ADC calibration device and ADC calibration method

Abstract

To provide an ADC calibration device and the like capable of calibrating a time-interleaved ADC at high speed.SOLUTION: A calibration device for a TI-ADC having a plurality of ADC includes: a calibration signal generator 21 for generating a calibration signal to be inputted to the TI-ADC; a sample signal storing part 22 for storing sample signals obtained by AD conversion for the calibration signal at a predetermined sampling frequency in the TI-ADC; and a correction information calculation part 23 for calculating a frequency characteristic of a mismatch characteristic among ADC from the stored sample signals and calculating correction information for correcting mismatch among the ADC from the frequency characteristic of the mismatch characteristic. The calibration signal generator generates a multi-carrier signal, in which a plurality of carrier signals mutually having different frequencies within a frequency range whose range is 1 / 2 or less of the sampling frequency are combined, as the calibration signal, and each carrier signal has a frequency different from any alias included in the sample signal.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an apparatus and method for calibrating an ADC, and more particularly to an apparatus and method for calibrating a time-interleaved ADC. [Background technology]

[0002] Among devices that convert analog signals into digital signal strings for various processing, time-interleaved AD conversion devices (hereinafter also referred to as time-interleaved ADCs or TI-ADCs) are used for high-speed analog signal processing. In a time-interleaved ADC, the analog signal to be converted is input to multiple AD converters (hereinafter also referred to as ADCs), and a sampling clock with a predetermined period is given to each ADC with a slightly delayed timing, thereby performing equivalent high-speed sampling digital conversion.

[0003] In a time-interleaved ADC, a technique is known in which the mismatch is corrected by calibration in order to reduce fluctuations in sample values ​​and occurrence of spurious signals due to mismatches between individual ADCs (see, for example, Patent Document 1).

[0004] Patent Document 1 discloses a method for correcting mismatches in a time-interleaved ADC. Conventional calibrations of time-interleaved ADCs, such as those described in Patent Document 1, use an unmodulated CW signal (a tone signal with a single frequency) as a calibration signal. Specifically, a sine wave signal is input to the time-interleaved ADC as the unmodulated CW signal, and the outputs from each ADC are compared to measure the degree of mismatch between the ADCs, thereby correcting the mismatch. Generally, mismatches between ADCs have frequency characteristics that vary depending on the frequency. Therefore, in measuring the mismatch, it is necessary to acquire the mismatch characteristics over the entire bandwidth used by changing the frequency of the unmodulated CW signal. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Patent No. 6508665 Summary of the Invention [Problem to be solved by the invention]

[0006] However, the conventional calibration of time-interleaved ADCs described in Patent Document 1 and elsewhere uses a single-tone unmodulated CW signal as the calibration signal, which poses a problem of requiring a relatively long time to acquire the frequency characteristics of mismatches between ADCs. This long calibration time is particularly significant when the frequency resolution is high over a wide band.

[0007] The present invention has been made to solve the above-mentioned problems, and has an object to provide an ADC calibration device and an ADC calibration method that can perform high-speed calibration of a time-interleaved ADC. [Means for solving the problem]

[0008] To achieve the above object, the calibration device for an AD conversion device according to the present invention is a calibration device for an AD conversion device that operates a plurality of AD converters in a time interleaved manner, and comprises: a calibration signal generator (21) that generates a calibration signal to be input to the AD conversion device; a sample signal storage unit (22) that stores sample signals obtained by AD converting the calibration signal at a predetermined sampling frequency in the AD conversion device; and a correction information calculation unit (23) that calculates, from the sample signals stored in the sample signal storage unit, frequency characteristics of mismatch characteristics that characterize mismatch between the plurality of AD converters, and calculates correction information for correcting the mismatch between the plurality of AD converters from the frequency characteristics of the mismatch characteristics, wherein the calibration signal generator generates, as the calibration signal, a multicarrier signal obtained by combining a plurality of carrier signals having different frequencies within a frequency range not exceeding half the sampling frequency, and each of the carrier signals has a frequency different from that of any of the aliases contained in the sample signals.

[0009] This configuration uses a multi-carrier signal instead of the conventional unmodulated CW signal, and by arranging multiple carrier signals over a wide range to avoid the effects of aliasing, it is possible to simultaneously detect characteristic differences at multiple frequency points and reduce the need to switch the frequency of the calibration signal, enabling high-speed calibration.

