calibration value, is the gain mismatch calibration value, is the timing mismatch calibration value. It is worth mentioning that the offset mismatch calibration value in formula (5) is can be equated to the offset mismatch calibration estimate obtained in the above formula (2): , the gain mismatch calibration value in formula (5) It can be equated to the gain mismatch calibration estimate obtained in the above formula (4): . Timing mismatch calibration value Same reason.

[0093] In this embodiment, after each sub-channel of the time-interleaved analog-to-digital converter is respectively connected to the first injection signal, each sub-channel is calibrated for offset mismatch according to the first signal output by each sub-channel. After the calibration is completed, the second injection signal is injected into each sub-channel, and the gain mismatch calibration is performed on each sub-channel according to the second signal output by the reference sub-channel and the non-reference sub-channel. After the calibration is completed, the third injection signal is injected into each sub-channel, and the first reference sub-channel, the second reference sub-channel and the sub-channel to be calibrated are iteratively determined, and then the timing mismatch calibration is performed on the sub-channel to be calibrated according to the third signal output by the first reference sub-channel, the second reference sub-channel and the sub-channel to be calibrated. In this embodiment, the reference sub-channel or the first reference sub-channel and the second reference sub-channel are determined in each sub-channel during the gain mismatch calibration and the timing mismatch calibration process, so there is no need to use a channel other than the original sub-channel of the time-interleaved analog-to-digital converter as a reference sub-channel, so that the circuit design is simple, the area occupied is small, and the power consumption is low. Furthermore, by iteratively determining the first reference sub-channel, the second reference sub-channel and the sub-channel to be timing calibrated in each sub-channel on the basis of completing the offset mismatch calibration and the gain mismatch calibration in each sub-channel, since in the process of iteratively determining the first reference sub-channel, the second reference sub-channel and the sub-channel to be timing calibrated, except for the first determination process, the remaining determination processes can simultaneously determine multiple sub-channels to be timing calibrated, that is, calibration of multiple sub-channels to be timing calibrated is completed at the same time, thereby improving the timing calibration rate, and further improving the efficiency of the foreground calibration of the time-interleaved analog-to-digital converter.

[0094] As an optional implementation manner, determining the offset mismatch value of each sub-channel according to the first signal output by each sub-channel in the above step 601 also includes the following implementation manner:

[0095] Optionally, the expected value of each first signal is used as the offset mismatch value of each sub-channel.

[0096] The expected value may describe the average situation of each sampling point in the first signal, and the expected value of the first signal is used as the offset mismatch value of each sub-channel, so as to accurately calibrate the offset mismatch of each sub-channel.

[0097] As an optional implementation, Fig. 9 A schematic flow chart of a method for determining the gain mismatch value of each non-reference sub-channel provided by the present invention. Fig. 9 As shown, the above step 602 of determining the gain mismatch value of each non-reference sub-channel according to the second signal output by the reference sub-channel and the non-reference sub-channel also includes the following implementation methods:

[0098] Step 901: Determine a reference variance of a second signal output by a reference subchannel.

[0099] Specifically, a reference variance of each sampling point in the second signal output by the reference subchannel is determined.

[0100] Step 902: Determine a plurality of first variances of the second signal output by each non-reference sub-channel.

[0101] Specifically, a reference variance of each sampling point in the second signal output by each non-reference sub-channel is determined.

[0102] Step 903: Calculate the difference between the reference variance and each first variance respectively, and use each difference as the gain mismatch value corresponding to each non-reference sub-channel.

[0103] Specifically, a first variance may be calculated for each non-reference sub-channel, and thus a difference may be calculated for each non-reference sub-channel.

[0104] In this embodiment, by determining the reference variance of the reference subchannel and the first variance corresponding to each non-reference subchannel, the difference between the reference variance and each first variance is used as the gain mismatch value of each non-reference subchannel, thereby accurately calibrating the gain mismatch of each subchannel.

[0105] Next, how to perform timing mismatch calibration is described in detail.

