Multi-channel data alignment method, device and multi-channel time-interleaved analog-to-digital converter

By aligning the data of a multi-channel time-interleaved ADC using a multi-level, step-by-step synchronization method, the problem of data timing errors caused by phase shift was solved, achieving high-precision and stable data alignment and improving the overall stability and linearity of the system.

CN123122546APending Publication Date: 2026-11-17NO 24 RES INST OF CETC
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611420674.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-14
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In multi-channel time-interleaved ADCs, due to process deviations, voltage fluctuations, and temperature changes, there is an unavoidable phase shift between the accompanying clock and data output of each channel. Traditional single-stage synchronization or globally unified clock alignment schemes cannot effectively solve data timing errors under high channel number and high speed, resulting in a decrease in ADC linearity or even functional problems.

Method used

A multi-level, step-by-step synchronization method is adopted. By acquiring the data output from each channel and its accompanying clock, pairwise synchronization latching is performed using the accompanying clocks of paired channels, group synchronization is performed, and alignment of all channels is completed by synthesizing a clock. This includes at least three steps: pairwise pairing, grouping, and synthesizing a clock.

Benefits of technology

It effectively reduces the risk of inconsistent timing of multi-channel data, improves system stability and circuit linearity, avoids global clock transmission delay deviation, and ensures the accuracy of data alignment and the stability of the system under process changes and fluctuations in operating conditions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123122546A_ABST
    Figure CN123122546A_ABST
Patent Text Reader

Abstract

This invention discloses a multi-channel data alignment method, apparatus, and multi-channel time-interleaved analog-to-digital converter. The method includes the following steps: acquiring the data output from each channel of the multi-channel time-interleaved analog-to-digital converter and its corresponding accompanying clock, wherein the accompanying clock is output synchronously with the data of each channel, and its timing is consistent with the timing of the data of the corresponding channel; aligning the data of each channel based on the phase difference between the accompanying clocks of each channel using a multi-level successive synchronization method; wherein the multi-level successive synchronization includes at least two levels of synchronization operation, and at least the last level of synchronization operation is based on at least two accompanying clocks to generate a synthesized clock to complete the alignment. This invention, through multi-level successive synchronization processing, progressively compresses the timing deviation between channels, greatly reducing the risk of inconsistent timing of multi-channel data, and does not require the introduction of an additional clock module, resulting in a simple structural framework.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of integrated circuit technology, and in particular to a multi-channel data alignment method, apparatus, and multi-channel time-interleaved analog-to-digital converter. Background Technology

[0002] In the field of high-speed data acquisition and signal processing, analog-to-digital converters (ADCs) play a crucial role. To overcome the physical limits of single-channel ADC sampling rates, multi-channel time-interleaved architectures are widely used. However, with the ever-increasing demand for bandwidth, multi-channel time-interleaved ADCs are not only becoming faster in terms of single-channel speed, but also rapidly increasing in the number of channels. The data rate increases significantly with the increase in single-channel speed, causing the window for data timing alignment to narrow drastically.

[0003] Ideally, the data clock cycles of all channels should be perfectly synchronized to properly reconstruct interleaved data. However, in actual chip design and manufacturing, due to process variations, voltage fluctuations, and temperature changes, an unavoidable phase shift exists between the accompanying clock and data output of each channel. If a traditional single-stage synchronization or globally unified clock alignment scheme is used, even small phase errors will be amplified at high channel counts and high data rates. Once an error occurs that leads to data timing errors, it will cause a decrease in ADC linearity and even functional problems. Summary of the Invention

[0004] In view of the shortcomings of the prior art, the technical problem to be solved by the present invention is to provide a multi-channel data alignment method, device and multi-channel time-interleaved analog-to-digital converter that can effectively realize the synchronization of data in each channel.

[0005] To solve the above-mentioned technical problems, one technical solution adopted by the present invention is to provide a multi-channel data alignment method, comprising the following steps: Acquire the data output from each channel of the multi-channel time-interleaved analog-to-digital converter and its corresponding accompanying clock. The accompanying clock is output synchronously with the data of each channel, and its timing is consistent with the timing of the data of the corresponding channel. Based on the phase difference between the accompanying clocks of each channel, the data of each channel is aligned by a multi-level step-by-step synchronization method; wherein, the multi-level step-by-step synchronization includes at least two levels of synchronization operation, and at least the last level of synchronization operation is based on at least two accompanying clocks to generate a composite clock to complete the alignment.