[0010] To achieve the above object, the calibration device for an AD conversion device according to the present invention is a calibration device for an AD conversion device that operates a plurality of AD converters in a time interleaved manner, and comprises: a calibration signal generator (21) that generates a calibration signal to be input to the AD conversion device; a sample signal storage unit (22) that stores sample signals obtained by AD converting the calibration signal at a predetermined sampling frequency in the AD conversion device; and a correction information calculation unit (23) that calculates, from the sample signals stored in the sample signal storage unit, frequency characteristics of mismatch characteristics that characterize mismatch between the plurality of AD converters, and calculates correction information for correcting the mismatch between the plurality of AD converters from the frequency characteristics of the mismatch characteristics, wherein the calibration signal generator generates, as the calibration signal, a multicarrier signal obtained by combining a plurality of carrier signals having different frequencies within a frequency range not less than half the sampling frequency and not more than the sampling frequency, and wherein each of the carrier signals has a frequency different from that of any of the aliases contained in the sample signals.

[0011] As described above, in the calibration device for an AD conversion device according to the present invention, the calibration signal generator generates as a calibration signal a multi-carrier signal obtained by combining multiple carrier signals having different frequencies within a frequency range equal to or greater than half the sampling frequency and equal to or less than the sampling frequency, and each carrier signal is set to have a frequency different from any of the aliases contained in the sample signal. Thus, even in the case of a frequency range exceeding the Nyquist frequency, i.e., in the case of undersampling, by allocating multiple carrier signals over a wide range while avoiding the effects of aliasing, as in the case of allocating multiple carrier signals within a frequency range equal to or less than the Nyquist frequency, it is possible to simultaneously detect characteristic differences at multiple frequency points and reduce the need to switch the frequency of the calibration signal, thereby enabling high-speed calibration.

[0012] In the calibration device for AD conversion devices according to the present invention, the Nyquist frequency f of each of the AD converters is set within a frequency range equal to or less than half the sampling frequency. s Each Nyquist zone R defined by / 2m i The same number of carrier signals are arranged at equal intervals within each Nyquist region R i The frequency difference between the minimum frequency of the carrier signal with the lowest frequency among the carrier signals arranged within each Nyquist zone R and the frequency at one end of the Nyquist zone to which the carrier signal with the lowest frequency belongs is i However, f s is the sampling frequency, m is the number of AD converters, and the Nyquist region R i (i=1,2,...,m) is (f s / 2m)·(i-1) or more and (f s / 2m)·i.

[0013] This configuration makes it possible to distribute multiple carrier signals almost evenly across the entire frequency range below half the sampling frequency, i.e., the entire frequency range in which the original signal can be reproduced based on the sampling theorem, while avoiding the effects of aliasing. This enables accurate and efficient measurement of the frequency characteristics of mismatch between AD converters.

[0014] In the calibration device for an AD conversion device according to the present invention, the frequencies of the plurality of carrier signals that make up the multicarrier signal may be set according to the following equation:

number

[0015] This configuration ensures that the frequencies of the carrier signals that make up the calibration signal do not overlap with those of the aliases that are aliased. It also makes it possible to simultaneously obtain mismatch characteristics between AD converters across multiple Nyquist zones. This allows accurate and efficient measurement of the frequency characteristics of mismatch characteristics between AD converters.

[0016] The calibration device for an AD conversion device according to the present invention may be configured to measure mismatch characteristics between the AD converters while changing the frequency range in which the plurality of carrier signals are allocated, and to measure the frequency characteristics of the mismatch characteristics over the entire frequency range not exceeding half the sampling frequency.

[0017] With this configuration, even if the frequency range in which the AD conversion device is expected to be used is wide, it is possible to measure the frequency characteristics of the mismatch characteristics between AD converters over the entire frequency range.

[0018] In the calibration device for an AD conversion device according to the present invention, the mismatch characteristics may include at least one of a relative amplitude ratio and a relative phase difference of each of the AD converters with respect to one of the plurality of AD converters as a reference, and a DC offset.

[0019] This configuration makes it possible to properly grasp the mismatch between the AD converters, and therefore to effectively correct the mismatch between the AD converters.

[0020] In a calibration device for an AD conversion device according to the present invention, the AD conversion device may include one or more AD converter cores, each of which may include a plurality of sub-AD converters, and the correction information calculation unit may be configured to calculate at least one of the frequency characteristics of the mismatch characteristics between the AD converter cores and the frequency characteristics of the mismatch characteristics between the sub-AD converters.