[0106] As an optional implementation manner, in the above step 603, the first reference sub-channel, the second reference sub-channel and the sub-channel to be timing calibrated are iteratively determined, and, according to the third signal output by the first reference sub-channel, the third signal output by the second reference sub-channel and the third signal output by the sub-channel to be timing calibrated, the specific method of performing timing mismatch calibration on the sub-channel to be timing calibrated includes:

[0107] Optionally, in each iteration, multiple calibration groups are determined based on the dichotomy method, each calibration group includes: a first reference subchannel, a second reference subchannel and a subchannel to be timing calibrated, and the first reference subchannel and the second reference subchannel have the same timing mismatch.

[0108] As an optional implementation, multiple calibration groups and the first reference sub-channel, the second reference sub-channel and the sub-channel to be calibrated can be determined in advance based on the binary method. Then, in each iteration, the timing mismatch value of the sub-channel to be calibrated is directly determined according to the third signal output by the first reference sub-channel, the second reference sub-channel and the sub-channel to be calibrated determined in each calibration group.

[0109] It is worth noting that the phase timing mismatches of the first reference subchannel and the second reference subchannel in each calibration group are the same, so that they can be used as reference subchannels, so that the timing mismatch of the calibrated timing calibration subchannel is the same as the timing mismatch of the first reference subchannel and the second reference subchannel. When the iteration is completed, the timing mismatches of each subchannel in the time-interleaved analog-to-digital converter are the same, thereby achieving timing mismatch calibration.

[0110] Optionally, in each calibration group, the timing mismatch value of the sub-channel to be calibrated is determined based on the third signal output by the first reference sub-channel, the third signal output by the second reference sub-channel, and the third signal output by the sub-channel to be calibrated, and the timing mismatch calibration is performed on the sub-channel to be calibrated based on the timing mismatch value of the sub-channel to be calibrated.

[0111] As an optional implementation, the timing mismatch value of the subchannel to be timing calibrated can be determined based on the variance of the difference between the third signal output by the second reference subchannel and the third signal output by the subchannel to be timing calibrated, and the variance of the difference between the third signal output by the subchannel to be timing calibrated and the third signal output by the first reference subchannel.

[0112] In this embodiment, in each iteration, multiple calibration groups are determined based on the binary method, and each calibration group includes: a first reference subchannel, a second reference subchannel, and a subchannel to be calibrated. In each calibration group, the timing mismatch value of the subchannel to be calibrated is determined according to the third signal output by the first reference subchannel, the third signal output by the second reference subchannel, and the third signal output by the subchannel to be calibrated, and the timing mismatch calibration is performed on the subchannel to be calibrated according to the timing mismatch value of the subchannel to be calibrated. In this embodiment, multiple calibration groups are determined by iteration, so that the calibration groups determined in each round can simultaneously realize the timing calibration of the subchannel to be calibrated, thereby improving the timing calibration rate.

[0113] Furthermore, the following steps illustrate how to determine multiple calibration groups based on the dichotomy method.

[0114] Optionally, each sub-channel to be time-series calibrated in the previous iteration is used as a reference calibration sub-channel.

[0115] The reference calibration sub-channels are all sub-channels that have undergone timing calibration.

[0116] Optionally, according to each reference calibration subchannel, a first reference subchannel, a second reference subchannel and a subchannel to be time-calibrated in a plurality of calibration groups corresponding to the reference calibration subchannel are determined, wherein the first reference subchannel in each calibration group is the reference calibration subchannel.

[0117] Specifically, one reference calibration sub-channel may correspond to determine the first reference sub-channel, the second reference sub-channel and the sub-channel to be calibrated in the two calibration groups.

[0118] Optionally, the second reference subchannel in each calibration group is the first reference subchannel or the second reference subchannel in the previous iteration.