[0006] Furthermore, the multi-level step-by-step synchronization includes: The first level of synchronization pairs each channel together and uses the accompanying clock of the paired channel to synchronously latch the data of each channel, forming multiple first-level data groups. The second-level synchronization involves grouping multiple first-level data groups into groups and using the same clock signal to synchronously latch each first-level data group within the same group, thus forming multiple second-level data groups. The third level of synchronization involves selecting at least two accompanying clocks from the accompanying clocks corresponding to each channel in a second-level data group, generating a composite clock based on the selected accompanying clocks, and using the composite clock to synchronously latch each second-level data group to achieve data alignment of all channels.

[0007] Furthermore, the first-level synchronization step includes the following sub-steps: Two channels whose sampling times differ by half a sampling period are paired together; The data of the two paired channels are first synchronized and latched using the accompanying clock of the channel with the later sampling time. Using the accompanying clock of the channel with the earlier sampling time in the two paired channels, the data after the first synchronous latch is synchronously latched a second time to form the aligned first-level data group.

[0008] Furthermore, the second-level synchronization step includes the following sub-steps: Multiple first-level data groups are grouped according to their time sequence, with each group containing the same number of first-level data groups; Select one accompanying clock from the original accompanying clocks corresponding to each channel in each group as the synchronization clock for that group; Each group's synchronization clock is used to synchronously latch each first-level data group within the group, aligning the timing of each first-level data group within the group to form multiple second-level data groups.

[0009] Furthermore, the effective edge of the synchronization clock is located within the time period when the data of each first-level data group in the same group is valid, and the effective edge has redundancy from the front boundary and the back boundary of the time period.

[0010] Furthermore, the third-level synchronization step includes the following sub-steps: Obtain the valid time period for each second-level data group; Select a second-level data group based on the effective data period of each second-level data group; Select at least two accompanying clocks from the original accompanying clocks corresponding to each channel in the selected second-level data group, perform logical operations on the selected accompanying clocks, and generate a synthesized clock. The synthesized clock is used to synchronously latch each of the second-level data groups to complete the data alignment of all channels.

[0011] Furthermore, the effective edge of the synthesized clock is located within the time period when the data of each second-level data group is valid, and the effective edge has redundancy from the front boundary and the back boundary of the time period.

[0012] Furthermore, the at least two accompanying clocks are limited to three accompanying clocks, and the logical operation is limited to an OR logical operation.

[0013] To solve the above-mentioned technical problems, another technical solution adopted by the present invention is: providing a multi-channel data alignment device for implementing the multi-channel data alignment method, comprising: The acquisition module is used to acquire the data output by each channel of the multi-channel time-interleaved analog-to-digital converter and its corresponding accompanying clock. The accompanying clock is output synchronously with the data of each channel, and its timing is consistent with the timing of the data of the corresponding channel. The alignment module is used to align the data of each channel based on the phase difference between the accompanying clocks of each channel, using a multi-level, step-by-step synchronization method.

[0014] To solve the above-mentioned technical problems, another technical solution adopted by the present invention is: to provide a multi-channel time-interleaved analog-to-digital converter, including multiple sub-channel analog-to-digital converters and the multi-channel data alignment device; the multi-channel data alignment device is used to receive the data output by the multiple sub-channel analog-to-digital converters and their corresponding accompanying clocks, and to align the data in a multi-level step-by-step synchronization manner.

[0015] The multi-channel data alignment method, apparatus, and multi-channel time-interleaved analog-to-digital converter of the present invention have at least the following beneficial effects: (i) Data is aligned by three-level synchronization. First, the synchronization range is narrowed by pairing two by two, then the synchronization is grouped, and finally the final alignment of all channels is completed by synthesizing the clock. The timing deviation between each channel is compressed step by step, which greatly reduces the risk of inconsistent timing of multi-channel data, increases the overall stability of the system, and improves the linearity of the circuit.

[0016] (ii) Using the accompanying clock as the synchronous clock source in each channel avoids the delay deviation caused by long-distance transmission of the global clock, ensures that the data of each channel has a fixed phase relationship with its corresponding clock, and improves the accuracy of data alignment.

[0017] (iii) In the third-level synchronization, a composite clock is generated by performing logical operations on at least two accompanying clocks without introducing additional clock modules. This saves on components and power consumption, and can actively generate new clock edges within the data validity window through logical operations when the existing clock edges cannot meet the alignment requirements, ensuring accurate alignment in high-speed multi-channel scenarios.

[0018] (iv) By configuring the effective edge of the synthesized clock to be within the period when all data in each data group is valid, and leaving sufficient redundancy at the start and end times of that period, the timing redundancy of the aligned clock is maximized, which significantly improves the stability and reliability of the system under the conditions of process angle changes and operating condition fluctuations. Attached Figure Description

[0019] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart of one embodiment of the multi-channel data alignment method of the present invention.

[0020] Figure 2 for Figure 1 Flowchart of step S200.

[0021] Figure 3 for Figure 2 Flowchart of step S210.

