Correction device for a multichannel time-interleaved adc system

Abstract

A correction device (100) for a multi-channel time-interleaved ADC system (300) is disclosed. The correction device (100) is configured to correct amplitude response mismatch errors of an ADC output sequence v using a first time-varying function B and further to correct phase response mismatch errors of the ADC output sequence using a second time-varying function C. Thus, due to the use of the first time-varying function B and the second time-varying function C, low complexity converter performance is improved.

Description

TECHNICAL FIELD

[0001] Embodiments of the present invention relate to a correction device for a multi-channel time-interleaved ADC system. Furthermore, embodiments of the present invention also relate to a corresponding method and a multi-channel time-interleaved ADC system comprising such a correction device. BACKGROUND

[0002] Time-interleaving of multiple parallel analogue to digital converters (ADCs) is a technique to increase the effective sampling rate of an overall digital converter. By using Channel time-interleaved ADCs (TIADCs), coefficients The effective sampling rate is increased. Unfortunately, the effective resolution (i.e. the number of bits per sample) of the individual channel converters is not preserved throughout the converter due to channel mismatch errors. It is therefore necessary to compensate for these errors to recover the resolution. The errors can be broadly classified into linear mismatch errors and non-linear mismatch errors.

[0003] At a certain resolution, it can be assumed that the channel frequency responses have a frequency independent amplitude and phase delay response, which corresponds to static gain and linear phase (time skew) mismatch errors. In the absence of gain errors, this case corresponds to non-uniform sampling, and the problem is to recover a uniform sampling sequence from a non-uniform sampling sequence. However, to achieve very high resolutions at high conversion speeds, it is necessary to extend the channel model to a general frequency response, which has a frequency dependent amplitude and phase delay response. In this case, these general frequency response mismatch errors have to be compensated. One special case is referred to as bandwidth mismatch, where the channel amplitude response is frequency dependent, while the phase delay is frequency independent. It should also be mentioned that the calibration of a time-interleaved ADC requires the estimation and correction / compensation of channel frequency response mismatches. SUMMARY

[0004] It is an object of embodiments of the present invention to provide a solution which reduces or solves the drawbacks and problems of the conventional solutions.

[0005] It is a further object of embodiments of the present invention to provide a solution which is cost-effective and high-performing compared to the conventional solutions.

[0006] The above and other objects are achieved by the subject-matter claimed by the independent claims. Further advantageous embodiments of the present invention are provided in the dependent claims.

[0007] According to a first aspect of the present application, the above mentioned and other objects are achieved by a correction device for a multi-channel time-interleaved analogue-to-digital converter (ADC) system, the correction device comprising:

[0008] an input for receiving a sequence of ADC outputs, ;

[0009] a correction block coupled to the input and configured to generate a first corrected output sequence of the sequence of ADC outputs by :

[0010] correcting amplitude response mismatch errors of the sequence of ADC outputs by a first time-varying function, correcting phase response mismatch errors of the sequence of ADC outputs by a second time-varying function,

[0011] an output coupled to the correction block and configured to output the corrected output sequence.

[0012] An advantage of the correction device according to the first aspect is that, due to the use of the first time-varying function

[0013] and the second time-varying function , the correction device disclosed herein is able to achieve a lower implementation cost for correction and estimation in a multi-channel time-interleaved ADC system compared to conventional solutions. In an implementation form of the correction device according to the first aspect, the first time-varying function

[0014] and the second time-varying function are real-valued functions.

[0015] An advantage of this implementation form is that it is able to achieve a lower implementation cost compared to the case with complex-valued functions.

[0016] In an implementation form of the correction device according to the first aspect, the first time-varying function and the second time-varying function

[0017] An advantage of this implementation form is that it is able to achieve a simpler estimation and design procedure when estimating the channel frequency response mismatches and designing the corresponding digital time-varying functions.

[0018] According to the first aspect, in an implementation form of the correction device,

[0019] ​​​​​​​a first time-varying function is a first finite impulse response (FIR) filter with symmetric coefficients,

[0020] a second time-varying function multiplied by the imaginary unit is a second FIR filter with anti-symmetric coefficients.

[0021] An advantage of this implementation is that the implementation complexity can be reduced, since only half of the multipliers in the implementation are needed for the symmetric filter and the anti-symmetric filter compared to a general asymmetric filter.

[0022] In an implementation of the correction device according to the first aspect, the first FIR filter and the second FIR filter are linear phase FIR filters.