[0119] In this embodiment, each sub-channel to be timing calibrated in the previous iteration is used as a reference calibration sub-channel respectively, and according to each reference calibration sub-channel, the first reference sub-channel, the second reference sub-channel and the sub-channel to be timing calibrated in the multiple calibration groups corresponding to the reference calibration sub-channel are determined, so that the sub-channels to be timing calibrated can all perform timing calibration with reference to the sub-channels with the same timing mismatch, thereby meeting the accuracy requirement of the timing calibration.

[0120] Further, Fig.10 This is a flow chart of a method for determining the first reference sub-channel, the second reference sub-channel and the sub-channel to be calibrated in a plurality of calibration groups corresponding to the reference calibration sub-channel proposed by the present invention. Fig.10 The specific steps of determining the first reference sub-channel, the second reference sub-channel and the sub-channel to be time-series calibrated in the multiple calibration groups corresponding to the reference calibration sub-channel are introduced.

[0121] Step 1001: Use a reference calibration subchannel as a first reference subchannel in a first calibration group and a first reference subchannel in a second calibration group corresponding to the reference calibration subchannel.

[0122] The calibration groups on both sides of the reference calibration subchannel are used as the first calibration group and the second calibration group.

[0123] As an optional implementation, in the first round of iterations, any two sub-channels with consistent timing mismatch in each sub-channel in the time-interleaved analog-to-digital converter are used as the first reference sub-channel and the second reference sub-channel.

[0124] Step 1002: Use two subchannels that meet preset conditions as the second reference subchannel in the first calibration group and the second reference subchannel in the second calibration group, respectively, wherein the preset conditions include: the channel interval between the subchannel and the reference calibration subchannel is half of the channel interval used in the previous iteration.

[0125] The channel interval is the phase shift of the clock signal referenced by each channel. As an example, if the time-interleaved analog-to-digital converter includes sub-channel 0, sub-channel 1, and sub-channel 2, and the phase difference between sub-channel 0 and sub-channel 2 is 1 clock cycle, that is, 2π, then the channel intervals of the three sub-channels are .

[0126] For example, if the channel interval between the subchannel in the previous iteration and the reference calibration subchannel is , then in the current iteration, the channel interval between the subchannel and the reference calibration subchannel is .

[0127] Step 1003: use the subchannel located between the first reference subchannel and the second reference subchannel in the first calibration group as the subchannel to be timing calibrated in the first calibration group, and use the subchannel located between the first reference subchannel and the second reference subchannel in the second calibration group as the subchannel to be timing calibrated in the second calibration group.

[0128] Specifically, the phase shift of the clock signal referenced by the subchannel between the first reference subchannel and the second reference subchannel in the calibration group is half of the phase shift difference between the clock signals referenced by the first reference subchannel and the second reference subchannel.

[0129] As an example, if the time-interleaved analog-to-digital converter includes five subchannels, namely, subchannel 0, subchannel 1, subchannel 2, subchannel 3 and subchannel 4, the timing mismatch of subchannel 0 and subchannel 4 is consistent. In the first round of iteration, subchannel 0 is used as the first reference subchannel and subchannel 4 is used as the second reference subchannel, and subchannel 2 is the subchannel to be calibrated for timing. After timing calibration of subchannel 2, it is used as the reference calibration subchannel. The calibration group adjacent to the reference calibration subchannel is used as the first calibration group and the second calibration group, and the reference calibration subchannel group is used as the first reference subchannel of the first calibration group and the second calibration group. Then, according to the phase shift of the reference clock signal between subchannel 0 and subchannel 4, the second reference subchannel of the first calibration group and the second reference subchannel of the second calibration group, as well as subchannel 0 and subchannel 4 are determined, and then the channel between subchannel 0 and subchannel 2 is used as the subchannel to be calibrated for timing of the first calibration group, that is, subchannel 2, and the channel between subchannel 2 and subchannel 4 is used as the subchannel to be calibrated for timing of the second calibration group, that is, subchannel 3. This process continues in this way until the timing mismatches of all sub-channels are consistent.