[0021] This configuration allows calibration even if the AD conversion device has a hierarchical structure including an AD converter core and sub AD converters.

[0022] In the calibration device for an AD conversion device according to the present invention, the calibration signal generator may be configured to adjust the crest ratio of the sample signal by adjusting the phase of each of the carrier signals.

[0023] This configuration makes it possible to suppress the crest ratio (ie, peak value / effective value).

[0024] A method for calibrating an AD conversion device according to the present invention is a method for calibrating an AD conversion device in which a plurality of AD converters operate in a time interleaved manner, and includes: a calibration signal generating step of generating a calibration signal to be input to the AD conversion device; a storage step of storing in a sample signal storage unit a sample signal obtained by AD converting the calibration signal at a predetermined sampling frequency in the AD conversion device; and a correction information calculating step of calculating, from the sample signal stored in the sample signal storage unit, a frequency characteristic of mismatch characteristics that characterize mismatch between the plurality of AD converters, and calculating correction information for correcting the mismatch between the plurality of AD converters from the frequency characteristic of the mismatch characteristics, wherein the calibration signal generating step generates, as the calibration signal, a multicarrier signal obtained by multiplexing a plurality of carrier signals having different frequencies within a frequency range not more than half the sampling frequency, and each of the carrier signals has a frequency different from that of any of the aliases contained in the sample signal.

[0025] This configuration uses a multi-carrier signal instead of the conventional unmodulated CW signal, and by arranging multiple carrier signals over a wide range to avoid the effects of aliasing, it is possible to simultaneously detect characteristic differences at multiple frequency points and reduce the need to switch the frequency of the calibration signal, enabling high-speed calibration.

[0026] To achieve the above object, a method for calibrating an AD conversion device according to the present invention is a method for calibrating an AD conversion device in which a plurality of AD converters are operated in a time interleaved manner, and includes the following steps: a calibration signal generating step of generating a calibration signal to be input to the AD conversion device; a storage step of storing sample signals obtained by AD converting the calibration signal at a predetermined sampling frequency in the AD conversion device in a sample signal storage unit; and a correction information calculating step of calculating, from the sample signals stored in the sample signal storage unit, frequency characteristics of mismatch characteristics that characterize mismatch between the plurality of AD converters, and calculating correction information for correcting the mismatch between the plurality of AD converters from the frequency characteristics of the mismatch characteristics, wherein the calibration signal generating step generates, as the calibration signal, a multicarrier signal obtained by multiplexing a plurality of carrier signals having different frequencies within a frequency range not less than half the sampling frequency and not more than the sampling frequency, and each of the carrier signals has a frequency different from that of any of the aliases contained in the sample signals.

[0027] In this way, even when the calibration signal is placed within a frequency range above the Nyquist frequency, i.e., in the case of undersampling, similar to the case where multiple carrier signals are placed within a frequency range equal to or lower than the Nyquist frequency, by placing multiple carrier signals over a wide range while avoiding the effects of aliasing, it is possible to simultaneously detect characteristic differences at multiple frequency points and reduce the need to switch the frequency of the calibration signal, thereby enabling high-speed calibration. [Effects of the Invention]

[0028] According to the present invention, it is possible to provide an ADC calibration device and an ADC calibration method that can perform high-speed calibration of a time-interleaved ADC. [Brief explanation of the drawings]

[0029] [Figure 1] 1 is a diagram illustrating a configuration of an ADC calibration device according to an embodiment of the present invention. [Figure 2]FIG. 1 is a diagram illustrating an example of the configuration of a time-interleaved ADC. [Figure 3] FIG. 10 illustrates the frequency spectrum of a sample signal output from a time-interleaved ADC when a multi-carrier signal is used as a calibration signal in accordance with one embodiment of the present invention. [Figure 4] 1A is a diagram showing the frequency spectrum of a multi-carrier signal used as a calibration signal in one embodiment of the present invention, and FIG. 1B is a diagram showing the frequency spectrum of a sample signal output from a time-interleaved ADC. [Figure 5] FIG. 2 is a diagram illustrating the configuration of an ADC core. [Figure 6] 1A and 1B are diagrams showing the connection between an ADC core and a calibration signal generator, where (a) shows the case where the time-interleaved ADC consists of a single ADC core, and (b) shows the case where the time-interleaved ADC consists of multiple ADC cores. DETAILED DESCRIPTION OF THE INVENTION

[0030] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0031] (Time-interleaved AD conversion device) First, the time-interleaved ADC 10 will be described.