[0023] An advantage of this implementation is that the implementation complexity can be reduced, since only half of the multipliers in the implementation are needed for the symmetric filter and the anti-symmetric filter compared to a general asymmetric filter.

[0024] According to the first aspect, in an implementation of the correction device,

[0025] The correction block is a first correction block, and the correction device comprises a second correction block coupled to the first correction block and to the output, respectively,

[0026] The second correction block is configured to generate a second correction output sequence by :

[0027] correcting an amplitude response mismatch error of the first correction output sequence by the first time-varying function ,

[0028] correcting a phase response mismatch error of the first correction output sequence by the second time-varying function .

[0029] An advantage of this implementation is that the correction is improved for each correction stage.

[0030] In an implementation of the correction device according to the first aspect, the correction device comprises a set of correction blocks, wherein is an integer greater than or equal to 2, wherein the set of correction blocks comprises a first correction block and a second correction block, and wherein each correction block in the set of correction blocks is configured to implement an error estimation function

[0031] ,

[0032] wherein denotes the angular frequency, denotes the time index, denotes the sampling period, denotes the summation index, denotes the index of the respective correction block, denotes the total number of correction blocks of the correction device, denotes the imaginary unit, is the power of

[0033] In an implementation form of the correction device according to the first aspect, the correction device comprises a set of correction blocks, wherein is an integer greater than or equal to 2, wherein the set of correction blocks comprises a first correction block and a second correction block, and wherein

[0034] the first correction block is configured to implement a first error correction function

[0035] ,

[0036] the -th correction block of the set of correction blocks is configured to implement a -th error correction function

[0037] wherein ,

[0038] wherein denotes the angular frequency, denotes the time index, denotes the sampling period, denotes the index of the respective correction block, denotes the summation index, denotes the total number of correction blocks of the correction device, denotes the imaginary unit, is the power of

[0039] This implementation form has the advantage of reducing complexity.

[0040] According to the first aspect, in an implementation form of the correction device,

[0041] the first time-varying function is polynomial of order ,

[0042] Second time-varying function for polynomial of order ,

[0043] in, Represents angular frequency. Indicates a time index. Indicates the summation index.

[0044] The advantage of this implementation method is that it is low-cost based on a polynomial structure.

[0045] According to the first aspect, in one implementation of the calibration device,

[0046] Second time-varying function for ,

[0047] in, Indicates a time index. Indicates phase shift, This represents a negative value for the phase delay of multiple channels.

[0048] The advantage of this implementation is reduced complexity, because it allows for the implementation of a second time-varying function. Public parts, for example, like This part is a fixed common part of all channels, so only one fixed instance is needed in the implementation.

[0049] According to a second aspect of the invention, the above and other objectives are achieved by a multi-channel time-interleaved ADC system comprising one or more correction devices according to any implementation of the correction device of the first aspect.

[0050] According to a third aspect of the invention, the above and other objectives are achieved by a method for calibrating a device, the method comprising:

[0051] Receive time-interleaved ADC output sequence ;

[0052] Generate the ADC output sequence using the following steps. First corrected output sequence :

[0053] Through the first time-varying function Correcting the ADC output sequence The amplitude response mismatch error,

[0054] Through the second time-varying function correcting the phase response mismatch error of the ADC output sequence;

[0055] outputting the corrected ADC output sequence .

[0056] The method according to the third aspect can be extended to implementation manners corresponding to implementation manners of the correction device according to the first aspect. Thus, implementation manners of the method comprise one or more features of the corresponding implementation manners of the correction device.

[0057] The advantages of the method according to the third aspect are the same as the advantages of the corresponding implementation manners of the correction device according to the first aspect.

[0058] The present application also relates to a computer program which, when the program code is executed by at least one processor, causes the at least one processor to carry out any method according to embodiments of the present application. Furthermore, the present application also relates to a computer program product comprising a computer readable medium and the above-mentioned computer program, wherein the computer program is included in the computer readable medium, and comprises one or more of the following group: a read-only memory (ROM), a programmable ROM (PROM), an erasable PROM (EPROM), a flash memory, an electrically EPROM (EEPROM), and a hard disk drive.

[0059] Further applications and advantages of embodiments of the present application will be apparent from the following detailed description. BRIEF DESCRIPTION OF DRAWINGS

[0060] The accompanying drawings are intended to illustrate and explain different embodiments of the present application.

[0061] Figure 1 A multi-channel time-interleaved ADC system without correction is shown.

[0062] Figure 2a and Figure 2b Periodic non-uniform sampling and reconstruction problem is shown.

[0063] Figure 3 A correction device according to an embodiment of the present application is shown.