[0130] In this embodiment, the benchmark calibration subchannel is used as the first reference subchannel in the first calibration group and the first reference subchannel in the second calibration group corresponding to the benchmark calibration subchannel, the two subchannels that meet the preset conditions are used as the second reference subchannels in the first calibration group and the second calibration group respectively, and the subchannel located between the first reference subchannel and the second reference subchannel in the first calibration group and the second calibration group is used as the subchannel to be timing calibrated in the first calibration group and the second calibration group, so as to ensure that the timing mismatches of the first reference subchannel and the second reference subchannel of each calibration group are the same, and after the timing calibration subchannel to be timing calibrated is performed based on the first reference subchannel and the second reference subchannel, the timing mismatches of each subchannel are the same.

[0131] After describing how to determine the first reference sub-channel, the second reference sub-channel, and the sub-channel to be calibrated, as an optional implementation, the following refers to Fig.11 How to determine the timing mismatch value of the sub-channel to be calibrated is introduced. Fig.11 It is a flow chart of a method for determining a timing adaptation value of a sub-channel to be timing calibrated provided by the present invention.

[0132] Step 1101: Determine a first expected value according to a third signal of the sub-channel to be time-calibrated and a third signal of the first reference sub-channel.

[0133] As an example, taking the first reference sub-channel as sub-channel 0 and the sub-channel to be time-series calibrated as sub-channel 1 as an example, the first expected value can be determined by referring to formula (6):

[0134] (6)

[0135] in, The third signal output by subchannel 0. is the third signal output by subchannel 1, is the first expected value. is the variance of each sampling point in the third signal output by subchannel 0, is the variance of each sampling point in the third signal output by subchannel 1, R is the autocorrelation function of the timing error of subchannel c in the time-interleaved analog-to-digital converter, is the timing error of sub-channel c, and sub-channel c is the sub-channel to be calibrated. is the equivalent system sampling interval of the time-interleaved analog-to-digital converter. Among them, the autocorrelation function of the timing error of sub-channel c can describe the linear correlation degree of the timing error of sub-channel c at different times.

[0136] Step 1102: Determine a second expected value according to the third signal of the sub-channel to be time-calibrated and the third signal of the second reference sub-channel.

[0137] As an example, taking the second reference sub-channel as sub-channel 2 and the sub-channel to be time-series calibrated as sub-channel 1 as an example, the first expected value can be determined by referring to formula (7):

[0138] (7)

[0139] in, is the third signal output by subchannel 2, is the third signal output by subchannel 1, is the second expected value. is the variance of each sampling point in the third signal output by subchannel 2, is the variance of each sampling point in the third signal output by subchannel 1, R is the autocorrelation function of the timing error of subchannel c in the time-interleaved analog-to-digital converter, is the timing error of sub-channel c, and sub-channel c is the sub-channel to be calibrated. is the equivalent system sampling interval of the time-interleaved analog-to-digital converter.

[0140] It is worth mentioning that since the third injection signal is a triangular wave signal of a specific frequency, the relationship described in formula (8) can be obtained:

[0141] (8)

[0142] Step 1103: determine the timing mismatch value of the sub-channel to be timing calibrated based on the first expected value, the second expected value and the preset constraint condition, wherein the preset constraint condition indicates that the frequency of the third injection signal is equal to the ratio of the frequency of the reference clock signal to the preset value, and the lowest common divisor of the preset value and the number of sub-channels of the time-interleaved analog-to-digital converter is less than or equal to 2.

[0143] As an optional implementation, according to formula (6) and formula (7), the cost function of the timing mismatch of the sub-channel to be time-calibrated can be obtained. The cost function of the timing mismatch can be expressed by the following formula (9):

[0144] (9)

[0145] As an optional implementation, the autocorrelation function of the sub-channel timing error is related to the frequency of the third injection signal. Specifically, if the ratio of the frequency of the third injection signal to the frequency of the reference clock signal is a preset ratio of 1, and the greatest common divisor of the preset ratio and the number of sub-channels is less than and equal to 2, then the autocorrelation function of the sub-channel in the time-interleaved analog-to-digital converter is in a monotonically increasing or monotonically decreasing state.