[0032] FIG. 2 shows an example of the basic configuration of the time-interleaved ADC 10.

[0033] In this time-interleaved ADC 10, an analog input signal IN(t) input to an input terminal 10a is branched into m signal paths by a signal distributor 11 such as a power divider, and m ADCs 120 to 12 m-1 Enter the following information into the boxes.

[0034] The sampling control unit 13 generates sampling clocks C0 to C1 each having a period T and phases shifted by ΔT (=T / m). m-1 and ADC120~12 respectively.m-1 and each ADC120~12 m-1 A designation signal ADNUM that designates the ADC that will perform sampling is provided to the signal switch 14.

[0035] Each ADC120~12 m-1 are clocks C0 to C m-1 The input value IN is sampled and converted to a digital value, and each sample value X0, X1, ..., X m-1 are output to the signal switcher 14.

[0036] The signal switch 14 is connected to each of the ADCs 120 to 122. m-1 The sample values ​​X0, X1, X2, ... output from the ADC specified by the specification signal ADNUM are selected in order, and a digital signal sequence OUT(n) in which the sample values ​​are arranged in the sampling order is output to the output terminal 10b.

[0037] The digital signal sequence OUT(n) obtained in this way is equivalent to that obtained by sampling the input signal IN(t) at a sampling period ΔT that is 1 / m of the clock period T, allowing high-speed sampling to be performed with a low-speed ADC.

[0038] However, as in the above-mentioned time-interleaved ADC 10, the input signal IN(t) is input to a plurality of ADCs 120 to 122. m-1 When the signal is distributed to the ADCs 120 to 122, the signal divider 12 itself has different distribution characteristics, the frequency characteristics of the distribution paths are different, and the ADCs 120 to 122 have different distribution characteristics. m-1 Due to the difference in frequency characteristics, errors occur in the results of signal processing of the obtained sample values.

[0039] Also, each ADC 120~12 m-1 Regarding the clock that determines the sampling timing, timing errors occur due to differences in the signal path length and differences in the delay characteristics of each ADC relative to the sampling clock, and errors occur in the results of signal processing of the obtained sample values.

[0040] To reduce such errors, it is necessary to reduce the influence of the non-uniformity (mismatch) of the characteristics from these input terminals to the ADC. m-1 A correction processing unit for correcting mismatch is provided between the signal switch 14 and the time interleaved ADC 10, thereby improving the accuracy of the time interleaved ADC 10.

[0041] The ADC calibration device 20 according to the embodiment of the present invention acquires correction information required to correct mismatches between ADCs.

[0042] (ADC calibration device) Next, an ADC calibration device 20 according to an embodiment of the present invention will be described.

[0043] FIG. 1 is a diagram showing the configuration of an ADC calibration device 20 according to an embodiment of the present invention.

[0044] The ADC calibration device 20 includes a plurality of ADCs 120 to 122. m-1 The calibration device is for calibrating a time-interleaved ADC 10 that operates in a time-interleaved manner, and includes a calibration signal generator 21, a sample signal storage unit 22, and a correction information calculation unit 23.

[0045] Calibration signal generator 21 generates a calibration signal to be input to time-interleaved ADC 10. Specifically, calibration signal generator 21 generates a multi-carrier signal, which is a combination of multiple carrier signals having different frequencies, as the calibration signal. In other words, calibration signal generator 21 is a generator of a multi-carrier signal, and an arbitrary waveform signal generator, for example, can be used as calibration signal generator 21.

[0046] The sample signal storage unit 22 stores sample signals obtained by AD converting the calibration signal at a predetermined sampling frequency in the time-interleaved ADC 10. The sample signal storage unit 22 functions as a waveform memory that stores the output waveform of the time-interleaved ADC 10.

[0047] The correction information calculation unit 23 calculates the frequency characteristics of the mismatch characteristics that characterize the mismatch between the ADCs from the sample signals stored in the sample signal storage unit 22, and calculates correction information for correcting the mismatch between the ADCs from the frequency characteristics of the mismatch characteristics.

[0048] The ADC calibration device 20 may be configured in part or in whole by a computer having a CPU, memory, and the like.