[0064] Figure 4 A flowchart of a method according to an embodiment of the present application is shown.

[0065] Figure 5 A correction device according to another embodiment of the present application is shown.

[0066] ​​Figure 6 A correction device according to yet another embodiment of the application is shown.

[0067] Figure 7 A first order correction device according to yet another embodiment of the application is shown.

[0068] Figure 8 An exemplary implementation of a system according to an embodiment of the application is shown. DETAILED DESCRIPTION

[0069] Reference Figure 1 The starting point is that a continuous-time (analog) signal is band-limited to This means that the Nyquist criterion for uniform sampling without aliasing is met with a sampling frequency of That is, uniform sampling according to does not introduce aliasing and the continuous-time signal can be recovered from .

[0070] With the inverse Fourier transform, the required uniform sampling sequence can be expressed as:

[0071] (1)

[0072] In a channel time-interleaved ADC without correction , the required sequence is not obtained, but another sequence is obtained by interleaving of the sequence, namely . As shown in Fig. 1, where a linear model of a Figure 1 channel TIADC with different channel frequency responses ( ) is shown, as well as a periodic time-varying correction filter system with impulse response .

[0073] Considering only linear channel mismatches, the sequence can be expressed by the inverse Fourier transform as:

[0074] (2)

[0075] where is the periodic time-varying system frequency response, and denotes the ADC channel frequency response. For example, in the special case of frequency-independent mismatches with a static gain constant and a static time skew ​in percentage of the sampling period , so the channel frequency response is modeled as . For , this corresponds to ( ), so to periodic non-uniform sampling and reconstruction problem.

[0076] Figure 2a and Figure 2b show the case when , where Fig. 2(a) shows uniform sampling, and Fig. 2(b) shows double-periodic non-uniform sampling corresponding to a two-channel TIADC with static time-skew error and without additional mismatch error. It is also noted that the above model also applies to the more general case, and thus to a single-channel ADC with time-varying frequency response, when .

[0077] In addition to channel mismatch, there is also DC offset mismatch, but these are more easily compensated by measuring and subtracting the sample mean in each channel.

[0078] Due to channel mismatch and DC offset mismatch, the sequence includes distortion that needs to be compensated to reach the resolution of a single-channel ADC.

[0079] Given the sequence , the digital correction / compensation (reconstruction) amounts to forming a new sequence that should be as close as possible to the desired sequence . Regardless of implementation, the correction can be represented and designed with a time-varying discrete-time compensation filter or system with periodic impulse response (e.g., ). In practice, a reference channel can be assumed, and the relative differences between channels are considered. Then, in the design, one of the is set to unity, which means that the corresponding impulse response will be the unit impulse . In the following equations, the general case is considered for simplicity.

[0080] The output of the correction filter, i.e. the sequence , is given by the linear convolution of

[0081] (3)

[0082] Inserting equation (2) into equation (3), one obtains

[0083] (4)

[0084] where

[0085] (5)

[0086] is a periodic time-frequency function. Comparing equation (1) and equation (3), it can be seen that if where then the desired result is obtained: for all , . Thus, for a channel TIADC, the design challenge is to determine the impulse response such that the corresponding approximates unity in some sense. One way to achieve this goal is to minimize in the least-squares sense, which means that are minimized separately or simultaneously by

[0087] (6)

[0088] Minimizing them separately, each can be obtained in closed form.

[0089] The basic result applies to all linear convolution compensation filters, i.e. finite-length impulse response (FIR) and infinite-length impulse response (IIR) filters . In the case of FIR (or IIR), the sum in equation (3) is finite (or infinite for IIR). In practice, FIR compensation filters are usually preferred because their design and implementation are less complex and enable an inherently robust and high-speed implementation.

[0090] A drawback of using the above described general time-varying correction filter is that all filter coefficients will subsequently be implemented using general multipliers, which can lead to unnecessarily high implementation costs. Furthermore, online filter design can become expensive. Therefore, in order to reduce complexity, specialized correctors have been developed. These correctors are based on representing the channel response as a polynomial in . At a certain resolution and bandwidth, it can be assumed that the channel frequency response has a frequency-independent amplitude and phase delay response, which corresponds to a static gain and time skew (linear phase) mismatch. Using a static gain constant and a static time skew ​in percent of the sampling period , the channel frequency responses are then modeled as . For , this corresponds to , hence to a non-uniform sampling and reconstruction problem. The static gain constant is easily compensated by multiplying . Then, the time skew error can be compensated by first expanding in a Taylor series of order as follows

[0091] (7)

[0092] However, for higher resolutions and bandwidths, the amplitude and phase responses must be modeled as frequency dependent. Then, has been extended to be represented by a general order polynomial according to

[0093] (8)

[0094] where

[0095] (9)

[0096] The amplitude of corresponds roughly to the resolution of the uncompensated TIADC. According to the reasoning above and the Richardson iteration, the correction can be done in steps. In each step, the approximation of the required sequence is modified. Typically, in practice , depends on the TIADC match and the required resolution. Furthermore, in practice the implementation can be different in different stages, since in each stage the approximation of needs to be modified.