[0022] Figure 4 This is the timing diagram of the accompanying clocks for the paired channels in the first-level synchronization.

[0023] Figure 5 This is a timing waveform diagram for aligning data between two channels during the first-level synchronization.

[0024] Figure 6 This is a circuit schematic for aligning data between two channels in the first stage of synchronization.

[0025] Figure 7 for Figure 2 Flowchart of step S220.

[0026] Figure 8 This is a timing diagram of the valid data window and synchronization clock for each channel in group 0 of the second-level synchronization.

[0027] Figure 9 for Figure 2 Flowchart of step S230.

[0028] Figure 10 This is a timing diagram for synthesizing the three accompanying clocks into the aligned clock in the third-level synchronization.

[0029] Figure 11 This is a timing diagram for the final alignment of all channel data in the third-level synchronization.

[0030] Figure 12 This is a block diagram of one embodiment of the multi-channel data alignment device of the present invention. Detailed Implementation

[0031] The following disclosure provides various embodiments or examples for implementing different features of the present invention. Specific examples will be described below to simplify the invention. Of course, these are merely examples and are not intended to limit the invention.

[0032] Although the numerical ranges and parameter settings presented in this invention are approximations, the numerical settings in specific instances are reported as precisely as possible. Any numerical value inherently contains some necessary error arising from the standard deviation found in the respective test measurements. Similarly, as used herein, the term "about" generally refers to within 10%, 5%, 1%, or 0.5% of a given value or range. Alternatively, the term "about" means within an acceptable average standard error that can be conceived by one of ordinary skill in the art. Except in instances of operation / work, or unless expressly stated otherwise, all numerical ranges, totals, values, and percentages, such as those for material quantities, durations, temperatures, operating conditions, amounts, and other similarities disclosed herein, should be understood to be modified by the term "about" in all cases. Therefore, unless otherwise stated, the numerical parameter settings set forth in this invention and the appended claims are approximations that can be changed upon request. At a minimum, each numerical parameter should be interpreted based on the number of significant figures reported and the application of ordinary rounding techniques. A range herein may be expressed as from one endpoint to another or between two endpoints. All scopes disclosed herein include endpoints unless otherwise stated.

[0033] Furthermore, the technical parts described in this invention and the appended claims are primarily the improved technical parts of this invention, and do not limit the object protected by this invention to only having these technical parts. Other known essential components (structures and / or methods) and / or non-essential components of the object protected, besides the technical parts described in this invention and the appended claims, are not included in this invention and the appended claims because they do not fall within the scope of improvements of this invention; however, this does not mean that the object protected by this invention does not possess these known components.

[0034] The following explanation uses a 64-channel time-interleaved analog-to-digital converter as an example to illustrate this solution.

[0035] Please see Figure 1 This is a flowchart of an embodiment of the multi-channel data alignment method of the present invention. This embodiment includes the following steps: S100. Obtain the data output from each channel of the multi-channel time-interleaved analog-to-digital converter and its corresponding accompanying clock. The accompanying clock is output synchronously with the data of each channel, and its timing is consistent with the timing of the data of the corresponding channel.

[0036] Specifically, the multi-channel time-interleaved analog-to-digital converter includes multiple sub-channel ADCs. Each sub-channel ADC alternately samples the input analog signal and converts the sampling results into digital data output. In this embodiment, the 64-channel time-interleaved analog-to-digital converter includes 64 sub-channel ADCs, denoted as channel 0 to channel 63. The data output by each channel is an 11-bit digital signal, denoted as DOUT_x<10:0>, where x represents the channel number, x=0,1,...,63.

[0037] Each channel outputs a corresponding accompanying clock, denoted as CKAMP, along with its data output. The accompanying clock is output synchronously with the data of its corresponding channel; that is, the timing of the accompanying clock is consistent with the timing of the data of that channel. Specifically, the data signal and the accompanying clock signal experience the same transmission delay, and regardless of changes in process technology, voltage, and temperature, a fixed phase relationship is maintained between the data and its accompanying clock. The frequency of the accompanying clock is consistent with the total sampling rate of the multi-channel time-interleaved analog-to-digital converter; in this embodiment, the frequency of the accompanying clock is 1 GHz.

[0038] S200. Based on the phase difference between the accompanying clocks of each channel, the data of each channel is aligned using a multi-level successive synchronization method. The multi-level successive synchronization includes at least two levels of synchronization operations, and at least the last level of synchronization operation is based on at least two accompanying clocks to generate a synthesized clock to complete the alignment.