[0049] (mismatch characteristics) Multiple ADCs 120,...,12 m-1 A sinusoidal CW signal (frequency f ref Consider the case where a single carrier signal (frequency [Hz], amplitude B [V]) is input as the calibration signal. m-1 are sampled in order at equal intervals.

[0050] The correction information calculation unit 23 calculates the correction information for the ADCs 120, . . . , 12 m-1 From the sample signals captured from the respective channels, for example, (i) relative amplitude ratio, (ii) relative phase difference (relative timing difference), and (iii) DC offset are calculated as mismatch characteristics.

[0051] The sample signal waveform sequence (real number) obtained by the time-interleaved ADC10 (sampling frequency: A [Hz]) x(k), k: 0,...,N-1 Then, each ADC120,...,12 m-1 The sample signal waveform series x0(l)=x(m1+0),l:0,...,L-1 ............... x m-1 (l)=(ml+m-1),l:0,...,L-1 Here, N and L represent the number of each sequence, and N is a multiple of m. N = mL It is assumed that:

[0052] Sample signal sequence x i DC offset DC in (k) i ,Gain i ,Phase i The gain can be calculated using the following formula: i Relative amplitude ratio and phase i The relative phase difference (relative timing difference) can be calculated from

[0053]

number

[0054]

number

[0055]

number

[0056] (mismatch correction value) The correction information calculation unit 23 calculates correction information for correcting mismatches between ADCs based on the frequency characteristics of mismatch characteristics such as the relative amplitude ratio, relative phase difference (relative timing difference), and DC offset between the ADCs. For example, when correcting mismatches between ADCs using a digital filter such as an FIR filter, filter coefficients are calculated for each ADC from the frequency characteristics of the mismatch characteristics using a known method, and the calculated coefficients are stored in an internal or external storage device as correction information (mismatch correction values). When interleaved ADCs are used, filtering is performed using the filter coefficients obtained for each ADC to correct mismatches between the ADCs.

[0057] (calibration signal) Generally, mismatch characteristics have frequency characteristics, so it is necessary to acquire the mismatch characteristics at each frequency. When acquiring mismatch characteristics for multiple frequencies, a conventional method has been to use a single CW signal as a calibration signal and vary its frequency. While this method is good in terms of performance and cost, it suffers from the problem of slow measurement speed. In an embodiment of the present invention, a multicarrier signal is generated by combining multiple carrier signals based on an arbitrary waveform, and this signal is used as the calibration signal.

[0058] When a multi-carrier signal is used as a calibration signal according to an embodiment of the present invention, multiple frequency points can be measured simultaneously, significantly reducing the time required for calibration.

[0059] On the other hand, since a multi-carrier signal obtained by combining multiple carrier signals is used, it is necessary to avoid the influence of aliasing when measuring the frequency characteristics of the mismatch characteristics. For this reason, in the embodiment of the present invention, carrier signals are not placed at frequencies that are in an aliasing relationship.

[0060] Figure 3 shows the frequency spectrum of the sample signal output from the time-interleaved ADC when a multi-carrier signal is used as the calibration signal according to this embodiment. The example in Figure 3 shows the case of a time-interleaved ADC (sampling rate Fs) consisting of two ADCs. Each ADC has the following characteristics: (i) The sampling rate is Fs / 2 (=Fs / m, m: number of ADCs), (ii) The Nyquist zone includes DC to Fs / 4 (first Nyquist zone) and Fs / 4 to Fs / 2 (second Nyquist zone), (iii) The spectra of adjacent Nyquist zones are in a folded relationship.

[0061] In the example shown in Figure 3, either the 10 signals indicated by the solid lines or the 10 signals indicated by the dotted lines can be used as calibration signals. In this way, mismatch characteristics can be obtained using calibration signals spanning multiple Nyquist zones. In this case, if it is not possible to cover the required frequency range at once, it is a good idea to change the measurement range while making measurements to cover the entire range.

[0062] Specifically, for example, calibration signal generator 21 generates a multicarrier signal as the calibration signal, which is a combination of multiple carrier signals with different frequencies within a frequency range equal to or less than half the sampling frequency. Each carrier signal has a different frequency from any alias included in the sample signal. This configuration uses a multicarrier signal instead of a conventional unmodulated CW signal, and by distributing multiple carrier signals over a wide range to avoid the effects of aliasing, it is possible to simultaneously detect characteristic differences at multiple frequency points and reduce the need to switch the frequency of the calibration signal, thereby enabling high-speed calibration.