[0097] Finally, it is important to emphasize that to be able to perform the corrections outlined in this section, the needs to be known. In practice, these channel responses must be estimated, which can constitute a significant part of the overall calibration (i.e. estimation and correction) of the TIADC.

[0098] In the conventional correctors, the correction structure corresponds to a general time-varying filter, or to a general time-varying filter with the channel frequency responses modeled as in The disadvantage of using a general time-varying filter is that all filter coefficients will subsequently be implemented using general multipliers, which can lead to unnecessarily high implementation costs. Furthermore, the online filter design can become expensive. It is noted here that Periodic time-varying filters can be implemented as Channel systems (filter banks), but it does not change the basic principle or the implementation costs. For the existing polynomial-based techniques, one disadvantage is that one complex function is used to model the amplitude and phase response in each channel. This can lead to unnecessarily high model order for either the amplitude mismatch error or the phase mismatch error when one of these mismatch errors dominates. Finally, this leads to unnecessarily high implementation complexity of the corrector. Furthermore, since the amplitude and phase response in each channel is modeled by one function, the channel estimation becomes more complex.

[0099] As mentioned above, in the existing correctors for ADC systems, the correction structure corresponds to a general time-varying filter, or to a structure derived or equivalently derived in one time-varying function by modeling the channel frequency response as a polynomial as in On the other hand, in the provided corrector, it is assumed that the channel frequency response is described in terms of its amplitude response and phase response. Therefore, in the provided scheme, two time-varying functions are used for channel amplitude and phase response modeling and the corresponding correction. This makes the implementation cost of the correction and estimation lower compared to the conventional schemes.

[0100] In general, embodiments of the invention address linear mismatch errors, in which case the channels are modeled as linear systems, and thus have a specific frequency response. There are also static DC offset mismatch errors, but they are signal-independent and easy to compensate for, and are thus not included in the formulas given below. The general problem is addressed in which both the amplitude and phase response are frequency-dependent. In this respect, corrections are also considered in which an exact model of the channel frequency response is available. However, the estimation benefits from an efficient correction technique, for example, when the calibration is done by simultaneously estimating and correcting by minimizing an appropriate cost metric. Therefore, the provided correction technique can be used after estimating the channel frequency response or as part of the calibration technique.

[0101] Figure 3 A correction device 100 for a multi-channel time-interleaved ADC system 300 according to an embodiment of the invention is shown. An input analog signal is received by a multi-channel ADC device 210, which converts the input analog signal to a time-interleaved ADC output sequence According to embodiments of the application, the correction device 100 comprises an input 110 for receiving a time-interleaved ADC output sequence The correction device 100 further comprises correction blocks 120a, 122a coupled to the input 110 and configured to generate a first corrected output sequence of the ADC output sequence by correcting amplitude response mismatch errors of the ADC output sequence by a first time-varying function and correcting phase response mismatch errors of the ADC output sequence by a second time-varying function The correction device 100 further comprises an output 130 coupled to the correction blocks 120a, 122a and configured to output the corrected output sequence .

[0102] Figure 4 A flowchart of a method 200 according to embodiments of the application is shown, which can be implemented in a correction device 100 as shown in Figure 3 The method 200 comprises receiving (202) a time-interleaved ADC output sequence The method 200 further comprises generating (204) a first corrected output sequence of the ADC output sequence by correcting (206) amplitude response mismatch errors of the ADC output sequence by a first time-varying function and correcting (208) phase response mismatch errors of the ADC output sequence by a second time-varying function The method 200 further comprises outputting (210) the corrected output sequence Steps 206 and 208 of the method 200 can be performed in any suitable order, and are thus not limited to a particular order.

[0103] The correction herein can be performed by performing a weighted sum, linear filtering, linear convolution or any other suitable method using the first time-varying function and the second time-varying function

[0104] In embodiments of the application, the first time-varying function and the second time-varying function are real-valued functions. Furthermore, the functions and may be periodic, where corresponds to the number of channels of the multi-channel ADC.