[0039] Since the 64 channels sample in turn in a time-interleaved architecture, the sampling times of adjacent channels differ by a fixed time interval, which is equal to the total sampling period T divided by the total number of channels. In this embodiment, the total sampling period T = 1000 ps (corresponding to a 1 GHz sampling rate), and the total number of channels is 64. Therefore, the sampling times of adjacent channels differ by T / 64 = 15.625 ps. Consequently, there is also a corresponding phase difference between the accompanying clocks of each channel. This invention utilizes the phase difference between the accompanying clocks of each channel to align the data of the 64 channels step by step through a multi-level, step-by-step synchronization method.

[0040] Please see Figure 2 This step S200 includes the following sub-steps: S210, Level 1 Synchronization: Each channel is paired up, and the data of each channel is synchronously latched using the accompanying clock of the paired channel, forming multiple Level 1 data groups. For details, please refer to [link to relevant documentation]. Figure 3 This step S210 includes the following sub-steps: S211. Pair two channels whose sampling times differ by half a sampling period into one group.

[0041] In this embodiment, the sampling period T = 1000 ps, ​​and half a sampling period T / 2 = 500 ps. Two channels with a sampling time difference of 500 ps are paired together. For example, channel 0 has a sampling time of 0 ps, ​​and channel 32 has a sampling time of 32 × 15.625 ps = 500 ps. Since they differ by T / 2, channels 0 and 32 are paired together. This process continues, resulting in 32 pairing groups for the 64 channels. Examples of partial pairings include channel 8 (sampling time 125 ps) and channel 40 (sampling time 625 ps), channel 16 (sampling time 250 ps) and channel 48 (sampling time 750 ps), and channel 24 (sampling time 375 ps) and channel 56 (sampling time 875 ps). Please refer to [link to relevant documentation]. Figure 4 The following are the timing diagrams for the accompanying clocks of channels 0, 8, 16, 24, 32, 40, 48, and 56, respectively, corresponding to the accompanying clock CKAMP_0. <0> CKAMP_8 <0> CKAMP_0 <1> CKAMP_8 <1> CKAMP_0 <2> CKAMP_8 <2> CKAMP_0 <3> CKAMP_8 <3> For example, CKAMP_0 <0> With CKAMP_0 <2> The difference is 500ps, which is T / 2, CKAMP_8 <0> With CKAMP_8 <2> They are also 500ps apart, or T / 2. The phase relationship of these two pairs of clocks corresponds to the pairing relationship between channel 0 and channel 32, and channel 8 and channel 40, respectively.

[0042] S212. Using the accompanying clock of the channel with the later sampling time in the two paired channels, perform the first synchronous latching of the data of the two channels.

[0043] Please see Figure 5 Taking channels 0 and 32 as examples, channel 0 has an earlier sampling time of 0 ps, ​​while channel 32 has a later sampling time of 500 ps. The accompanying clock for channel 32 is denoted as CKAMP_0. <2> Its rising edge occurs at 500ps, 1500ps, 2500ps, etc. In CKAMP_0 <2> At the first rising edge (500ps), DOUT_0<10:0> and DOUT_32<10:0> are latched to D0<10:0> and D32<10:0> respectively, at which point the data of channel 0 and channel 32 are time-aligned.

[0044] S213. Using the accompanying clock of the channel with the earlier sampling time in the two paired channels, perform a second synchronous latch on the data after the first synchronous latch to form an aligned first-level data group.

[0045] Please continue reading. Figure 5The accompanying clock for channel 0 is denoted as CKAMP_0. <0> Its rising edge occurs at 0ps, 1000ps, 2000ps, etc. After the first latch in step S212, the data in channel 0 and channel 32 are latched at 500ps. At 1000ps, CKAMP_0 <0> Upon the rising edge of the signal, D0<10:0> and D32<10:0> are further latched to DO_0<10:0> and DO_32<10:0>, completing the transfer to CKAMP_0. <0> The unification of the clock domain completes the first level of synchronization.

[0046] Please see Figure 6 The raw data DOUT_0<10:0> of channel 0 is input to the data terminal of the first-stage flip-flop, and the clock input of the first-stage flip-flop is connected to the accompanying clock CKAMP_0 of channel 32. <2> Its output forms intermediate data D0<10:0>; this intermediate data D0<10:0> is then connected to the data terminal of the second-stage flip-flop, and the clock input of the second-stage flip-flop is connected to the accompanying clock CKAMP_0 of channel 0. <0> The final output is aligned data DO_0<10:0>. The original data DOUT_32<10:0> from channel 32 is first processed by a first-level buffer to form data D32<10:0>; data D32<10:0> is then connected to the data terminal of the third-stage flip-flop, and the clock input of the third-stage flip-flop is also connected to the accompanying clock CKAMP_0 of channel 0. <0> The final output is the aligned data DO_32<10:0>. This achieves the first stage of synchronization between two channels with a time difference of T / 2.