[0063] More specifically, the multi-carrier signal is, for example, a signal having a frequency range equal to or less than half the sampling frequency, the Nyquist frequency f s Each Nyquist zone R defined by / 2m iThe same number of carrier signals are arranged at equal intervals d within each Nyquist region R i The frequency difference Δf between the minimum frequency of the carrier signal with the lowest frequency among the carrier signals arranged within the Nyquist zone and the frequency at one end of the Nyquist zone to which the carrier signal with the lowest frequency belongs i is the value of each Nyquist zone R i However, f s is the sampling frequency, m is the number of ADCs, and the Nyquist region R i (i=1,2,...,m) is (f s / 2m)·(i-1) or more and (f s / 2m)·i. Also, Δf i <dとする。

[0064] This configuration makes it possible to distribute multiple carrier signals almost evenly across the entire frequency range below half the sampling frequency, i.e., the entire frequency range in which the original signal can be reproduced based on the sampling theorem, while avoiding the effects of aliasing. This allows for accurate and efficient measurement of the frequency characteristics of mismatch between ADCs.

[0065] In this embodiment, when an arbitrary waveform signal generator is used as a calibration signal generator to generate a multicarrier signal based on an arbitrary waveform, a signal pattern can be easily generated by using FFT (Fast Fourier Transform) and DFT (Discrete Fourier Transform). At this time, the frequency at which the carrier signal is to be placed can be selected arbitrarily, and the phase of each carrier signal can also be adjusted to suppress the crest ratio.

[0066] In this embodiment, when mismatch characteristics are simultaneously acquired across multiple Nyquist zones, the carrier allocation of the calibration signal is set so that the frequency of the carrier signal and the frequency of aliasing noise (alias) that is in an alias relationship with the carrier signal do not overlap with each other. As an example, the carrier allocation that avoids the influence of aliasing can be determined as follows.

[0067] The frequencies of the multiple carrier signals that make up the multi-carrier signal may be set, for example, according to the following equation:

number

[0068] frequency f k If the carrier group (series) is extended to the second Nyquist zone or beyond according to the aliasing relationship, carriers with different variables k will not be in an aliasing relationship and will not be affected by aliasing. For example, k The multicarrier signal can be determined by setting the carrier signals belonging to the (k+1)th Nyquist zone to be included in the (k+1)th Nyquist zone.

[0069] As an example, Fig. 4 shows an example of carrier allocation when m=4 and n=4. Fig. 4(a) shows the frequency spectrum of a multi-carrier signal used as a calibration signal, and Fig. 4(b) shows the frequency spectrum of a sample signal output from a time-interleaved ADC. In Fig. 4, the carrier signals for k=0, 1, 2, and 3 are represented by solid lines in black, light gray, medium gray, and dark gray, respectively. As shown in Fig. 4(a), f k=0 The series of carrier signals are arranged in the second Nyquist zone, k=1 The carrier signal of the series is placed in the third Nyquist zone, and k=2 The carrier signal of the series is arranged in the fourth Nyquist zone, k=3A series of carrier signals are arranged. As shown in Figure 4(b), the carrier signal frequencies are set so as not to overlap with the alias frequencies indicated by the dotted lines. The mismatch characteristics are acquired for frequency points where the actual signal exists.

[0070] With the above configuration, the carrier signals constituting the calibration signal can be arranged so that the carrier signals and their aliases do not overlap in frequency. Furthermore, the mismatch characteristics between ADCs can be acquired simultaneously across multiple Nyquist zones. This allows accurate and efficient measurement of the frequency characteristics of the mismatch characteristics between ADCs.

[0071] The mismatch characteristics between ADCs may be measured while changing the frequency range in which the multiple carrier signals are allocated, so that the frequency characteristics of the mismatch characteristics are measured over the entire frequency range equal to or less than half the sampling frequency. With this configuration, even if the frequency range in which the time-interleaved ADC 10 is expected to be used is wide, the frequency characteristics of the mismatch characteristics between ADCs can be measured over the entire frequency range.

[0072] The mismatch characteristics may include at least one of a relative amplitude ratio, a relative phase difference, and a DC offset of each ADC relative to one of the ADCs, which allows for effective correction of mismatches between the ADCs as needed.