[0105] ​In a further embodiment of the application, the first time-varying function and the second time-varying function are independent functions. This can be understood as the first time-varying function and the second time-varying function are uncorrelated.

[0106] The provided correction device 100 can also correct signals whose frequency content lies in the higher Nyquist band. This is understandable because all equations hold true regardless of the actual value of the angular frequency However, when approximating the channel frequency response mismatch with digital filters in the correction device 100, the filter design is affected because different Nyquist bands correspond to different sets of filter coefficients.

[0107] Generally, in the provided correction device 100, the channel frequency response is described in terms of their amplitude response (e.g. ) and phase response (e.g. ), so the channel frequency response can be expressed as

[0108] (10)

[0109] Here, the first time-varying function and the second time-varying function represent the deviation from the ideal unity amplitude response and zero phase response, respectively. In practice, and are small (typically a few percent), usually frequency dependent. By exploiting the Taylor series expansion of , retaining the term, the in equation (10) can be written as

[0110] (11)

[0111] where

[0112] (12)

[0113] In an embodiment of the present invention, the calibration device 100 includes a plurality of calibration blocks, i.e., two or more calibration blocks. Therefore, in this case, the aforementioned calibration blocks 120a and 122a can be represented as first calibration blocks, and the calibration device 100 further includes at least one second calibration block 120b and 122b, respectively coupled to the first calibration blocks 120a and 122a and the output terminal 130. The output terminal 130 is coupled to the first calibration block via the second calibration blocks. The second calibration blocks 120b and 122b are further used to generate a second calibration output sequence through the following steps. : Through the first time-varying function Correcting the first correction output sequence The amplitude response mismatch error is determined by the second time-varying function. Correcting the first correction output sequence The phase response mismatch error.

[0114] Using Richardson iteration, the correction can be performed in principle as follows: Figure 5 Execute as shown. Figure 5 Embodiments of the present invention are disclosed, wherein the calibration device 100 includes a plurality of calibration blocks, and different calibration blocks include the same error estimation function. It should be noted that each correction block includes an error estimation function block 120p and an associated adder 122p. Different blocks of the correction device 100 are coupled or connected to each other via coupling / connection components for transmitting sequences within the correction device 100.

[0115] Therefore, the calibration device 100 may include a set of The calibration blocks 120a, 122a...120P, 122P can be indexed by letters. As an index, where It is an integer greater than or equal to 2. This group Each correction block includes first correction blocks 120a and 122a and second correction blocks 120b and 122b. The correction blocks are coupled to each other such that the corrected output sequence of a correction block is provided to the next subsequent correction block, and so on, thereby providing a corrected correction for each correction stage. For example, in... Figure 5 In the middle, provide ADC output sequence This serves as the input to the first correction blocks 120a and 122a. The error estimation function in block 120a... The output sequence is generated and taken from the ADC output sequence in adder 122a. Subtract from the corrected output sequence from adder 122a. The input sequences will be the second correction blocks 120b and 122b, which also include the error estimation function from block 120b. As described above, an error estimate function is output and subtracted from the ADC output sequence in adder 122b. The corrected output sequence from adder 122b will be the input sequence to the third correction block 120c, 122c, and the process is repeated for each stage in the correction device 100. The correction device 100 comprises stages or correction blocks, and finally, the corrected output sequence (i.e. Figure 5 in ) is output from the correction device 100.

[0116] As mentioned in the embodiments disclosed above in Figure 5 , the same error estimate function is used in each stage. Thus, each of the set of correction blocks 120a, 122a... 120P, 122P is configured to implement an error estimate function

[0117] ,

[0118] where denotes the angular frequency, denotes the time index, denotes the sampling period, denotes the summation index, denotes the index of the respective correction block, denotes the total number of correction blocks of the correction device 100, denotes the imaginary unit, is the th power of .

[0119] In Figure 5 , the error estimate function may correspond to a non-causal digital filter with a frequency response . In a practical causal implementation of the correction, an additional delay can be incorporated so that all parallel branches have the same delay, the outputs of which are added or subtracted. For digital filters, instead of the usual , the notation is used to indicate that they can be written in terms of a real-valued magnitude function and a phase function .

[0120] In the disclosed scheme of Figure 5 , two independent time-varying functions, i.e. and for channel amplitude and phase response modeling and corresponding correction. In the conventional scheme, a time-varying function for amplitude and phase response modeling and correction.