[0146] Therefore, according to formula (6), formula (7), formula (8) and formula (9), formula (10) can be obtained:

[0147] (10)

[0148] in, It can be the first-order Taylor expansion of the autocorrelation function of the timing error of subchannel c.

[0149] Furthermore, based on formula (10), the autocorrelation function of the timing error of subchannel c is guaranteed to be In the case of monotonically increasing or monotonically decreasing, the timing mismatch value can be determined by the following formula (11):

[0150] (11)

[0151] in, is the timing mismatch value of the sub-channel to be calibrated, N is the total number of sampling times, is the digital signal output by the nth sampling point of subchannel 0 of the time-interleaved analog-to-digital converter, is the digital signal output by the nth sampling point of subchannel 1 of the time-interleaved analog-to-digital converter, is the digital signal output by the nth sampling point of subchannel 2 of the time-interleaved analog-to-digital converter, Indicates that the sub-channel to be calibrated has undergone offset mismatch calibration and gain mismatch calibration.

[0152] If the time-interleaved analog-to-digital converter includes M sub-channels, the equivalent system sampling interval of the time-interleaved analog-to-digital converter is , then the output interval of the timing mismatch value is .

[0153] In this embodiment, the first expected value and the second expected value are determined according to the third signal of the sub-channel to be timing calibrated, the third signal of the first reference sub-channel and the third signal of the second reference sub-channel, and then the timing mismatch value of the sub-channel to be timing calibrated is determined according to the first expected value, the second expected value and preset constraints. This process does not require explicit solution of the autocorrelation function of the sub-channel timing error, thereby reducing the computational complexity.

[0154] As an optional implementation, to further describe the timing mismatch calibration process, Fig.12 It is a schematic diagram of an implementation architecture of a timing mismatch calibration provided by the present invention. Fig.12 In the example, the third injection signal of the triangular wave is first injected into the input end, and then the clock signal with a frequency of one-Mth of the reference clock signal is input into the demultiplexer, so that each sub-channel is sampled based on each clock signal with different phases. Taking the first reference sub-channel as sub-channel 0, the second reference sub-channel as sub-channel 2, and the sub-channel to be calibrated as sub-channel 1 as an example, the absolute value of the difference between the digital signal output by sub-channel 1 and the digital signal output by sub-channel 0 is calculated, and the absolute value of the difference between the digital signal output by sub-channel 2 and the digital signal output by sub-channel 1 is calculated, and the two absolute values ​​of each sampling point are cumulatively added and divided by the total number of sampling times N to obtain the timing mismatch value. . Then calculate the timing mismatch value With preset timing loop parameters The offset mismatch calibration estimate corresponding to the previous output interval is obtained by multiplying The sum of the values ​​is used as the current timing mismatch calibration estimate. . Then the current offset mismatch calibration estimate is The feedback is sent to the analog compensation circuit to perform timing mismatch compensation on the sub-channel to be calibrated.

[0155] Next, the process of performing timing mismatch calibration on the sub-channel to be timing calibrated according to the timing mismatch value of the sub-channel to be timing calibrated in the above steps is introduced.

[0156] Optionally, an alpha filtering process is performed on the timing mismatch value of the timing calibration sub-channel to obtain a timing mismatch calibration value, and timing calibration is performed on the timing calibration sub-channel based on the timing mismatch calibration value.

[0157] Specifically, after calculating the timing mismatch value of each sub-channel, the timing mismatch calibration value can be calculated based on the timing mismatch value of each sub-channel using the filtering method, and then the timing mismatch calibration value is fed back to the analog compensation circuit for timing mismatch compensation. In actual calculations, the timing mismatch calibration estimate is usually calculated to replace the timing mismatch calibration value. Exemplarily, the timing mismatch calibration estimate can be calculated based on alpha filtering by the following formula (12):

[0158] (12)

[0159] in, is the current timing mismatch calibration estimate, is the estimated value of the timing mismatch calibration corresponding to the previous output interval, To preset the timing loop parameters, is the timing mismatch value. It is worth noting that when sampling and calibration are performed for the first time, the current timing mismatch value calculation formula The value of is 0.