[0047] Similarly, for channels 8 and 40, the accompanying clock CKAMP_8 of channel 40 is used. <2> Perform the first synchronous latch, and then use the accompanying clock CKAMP_8 of channel 8. <0> A second synchronization latch is performed to form another first-level data group. For channels 16 and 48, the accompanying clock CKAMP_0 of channel 48 is used. <3> Perform the first synchronization latch, and then use the accompanying clock CKAMP_0 of channel 16. <1> Perform a second synchronization latch. For channels 24 and 56, use the accompanying clock CKAMP_8 of channel 56. <3> Perform the first synchronization latch, and then use the accompanying clock CKAMP_8 of channel 24. <1> A second synchronization latch is performed. In this way, the 64 channels are paired up and synchronized twice, forming 32 first-level data groups. The data of the two channels within each first-level data group is now fully aligned.

[0048] S220 second-level synchronization groups multiple first-level data groups and uses the same clock signal to synchronously latch each first-level data group within the same group, forming multiple second-level data groups.

[0049] Please see Figure 7 This step S220 includes the following sub-steps: S221. Group the multiple first-level data groups according to their time sequence, with each group containing the same number of first-level data groups.

[0050] Specifically, the 32 first-level data groups are divided into 8 groups according to their temporal sequence. Each group contains 4 first-level data groups, i.e., 8 raw channels. For example, group 0 contains the first-level data groups: channels 0 and 32, channels 8 and 40, channels 16 and 48, and channels 24 and 56, corresponding to channel numbers 0, 8, 16, 24, 32, 40, 48, and 56. Similarly, group 1 contains channels 1, 9, 17, 25, 33, 41, 49, and 57; ...; group 7 contains channels 7, 15, 23, 31, 39, 47, 55, and 63.

[0051] S222. Select one accompanying clock from the original accompanying clocks corresponding to each channel in each group as the synchronization clock for that group.

[0052] For details, please refer to Figure 8 Taking group 0 as an example, this group contains channels 0, 8, 16, 24, 32, 40, 48, and 56. Each of these channels has a corresponding original accompanying clock; for example, channel 0 corresponds to CKAMP_0. <0> Channel 32 corresponds to CKAMP_0 <2> Channel 8 corresponds to CKAMP_8 <0> Channel 40 corresponds to CKAMP_8 <2> Channel 16 corresponds to CKAMP_0 <1> Channel 48 corresponds to CKAMP_0 <3> Channel 24 corresponds to CKAMP_8 <1> Channel 56 corresponds to CKAMP_8 <3> One of these original accompanying clocks is selected as the synchronization clock for this group. The upper part of the figure shows the waveforms of the accompanying clocks CKAMP_0<0:3> to CKAMP_8<0:3> for the above 8 channels, and the lower part shows the valid windows of the data DO_0, DO_8, DO_16, DO_24, DO_32, DO_40, DO_48, and DO_56 for the 8 channels, labeled as T0, T8, T16, T24, T32, T40, T48, and T56, respectively.

[0053] The effective edge of the synchronization clock falls within the time period during which all first-level data groups within the same group are valid, and redundancy is maintained between the effective edge and both the front and back boundaries of the time period. Specifically, the front boundary of the time period during which all data are valid refers to the start time of the latest arriving first-level data group, and the back boundary refers to the end time of the earliest arriving first-level data group. The effective edge of the synchronization clock falls within this time period, and sufficient redundancy is maintained to both the front and back boundaries, thereby maximizing timing redundancy and ensuring reliable latching under various process angles and operating conditions. In this embodiment, CKAMP_8 is selected. <1> As the synchronization clock of Group 0, its effective edge is the falling edge, which falls within the period when all four first-level data groups in this group are valid. It is 375ps away from the end of the first data group and 250ps away from the beginning of the fourth data group, with the maximum time redundancy.

[0054] S223. Use the synchronization clock of each group to synchronously latch each first-level data group in the group, so that the first-level data groups in the group are aligned in time to form multiple second-level data groups.

[0055] Taking group 0 as an example, when selecting CKAMP_8 <1> After being used as a synchronization clock, CKAMP_8 <1> On the falling edge of the signal, the four first-level data groups (channels 0 and 32, channels 8 and 40, channels 16 and 48, and channels 24 and 56) in group 0 are simultaneously latched synchronously. After this latching, the eight channel data in group 0 are fully aligned in timing, forming a second-level data group, denoted as DO_0,8,16,24,32,40,48,56<10:0>. Similarly, the remaining seven groups are also aligned within their respective groups, forming eight second-level data groups.