[0073] As shown in FIG. 5 , the ADC core 30 includes multiple sub-ADCs 320, 321, etc. The time-interleaved ADC 10 may include one or multiple ADC cores. The sub-ADCs or ADC cores perform sampling operations at equal intervals in a predetermined order. Configuration information for the time-interleaved ADC 10 may be stored in a storage unit included in the correction information calculation unit 23 or in another storage unit. The correction information calculation unit 23 may calculate mismatch characteristics between ADC cores, mismatch characteristics between sub-ADCs, or both based on this configuration information. In this way, even if the time-interleaved ADC 10 has a hierarchical structure including ADC cores and sub-ADCs, the correction information calculation unit 23 can calculate mismatch characteristics and mismatch correction information by taking the hierarchical structure into consideration.

[0074] FIG. 6(a) is a diagram showing a connection between the time-interleaved ADC 10 and the calibration signal generator 21 when the time-interleaved ADC 10 includes a single ADC core 30, and FIG. 6(b) is a diagram showing a connection between the time-interleaved ADC 10 and the calibration signal generator 21 when the time-interleaved ADC 10 includes two ADC cores 300 and 301. As shown in FIG. 6(a), when the time-interleaved ADC 10 includes a single ADC core 30, a calibration signal is input to the ADC core 30. As shown in FIG. 6(b), when the time-interleaved ADC 10 includes two ADC cores 300 and 301, the time-interleaved ADC 10 includes a power divider 40, to which a calibration signal is input, and which distributes the calibration signal to each of the ADC cores 300 and 301. Although FIG. 6(b) shows a case where there are two ADC cores, the number of ADC cores may be any integer greater than two.

[0075] The calibration signal generator 21 may adjust the crest ratio of the sample signal by adjusting the phase of each carrier signal. This configuration makes it possible to suppress the crest ratio.

[0076] (Calibration method) Next, a calibration method will be described. In the following description, it is assumed that the time-interleaved ADC 10 includes one or more ADC cores, and each ADC core includes multiple sub-ADCs.

[0077] The purpose of the calibration is to obtain the frequency characteristics of the mismatch characteristics between sub-ADCs or between ADC cores, and to obtain the correction value when performing mismatch correction.

[0078] Specifically, the calibration involves the following two steps: (i) Obtain the frequency characteristics of the mismatch characteristics between sub-ADCs or between ADC cores. (ii) From the frequency characteristics of the acquired mismatch characteristics, various mismatch correction values ​​used during time interleaving operation are calculated.

[0079] More specifically, the calibration signal generator 21 generates a calibration signal and inputs it to the time-interleaved ADC 10. The calibration signal is a multi-carrier signal obtained by combining multiple carrier signals with different frequencies within a frequency range equal to or less than half the sampling frequency, and each carrier signal has a different frequency from any alias included in the sample signal. The time-interleaved ADC 10 outputs a sample signal by performing AD conversion on the calibration signal at a predetermined sampling frequency. The sample signal storage unit 22 stores the sample signal. Next, the correction information calculator 23 calculates the frequency characteristics of the mismatch characteristics between the sub-ADCs or the ADC cores from the sample signal stored in the sample signal storage unit 22, and calculates correction information for correcting the mismatch between the sub-ADCs or the ADC cores from the frequency characteristics of the mismatch characteristics.

[0080] When a single-tone unmodulated CW signal is used as the calibration signal as in the past, the calibration signal circuit can be simplified and high measurement accuracy and stable results can be obtained, but there is a problem that the measurement takes a long time because measurements are repeated for each frequency point, which is particularly noticeable when the frequency resolution is high over a wide band.In contrast, when a multicarrier signal is used as the calibration signal according to the embodiment of the present invention, (i) calibration can be performed quickly by simultaneously measuring a wide frequency range, (ii) there is no restriction (or the restriction is small) that measurements cannot be made at the boundaries of the Nyquist zone, and (iii) relatively high measurement accuracy and stable results can be obtained.

[0081] Furthermore, calibration signal generator 21 according to this embodiment may be configured to generate, as a calibration signal, a multi-carrier signal obtained by combining multiple carrier signals having different frequencies within a frequency range equal to or greater than half the sampling frequency but equal to or less than the sampling frequency, where each carrier signal is set to have a different frequency from any of the aliases contained in the sample signal. Thus, even in the case of a frequency range exceeding the Nyquist frequency, i.e., undersampling, by allocating multiple carrier signals over a wide range to avoid the effects of aliasing, as in the case of allocating multiple carrier signals within a frequency range equal to or less than the Nyquist frequency, it is possible to simultaneously detect characteristic differences at multiple frequency points and reduce the need to switch the frequency of the calibration signal, thereby enabling high-speed calibration.