[0160] In this embodiment, the timing mismatch value of the timing calibration subchannel is subjected to alpha filtering to obtain a timing mismatch calibration value, and the timing calibration is performed on the timing calibration subchannel based on the timing mismatch calibration value, thereby ensuring the stability and accuracy of the timing calibration feedback.

[0161] An embodiment of the present invention further provides a time-interleaved analog-to-digital converter, which includes a plurality of sub-channels, and each sub-channel is calibrated by the above-mentioned time-interleaved analog-to-digital converter sub-channel mismatch calibration method.

[0162] An embodiment of the present invention also provides a serializer and deserializer system, which includes a transmitter, a channel equalizer, an automatic gain controller, a time-interleaved analog-to-digital converter, a digital signal processor and a deserializer, wherein the time-interleaved analog-to-digital converter is the above-mentioned time-interleaved analog-to-digital converter.

[0163] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for calibrating subchannel mismatch of a time-interleaved analog-to-digital converter, characterized in that: The method comprises: After each sub-channel of the time-interleaved analog-to-digital converter is respectively connected to a first injection signal, an offset mismatch value of each sub-channel is determined according to the first signal output by each sub-channel, and offset mismatch calibration is performed on each sub-channel according to the offset mismatch value of each sub-channel, wherein the first injection signal is a 0 volt DC signal; After each sub-channel completes the offset mismatch calibration and is respectively connected to the second injection signal, the gain mismatch value of each non-reference sub-channel is determined according to the second signal output by the reference sub-channel and the non-reference sub-channel, and the gain mismatch calibration is performed on each non-reference sub-channel according to the gain mismatch value of each non-reference sub-channel, wherein the second injection signal is a DC signal; After each non-reference sub-channel completes the gain offset mismatch calibration and each sub-channel is respectively connected to the third injection signal, the first reference sub-channel, the second reference sub-channel and the sub-channel to be timing calibrated are iteratively determined, and, according to the third signal output by the first reference sub-channel, the third signal output by the second reference sub-channel and the third signal output by the sub-channel to be timing calibrated, the timing mismatch calibration is performed on the sub-channel to be timing calibrated.

2. The method for calibrating sub-channel mismatch of a time-interleaved analog-to-digital converter according to claim 1, characterized in that: The step of determining the offset mismatch value of each sub-channel according to the first signal output by each sub-channel comprises: The expected value of each of the first signals is used as the offset mismatch value of each sub-channel.

3. The method for calibrating sub-channel mismatch of a time-interleaved analog-to-digital converter according to claim 1, characterized in that: The step of determining the gain mismatch value of each non-reference sub-channel according to the second signal output by the reference sub-channel and the non-reference sub-channel comprises: Determining a reference variance of a second signal output by the reference subchannel; Determine a plurality of first variances of the second signal output by each non-reference subchannel; The differences between the reference variance and each first variance are calculated respectively, and each difference is used as a gain mismatch value corresponding to each non-reference sub-channel.

4. The method for calibrating sub-channel mismatch of a time-interleaved analog-to-digital converter according to claim 1, characterized in that: The iterative determination of the first reference sub-channel, the second reference sub-channel, and the sub-channel to be time-calibrated, and the timing mismatch calibration of the sub-channel to be time-calibrated according to the third signal output by the first reference sub-channel, the third signal output by the second reference sub-channel, and the third signal output by the sub-channel to be time-calibrated, includes: In each iteration, a plurality of calibration groups are determined based on the binary method, each of the calibration groups respectively comprising: a first reference sub-channel, a second reference sub-channel and a sub-channel to be time-calibrated, and the first reference sub-channel and the second reference sub-channel have the same timing mismatch; In each calibration group, the timing mismatch value of the sub-channel to be calibrated is determined according to the third signal output by the first reference sub-channel, the third signal output by the second reference sub-channel and the third signal output by the sub-channel to be calibrated, and the timing mismatch calibration is performed on the sub-channel to be calibrated according to the timing mismatch value of the sub-channel to be calibrated.