[0056] The S230 third-level synchronization selects at least two accompanying clocks from the accompanying clocks corresponding to each channel in a second-level data group, generates a composite clock based on the selected accompanying clocks, and uses the composite clock to synchronously latch each second-level data group, achieving data alignment across all channels. Please refer to [link to relevant documentation]. Figure 9 This step S230 includes the following sub-steps: S231. Obtain the valid time period of each second-level data group.

[0057] Specifically, after the second-level synchronization, eight second-level data groups are obtained, denoted as DO_0,8,16,24,32,40,48,56<10:0>, DO_1,9,17,25,33,41,49,57<10:0>, DO_2,10,18,26,34,42,50,58<10:0>, and DO_3,11,19,27,35,43. The timing of these eight second-level data groups is synchronized with the accompanying clock CKAMP_0. (DO_4,12,20,28,36,44,52,60,DO_5,13,21,29,37,45,53,61,DO_6,14,22,30,38,46,54,62,DO_7,15,23,31,39,47,55,63) <2> CKAMP_1 <2> CKAMP_2 <2> CKAMP_3 <2> CKAMP_4 <2> CKAMP_5 <2> CKAMP_6 <2> CKAMP_7 <2> The timing is consistent, with a maximum timing difference of 125ps.

[0058] S232. Select a second-level data group based on the effective data period of each second-level data group.

[0059] Specifically, the distribution of the effective data periods of each of the eight second-level data groups on the time axis is obtained through time series simulation. Using the central region of the overall time series window of the eight second-level data groups as a reference, the second-level data group whose effective data periods are closest to this central region is selected. In this embodiment, the third second-level data group is selected.

[0060] S233. Select at least two accompanying clocks from the original accompanying clocks corresponding to each channel in the selected second-level data group, perform logical operations on the selected accompanying clocks, and generate a synthesized clock.

[0061] Specifically, three accompanying clocks are selected from the original accompanying clocks corresponding to each channel in the selected second-level data group. These original accompanying clocks refer to the accompanying clocks before the first-level synchronization, thus achieving high-reliability alignment of the 64-channel data without introducing an additional clock module. These three accompanying clocks are input into an OR gate for OR logic operation to obtain a synthesized clock. The logic rule of the OR gate is: as long as one input is high, the output is high; only when all inputs are low, the output is low. The duty cycle of the synthesized clock is greater than the duty cycle of each of the selected accompanying clocks. In this embodiment, the duty cycle of each of the three accompanying clocks is 25%, and the duty cycle of the resulting synthesized clock is 50%.

[0062] Please see Figure 10The three accompanying clocks selected are CKAMP_2 <1> CKAMP_10 <1> CKAMP_2 <2> The rising edges of the three accompanying clocks are staggered sequentially, and the high-level intervals of adjacent clocks partially overlap. The high-level interval of the OR gate output is the union of the high-level intervals of the three clocks, forming a continuous high-level interval. Its starting boundary is the rising edge of the synthesized clock CK_OUT, and its ending boundary is the falling edge of the synthesized clock CK_OUT. The high-level interval of CK_OUT begins at CKAMP_2. <1> The rising edge terminates at CKAMP_2. <2> The falling edge of the clock forms a continuous high-level pulse. By selecting the phase relationship of the three accompanying clocks, the effective edge of the synthesized clock can be made to fall within the time period when the data of each second-level data group is valid, and redundancy is reserved at both the front and back boundaries of this time period, thereby maximizing timing redundancy. Here, the front boundary refers to the start time of the data of the latest second-level data group, and the back boundary refers to the end time of the data of the earliest second-level data group.

[0063] It should be noted that if only two accompanying clocks are selected for OR logic operations, although high-level splicing can be achieved under ideal fixed phase conditions, in the actual operation of high-speed multi-channel time-interleaved analog-to-digital converters, the high-level intervals of the two clocks are easily misaligned due to process, voltage, and temperature fluctuations. This results in separated double high-level pulses and low-level gaps at the OR gate output. These gaps cause the synthesized clock to generate multiple edges within one cycle, leading to uncertainty in the subsequent latch timing. Therefore, this embodiment selects three accompanying clocks for OR logic synthesis. By presetting the phase difference between adjacent clocks, their high-level intervals partially overlap, and the union of the three clocks can form a continuous high-level interval without low-level gaps. The resulting aligned clock has a unique and reliable valid edge, such as a falling edge, and this edge is located at the timing redundancy center of the multi-channel data valid window, significantly improving latch reliability.

[0064] S234. Use the synthesized clock to synchronously latch each of the second-level data groups to complete the data alignment of all channels.