[0121] Furthermore, in a further embodiment of the present application, the error estimation function is not identical in different correction stages, as it can ensure that the error estimation function is a order approximation in stage . Thus, the correction can be performed by a modified approximation of the error estimation function in each correction stage, as shown in Figure 6 , where, for correction block index , the error estimation function can be expressed as

[0122] (13)

[0123] For correction block index , the error estimation function can be expressed as

[0124] (14)

[0125] It should be noted that, since the functions and in the embodiments of the present application are real-valued functions, the first time-varying function can be approximated by a linear phase FIR filter with symmetric coefficients, while the second time-varying function multiplied by the imaginary unit (i.e. ) can be approximated by a linear phase FIR filter with non-symmetric coefficients. Compared with a general non-symmetric filter, these filters only need half of the multipliers in implementation.

[0126] In other embodiments of the present application, when the correction device 100 comprises a plurality of correction blocks, different correction blocks comprise different error estimation functions, instead of the same error estimation function as disclosed in the Figure 5 . Generally, the correction device 100 comprises a set of correction blocks 120a, 122a...120P, 122P, where is an integer greater than or equal to 2. The set of correction blocks comprises a first correction block 120a, 122a and a second correction block 120b, 122b. According to the present embodiment, the first correction block 120a, 122a is used to implement a first error correction function as ​​Furthermore, the set of correction blocks 120a, 122a,... 120p, 122p The correction block 120p, 122p is used to apply the error correction function is implemented as wherein wherein denotes the angular frequency, denotes the time index, denotes the sampling period, denotes the index of the respective correction block, denotes the summation index, denotes the total number of correction blocks of the correction device 100, denotes the imaginary unit, is the power of

[0127] According to the channel frequency response, it can be beneficial to model it according to a polynomial. Since and are real-valued functions, it is appropriate here to use polynomials in instead of the previously used Therefore, in embodiments of the invention, the first time-varying function is approximated by a polynomial of order and the second time-varying function is approximated by a polynomial of order and can be modeled as

[0128] (15)

[0129] wherein denotes the angular frequency, denotes the time index, denotes the summation index.

[0130] The correction device can then be implemented according to variable multipliers and and fixed sub-filters. All fixed sub-filters implementing the terms in can be designed as symmetric linear phase FIR filters, while all sub-filters resulting from the expansion of can be designed as symmetric or anti-symmetric linear phase FIR filters. The variable multipliers and are implemented with general multipliers, while all fixed sub-filters can be implemented with fixed multipliers, which are less costly to implement. According to the required polynomial order and ) and the filter order of the fixed sub-filters, the polynomial-based structure can be less costly than the general structure above. However, in terms of the function , this approach is useful when and / or is small. Otherwise, the expansion of the function in equation (14) will result in many fixed sub-filters that need to be implemented.

[0131] An alternative is to use phase delays in the channel model instead of using phases. In this case, the second time-varying function is expanded. The phase shift is included and then the channel is modeled as

[0132] (16)

[0133] where the second time-varying function is , denotes the phase shift, is a constant, denotes the sign of , corresponding to the negative value of the channel phase delay, can be expressed as where is the phase response, i.e., including phase delays of the multi-channel ADC system. In this case, the error estimation functions and in equations (13) and (14) are replaced by

[0134] (17)

[0135] and

[0136] (18)

[0137] This structure can have an advantage if the filter order required to implement is lower than the filter order required to implement . This is because the part like is a fixed common part for all channels, so only one fixed instance is needed in the implementation.

[0138] For well-matched TIADCs, only first-, second-, and third-order compensators are usually of practical interest, i.e., for , or ​In many cases, a first order corrector with good performance is sufficient. If the spurious-free dynamic range (SFDR) is dB full scale (dBFS) uncorrected, then after first order correction, approximately dBFS can be achieved. The exact numbers will depend on the channel frequency response and the number of channels of the ADC system.

[0139] Based on the channel model in equation (16), the provided first order correction device 100 can be designed as shown in Figure 7 where the expressions correspond to a Hilbert filter, while correspond to a differentiator, both of which can be implemented using anti-symmetric linear phase FIR filters. It is also noted that the functions and are periodic time-varying filters, which can be implemented in a time-interleaved fashion from time-invariant filters. Thus, in summary, each error estimate function can be implemented by four FIR filters, one for each of , , and . It is also noted that the blocks Figure 7 , , and in together represent the second time-varying function .

[0140] Furthermore, the disclosed approach also relates to a multi-channel time-interleaved ADC system 300 comprising one or more correction devices 100 according to embodiments of the invention. Figure 3 A multi-channel time-interleaved ADC system 300 is shown, wherein the multi-channel time-interleaved ADC 210 is coupled to the correction device 100.