5. The method for calibrating sub-channel mismatch of a time-interleaved analog-to-digital converter according to claim 4, characterized in that: The method of determining a plurality of calibration groups based on the dichotomy method comprises: Each sub-channel to be calibrated in the previous iteration is used as a reference calibration sub-channel; According to each benchmark calibration subchannel, the first reference subchannel, the second reference subchannel and the subchannel to be time-series calibrated in the multiple calibration groups corresponding to the benchmark calibration subchannel are determined, wherein the first reference subchannel in each calibration group is the benchmark calibration subchannel.

6. The method for calibrating sub-channel mismatch of a time-interleaved analog-to-digital converter according to claim 5, characterized in that: The step of determining, according to each reference calibration sub-channel, a first reference sub-channel, a second reference sub-channel, and a sub-channel to be time-series calibrated in a plurality of calibration groups corresponding to the reference calibration sub-channel comprises: Using the reference calibration subchannel as the first reference subchannel in a first calibration group and the first reference subchannel in a second calibration group corresponding to the reference calibration subchannel; Two subchannels satisfying preset conditions are used as the second reference subchannel in the first calibration group and the second reference subchannel in the second calibration group, respectively, wherein the preset conditions include: the channel interval between the subchannel and the reference calibration subchannel is half of the channel interval used in the previous iteration; The subchannel located between the first reference subchannel and the second reference subchannel in the first calibration group is used as the subchannel to be timing calibrated in the first calibration group, and the subchannel located between the first reference subchannel and the second reference subchannel in the second calibration group is used as the subchannel to be timing calibrated in the second calibration group.

7. The method for calibrating sub-channel mismatch of a time-interleaved analog-to-digital converter according to claim 4, characterized in that: Determining a timing mismatch value of the sub-channel to be time-calibrated according to a third signal output by the first reference sub-channel, a third signal output by the second reference sub-channel, and a third signal output by the sub-channel to be time-calibrated includes: Determining a first expected value according to the third signal of the sub-channel to be time-calibrated and the third signal of the first reference sub-channel; Determining a second expected value according to the third signal of the sub-channel to be time-calibrated and the third signal of the second reference sub-channel; According to the first expected value, the second expected value and the preset constraint condition, the timing mismatch value of the sub-channel to be timing calibrated is determined, the preset constraint condition indicates that the frequency of the third injection signal is equal to the ratio of the frequency of the reference clock signal to the preset value, and the lowest common divisor of the preset value and the number of sub-channels of the time-interleaved analog-to-digital converter is less than or equal to 2.

8. The method for calibrating sub-channel mismatch of a time-interleaved analog-to-digital converter according to claim 4, characterized in that: The step of performing timing mismatch calibration on the sub-channel to be calibrated according to the timing mismatch value of the sub-channel to be calibrated includes: An alpha filter process is performed on the timing mismatch value of the sub-channel to be timing-calibrated to obtain a timing mismatch calibration value, and timing calibration is performed on the sub-channel to be timing-calibrated based on the timing mismatch calibration value.

9. A time-interleaved analog-to-digital converter, characterized in that: The time-interleaved analog-to-digital converter includes a plurality of sub-channels, and each sub-channel is calibrated by the time-interleaved analog-to-digital converter sub-channel mismatch calibration method according to any one of claims 1-8.

10. A serializer and deserializer system, characterized in that: The serializer and deserializer system includes a transmitter, a channel equalizer, an automatic gain controller, a time-interleaved analog-to-digital converter, a digital signal processor and a deserializer, and the time-interleaved analog-to-digital converter is the time-interleaved analog-to-digital converter according to claim 9.

Citation Information

Patent Citations

  • Method and device for realizing self-calibration of time-interleaved analog-to-digital converter

    CN110048717A

  • Calibration method, calibration device, time-interleaved ADC, electronic equipment and readable medium

    CN113904683A