[0065] Please see Figure 11This synthesized clock is used to simultaneously latch eight secondary data groups. The accompanying clock CKAMP_0 corresponds to each of the eight secondary data groups. <2> To CKAMP_7 <2> All clocks operate at 1 GHz with adjacent clock edges spaced 125 ps apart. Eight data streams, DO_0,8,16,...<10:0> to DO_7,15,23,...<10:0>, have valid windows labeled T0, T1, T8, T40, T16, T48, T24, and T56, respectively. The valid edge of the synthesized clock CK_OUT is located in the center of the time period when all eight data streams are valid. This time period extends from the start time of the latest arriving second-level data stream (approximately 125 ps) to the end time of the earliest arriving second-level data stream (approximately 1000 ps), with a width of approximately 875 ps, which represents the final data alignment redundancy. Sufficient redundancy is maintained from the valid edge of the synthesized clock to the front and back boundaries of this time period, ensuring that all eight second-level data streams are correctly latched. Ultimately, all 64 channels are aligned, achieving a data alignment redundancy of 875 ps.

[0066] Please see Figure 12 This is a structural block diagram of an embodiment of the multi-channel data alignment device of the present invention. The multi-channel data alignment device of this embodiment is used to implement the multi-channel data alignment method described in the above embodiment. Specifically, the multi-channel data alignment device of this embodiment includes an acquisition module 100 and an alignment module 200. Wherein: The acquisition module 100 is used to acquire the data output by each channel of the multi-channel time-interleaved analog-to-digital converter and its corresponding accompanying clock. The accompanying clock is output synchronously with the data of each channel, and its timing is consistent with the timing of the data of the corresponding channel.

[0067] The alignment module 200 is used to align the data of each channel based on the phase difference between the accompanying clocks of each channel, employing a multi-level, step-by-step synchronization method. The alignment module 200 includes a first-level alignment submodule 210, a second-level alignment submodule 220, and a third-level alignment submodule 230. Wherein: The first-level alignment submodule 210 is used to pair each channel in pairs and synchronously latch the data of each channel using the accompanying clock of the paired channel to form multiple first-level data groups. Specifically, the first-level alignment submodule 210 first pairs two channels whose sampling times differ by half a sampling period into one group; then, it uses the accompanying clock of the channel with the later sampling time in the paired two channels to perform the first synchronous latching of the data of the two channels; finally, it uses the accompanying clock of the channel with the earlier sampling time in the paired two channels to perform the second synchronous latching of the data after the first synchronous latching, forming aligned first-level data groups.

[0068] The second-level alignment submodule 220 is used to group multiple first-level data groups and synchronously latch each first-level data group within the same group using the same clock signal to form multiple second-level data groups. Specifically, the second-level alignment submodule 220 first groups the multiple first-level data groups according to their timing sequence, with each group containing the same number of first-level data groups; then, it selects one accompanying clock from the original accompanying clocks corresponding to each channel in each group as the synchronization clock for that group; finally, it uses the synchronization clock of each group to synchronously latch each first-level data group within the group, so that the first-level data groups within the group are aligned in timing to form multiple second-level data groups.

[0069] The third-level alignment submodule 230 is used to select at least two accompanying clocks from the accompanying clocks corresponding to each channel in a second-level data group, generate a composite clock based on the selected accompanying clocks, and use the composite clock to synchronously latch each second-level data group to achieve data alignment of all channels. Specifically, the third-level alignment submodule 230 first obtains the valid data time period of each second-level data group; then, based on the valid data time period of each second-level data group, selects a second-level data group; then, selects at least two accompanying clocks from the original accompanying clocks corresponding to each channel in the selected second-level data group, performs logical operations on the selected accompanying clocks to generate a composite clock; finally, uses the composite clock to synchronously latch each second-level data group to complete the data alignment of all channels.

[0070] This invention also discloses a multi-channel time-interleaved analog-to-digital converter (ADC), comprising multiple sub-channel ADCs and the aforementioned multi-channel data alignment device. The multi-channel data alignment device receives the data output from the multiple sub-channel ADCs and their corresponding accompanying clocks, and aligns the data using the aforementioned multi-channel data alignment method.

[0071] This invention aligns data through a three-stage, progressive synchronization process. First, it narrows the synchronization range by pairing data in pairs. Then, it performs group synchronization. Finally, it completes the final alignment of all channels using a synthesized clock. This progressively reduces timing discrepancies between channels, significantly minimizing the risk of misalignment in multi-channel data, increasing overall system stability, and improving circuit linearity. This invention uses accompanying clocks as synchronization clock sources in each channel, avoiding delays caused by long-distance transmission of the global clock. This ensures a fixed phase relationship between the data in each channel and its corresponding clock, improving data alignment accuracy. In the third-stage synchronization, this invention generates a synthesized clock by performing logical operations on at least two accompanying clocks. This avoids introducing additional clock modules, saving components and power consumption. Furthermore, even when existing clock edges cannot meet alignment requirements, it actively generates new clock edges within the valid data window through logical operations, ensuring accurate alignment even in high-speed, multi-channel scenarios. This invention maximizes the timing redundancy of the aligned clock by configuring the effective edge of the synthesized clock to fall within the period when all data groups are valid, and by providing sufficient redundancy at both the start and end times of this period. This significantly improves the stability and reliability of the system under conditions of process angle variations and fluctuating operating conditions. Experimental verification shows that the scheme of this invention can achieve data alignment for a 64-channel analog-to-digital converter, with an alignment redundancy of up to 875 ps.