[0141] For such a multi-channel time-interleaved ADC system 300, many applications are possible, Figure 8Non-limiting example implementations of the correction device 100 and the multi-channel time-interleaved ADC system 300 are shown. For example, in a communication system 600, such as 3GPP LTE and 5G (also known as new radio (NR)), a network access node 500 (e.g., a base station) and a client device 400 (e.g., a UE) can include the correction device 100 and / or the multi-channel time-interleaved ADC system 300. Wireless signals are transmitted in downlink (DL) and uplink (UL) communications between the network access node 500 and the client device 400. Analog radio signals are received at the network access node 500 and the client device 400 and, in some cases, converted to digital sequences.

[0142] The client devices in the present disclosure can include, but are not limited to, the following UEs: a smartphone, a cellular phone, a cordless phone, a session initiation protocol (SIP) phone, a wireless local loop (WLL) station, a personal digital assistant (PDA), a handheld device having wireless communication function, a computing device or other processing device connected to a wireless modem, an in-vehicle device, a wearable device, an integrated access and backhaul (IAB) node (e.g., a mobile car or a device installed in a car), a drone, a device-to-device (D2D) device, a wireless video camera, a mobile station, an access terminal, a subscriber unit, a wireless communication device, a wireless local access network (WLAN) station, a tablet computer with wireless capability, a notebook computer embedded device, a universal serial bus (USB) dongle, a wireless customer-premises equipment (CPE), and / or a chipset. In an internet of things (IOT) scenario, a client device can represent a machine or another device or chipset that performs communication with another wireless device and / or network device. A UE can also be referred to as a mobile phone, a cellular phone, or a computer tablet or laptop with wireless capability. In this context, a UE can be, for example, a portable, pocket-storable, hand-held, computer-comprised, or vehicle-mounted mobile device, enabled to communicate voice and / or data with another entity, such as another receiver or server, through a radio access network. A UE can be a station (STA), which is any device that contains an IEEE 802.11 -conformant media access control (MAC) and physical layer (PHY) interface to the wireless medium (WM). A UE can also be used to communicate in 3GPP-related LTE and LTE-Advanced, in WiMAX and its evolution, and in fifth generation wireless technology, such as in NR.

[0143] The network access node in the present application can include, but is not limited to, a NodeB in a wideband code division multiple access (WCDMA) system, an evolved NodeB (eNB) or evolved NodeB (eNodeB) in an LTE system, a relay node, an access point, a vehicle-mounted device, a wearable device, or a gNB in a fifth generation (5G) network. Further, the network access node herein can be denoted as a radio network access node, an access network access node, an access point, or a base station, such as a radio base station (RBS), which in some networks can be referred to as a transmitter, a "gNB", a "gNodeB", an "eNB", an "eNodeB", a "NodeB" or a "B node", depending on the technology and terminology used. The radio network access node can be of different classes such as a macro eNodeB, a home eNodeB, or a micro base station, depending on transmission power and thereby also cell size. The radio network access node can be a station (STA), which is any device that contains an IEEE 802.11 -conformant MAC and PHY interface to the wireless medium. The radio network access node can also be a base station corresponding to a 5G wireless system.

[0144] Furthermore, any of the methods provided by the embodiments of the present application can be implemented in a computer program having a code segment for implementing the method steps. The computer program includes a computer-readable medium having stored the computer program. The computer-readable medium can include any memory, such as a read-only memory (ROM), a programmable read-only memory (PROM), an erasable PROM (EPROM), a flash memory, an electrically erasable PROM (EEPROM), or a hard disk drive.

[0145] Furthermore, it should be appreciated that an embodiment of the correction device 100 comprises the necessary capabilities to perform the scheme in the form of, for example, functions, means, units, elements, etc. Examples of these means, units, elements and functions are: processors, memories, buffers, control logic, etc., which are suitably arranged together to perform the scheme. In particular, the processor of the correction device 100 can comprise one or more instances of a central processing unit (CPU), a processing unit, a processing circuit, a processor, a microprocessor, an application-specific integrated circuit (ASIC), or other processing logic that can interpret and execute instructions. Hence, the expression "processor" can represent a processing circuit comprising a plurality of processing circuits, such as, for example, any, some or all of the processing circuits described above. The processing circuitry can further perform data processing functions for inputting, outputting, and processing of data comprising data buffering and device control functions, such as call processing control, user interface control, or the like.

[0146] Finally, it should be understood that the present application is not limited to the embodiments described above, but also relates to all embodiments within the scope of the appended independent claims.

Claims

1. A calibration device (100) for a multi-channel time-interleaved analog-to-digital converter (ADC) system (300), characterized in that, The calibration device (100) includes: Input terminal (110) is used to receive the time-interleaved ADC output sequence. ; Multiple calibration blocks (120a, 122a) are coupled to the input terminal (110). The multiple calibration blocks are coupled sequentially, and the calibration output sequence of the previous calibration block is provided to the next subsequent calibration block. The first calibration block among the multiple calibration blocks is used to generate the ADC output sequence through the following steps. First corrected output sequence : Through the first time-varying function Correcting the ADC output sequence The amplitude response mismatch error, Through the second time-varying function Correcting the ADC output sequence Phase response mismatch error; The output terminal (130) is coupled to the output of the plurality of correction blocks (120a, 122a) and is used to output the correction output sequence. .

2. The calibration device (100) according to claim 1, characterized in that, The first time-varying function and the second time-varying function It is a real-valued function.

3. The calibration device (100) according to claim 1, characterized in that, The first time-varying function and the second time-varying function It is an independent function.

4. The calibration device (100) according to claim 2 or 3, characterized in that, The first time-varying function It is the first FIR filter with symmetric coefficients. The second time-varying function multiplied by the imaginary unit It is a second FIR filter with antisymmetric coefficients.

5. The calibration device (100) according to claim 4, characterized in that, The first FIR filter and the second FIR filter are linear phase FIR filters.

6. The calibration device (100) according to claim 1, characterized in that, The calibration device (100) includes a second calibration block (120b, 122b) coupled to the first calibration block (120a, 122a) and the output terminal (130), respectively. The second correction block (120b, 122b) is used to generate the second correction output sequence through the following steps. : Through the first time-varying function Correcting the first correction output sequence The amplitude response mismatch error, Through the second time-varying function Correcting the first correction output sequence The phase response mismatch error.

7. The calibration device (100) according to claim 6, characterized in that, Including a group There are several calibration blocks (120a, 122a...120P, 122P), among which... The set is an integer greater than or equal to 2, wherein... The calibration blocks include the first calibration block (120a, 122a) and the second calibration block (120b, 122b), and wherein the set Each of the calibration blocks (120a, 122a...120P, 122P) is used to implement the following error estimation function. ， in, Represents angular frequency. Indicates a time index. Indicates the sampling period. Indicates the summation index. Indicates the index of the corresponding correction block. This indicates the total number of calibration blocks in the calibration device (100). Represents the imaginary unit. yes of Power of 1.

8. The calibration device (100) according to claim 6, characterized in that, Including a group There are several calibration blocks (120a, 122a...120P, 122P), among which... The set is an integer greater than or equal to 2, wherein... Each correction block includes the first correction block (120a, 122a) and the second correction block (120b, 122b), and wherein, The first correction blocks (120a, 122a) are used to implement the following first error correction function. , ， The group The first correction block (120p, 122p) (120a, 122a...120P, 122P) The correction block is used to achieve the following: Error correction function ,in, , in, Represents angular frequency. Indicates a time index. Indicates the sampling period. Indicates the index of the corresponding correction block. Indicates the summation index. This indicates the total number of calibration blocks in the calibration device (100). Represents the imaginary unit. yes of Power of 1.

9. The calibration device (100) according to claim 7 or 8, characterized in that, The first time-varying function for polynomial of order , The second time-varying function for polynomial of order , in, Represents angular frequency. Indicates a time index. Indicates the summation index.

10. The calibration device (100) according to claim 7 or 8, characterized in that, The second time-varying function for , in, Indicates a time index. Indicates phase shift, This represents a negative value for the phase delay of multiple channels.

11. A multi-channel time-interleaved ADC system (300), characterized in that, Includes one or more calibration devices (100) according to any one of claims 1 to 10.

12. A method (200) for a calibration apparatus (100) for a multi-channel time-interleaved ADC system (300), characterized in that, The method (200) is applied to the calibration device (100) according to any one of claims 1 to 10, the method comprising: Receive (202) time-interleaved ADC output sequence ; The ADC output sequence (204) is generated through the following steps. First corrected output sequence : Through the first time-varying function Correction (206) of the ADC output sequence The amplitude response mismatch error; Through the second time-varying function Correction (208) of the ADC output sequence Phase response mismatch error; Output (210) Corrected output sequence .

13. A computer program product having program code, characterized in that, When the computer program product is run on a computer, the program code is used to perform the method according to claim 12.

Images