Claims

1. A method of aligning multi-lane data, the method comprising: Includes the following steps: Acquire the data output from each channel of the multi-channel time-interleaved analog-to-digital converter and its corresponding accompanying clock. The accompanying clock is output synchronously with the data of each channel, and its timing is consistent with the timing of the data of the corresponding channel. Based on the phase difference between the accompanying clocks of each channel, the data of each channel is aligned by a multi-level step-by-step synchronization method; wherein, the multi-level step-by-step synchronization includes at least two levels of synchronization operation, and at least the last level of synchronization operation is based on at least two accompanying clocks to generate a composite clock to complete the alignment.

2. The multi-channel data alignment method as described in claim 1, characterized in that, The multi-level, step-by-step synchronization includes: The first level of synchronization pairs each channel together and uses the accompanying clock of the paired channel to synchronously latch the data of each channel, forming multiple first-level data groups. The second-level synchronization involves grouping multiple first-level data groups into groups and using the same clock signal to synchronously latch each first-level data group within the same group, thus forming multiple second-level data groups. The third level of synchronization involves selecting at least two accompanying clocks from the accompanying clocks corresponding to each channel in a second-level data group, generating a composite clock based on the selected accompanying clocks, and using the composite clock to synchronously latch each second-level data group to achieve data alignment of all channels.

3. The multi-channel data alignment method as described in claim 2, characterized in that, The first-level synchronization steps include the following sub-steps: Two channels whose sampling times differ by half a sampling period are paired together; The data of the two paired channels are first synchronized and latched using the accompanying clock of the channel with the later sampling time. Using the accompanying clock of the channel with the earlier sampling time in the two paired channels, the data after the first synchronous latch is synchronously latched a second time to form the aligned first-level data group.

4. The multi-channel data alignment method as described in claim 2, characterized in that, The second-level synchronization step includes the following sub-steps: Multiple first-level data groups are grouped according to their time sequence, with each group containing the same number of first-level data groups; Select one accompanying clock from the original accompanying clocks corresponding to each channel in each group as the synchronization clock for that group; Each group's synchronization clock is used to synchronously latch each first-level data group within the group, aligning the timing of each first-level data group within the group to form multiple second-level data groups.

5. The multi-channel data alignment method as described in claim 4, characterized in that: The effective edge of the synchronization clock is located within the time period when the data of each first-level data group in the same group is valid, and the effective edge has redundancy from the front boundary and the back boundary of the time period.

6. The multi-channel data alignment method as described in claim 2, characterized in that, The third-level synchronization step includes the following sub-steps: Obtain the valid time period for each second-level data group; Select a second-level data group based on the effective data period of each second-level data group; Select at least two accompanying clocks from the original accompanying clocks corresponding to each channel in the selected second-level data group, perform logical operations on the selected accompanying clocks, and generate a synthesized clock. The synthesized clock is used to synchronously latch each of the second-level data groups to complete the data alignment of all channels.

7. The multi-channel data alignment method as described in claim 6, characterized in that: The effective edge of the synthesized clock is located within the time period when all data in each of the second-level data groups is valid, and the effective edge has redundancy from the front and back boundaries of the time period.

8. The multi-channel data alignment method as described in claim 6, characterized in that: The at least two accompanying clocks are limited to three accompanying clocks, and the logical operation is limited to an OR logical operation.

9. A multi-channel data alignment device, characterized in that, A method for implementing the multi-channel data alignment method as described in any one of claims 1 to 8 includes: The acquisition module is used to acquire the data output by each channel of the multi-channel time-interleaved analog-to-digital converter and its corresponding accompanying clock. The accompanying clock is output synchronously with the data of each channel, and its timing is consistent with the timing of the data of the corresponding channel. The alignment module is used to align the data of each channel based on the phase difference between the accompanying clocks of each channel, using a multi-level, step-by-step synchronization method.

10. A multi-channel time-interleaved analog-to-digital converter, characterized in that, It includes multiple sub-channel analog-to-digital converters and a multi-channel data alignment device as described in claim 9; the multi-channel data alignment device is used to receive the data output by the multiple sub-channel analog-to-digital converters and their corresponding accompanying clocks, and to align the data using a multi-level, step-by-step synchronization method.