[0082] In addition, in implementing the present invention, it is also possible to configure the present invention so that "half or less the sampling frequency" is "less than half the sampling frequency," "more than half the sampling frequency" is "half or more the sampling frequency," and "less than or equal to the sampling frequency" is "less than the sampling frequency." [Industrial Applicability]

[0083] As described above, the present invention has the effect of enabling high-speed calibration of a time-interleaved ADC, and is useful for ADC calibration devices and ADC calibration methods in general. [Explanation of symbols]

[0084] 10 Time-interleaved ADC (Analog-to-Digital Converter) 11 Signal splitter 12 ADC (Analog-to-Digital Converter) 13 Sampling control section 14 Signal switch 20 ADC calibration device (calibration device) 21 Calibration signal generator 22 Sample signal storage unit 23 Correction information calculation unit 30 ADC cores 31, 40 Power divider 32 sub ADC

Claims

1. A calibration device for an AD conversion device that operates a plurality of AD converters in a time interleaved manner, comprising: a calibration signal generator (21) for generating a calibration signal to be input to the AD conversion device; a sample signal storage unit (22) for storing a sample signal obtained by AD converting the calibration signal at a predetermined sampling frequency in the AD conversion device; a correction information calculation unit (23) that calculates, from the sample signals stored in the sample signal storage unit, frequency characteristics of mismatch characteristics that characterize mismatches between the plurality of AD converters, and calculates correction information for correcting mismatches between the plurality of AD converters from the frequency characteristics of the mismatch characteristics; Equipped with The calibration signal generator generates a multi-carrier signal as the calibration signal by combining multiple carrier signals having different frequencies within a frequency range not exceeding half the sampling frequency, and each of the carrier signals has a frequency different from any of the aliases contained in the sample signal.

2. A calibration device for an AD conversion device that operates a plurality of AD converters in a time interleaved manner, comprising: a calibration signal generator (21) for generating a calibration signal to be input to the AD conversion device; a sample signal storage unit (22) for storing a sample signal obtained by AD converting the calibration signal at a predetermined sampling frequency in the AD conversion device; a correction information calculation unit (23) that calculates, from the sample signals stored in the sample signal storage unit, frequency characteristics of mismatch characteristics that characterize mismatches between the plurality of AD converters, and calculates correction information for correcting mismatches between the plurality of AD converters from the frequency characteristics of the mismatch characteristics; Equipped with The calibration signal generator generates a multi-carrier signal as the calibration signal by combining a plurality of carrier signals having different frequencies within a frequency range of more than half the sampling frequency and less than or equal to the sampling frequency, and each of the carrier signals has a frequency different from any of the aliases contained in the sample signal.

3. Within a frequency range equal to or less than half the sampling frequency, the Nyquist frequency f of each AD converter is ｓ Each Nyquist region R defined by / 2m ｉ The same number of carrier signals are arranged at equal intervals within each Nyquist region R ｉ The frequency difference between the minimum frequency of the carrier signal with the lowest frequency among the carrier signals arranged within each Nyquist zone R and the frequency at one end of the Nyquist zone to which the carrier signal with the lowest frequency belongs is ｉ are different from each other, except that f ｓ is the sampling frequency, m is the number of AD converters, and the Nyquist region R ｉ (i=1, 2, ..., m) is (f ｓ / 2m)·(i-1) or more and (f ｓ 2. The calibration device of claim 1, wherein the frequency range is less than 1 / 2m.

4. 2. The calibration device according to claim 1, wherein the frequencies of the plurality of carrier signals constituting the multicarrier signal are set according to the following equation: [Equation 1] (where f ｓ is the sampling frequency, m is the number of AD converters, and n is a parameter that determines the fineness of the carrier interval. ｋ = 0 is excluded.)

5. 2. The calibration device according to claim 1, wherein the mismatch characteristics between the AD converters are measured while changing the frequency range in which the plurality of carrier signals are arranged, and the frequency characteristics of the mismatch characteristics are measured over the entire frequency range equal to or less than half the sampling frequency.

6. 3. The calibration device according to claim 1, wherein the mismatch characteristics include at least one of a relative amplitude ratio and a relative phase difference of each of the plurality of AD converters with respect to one of the AD converters as a reference, and a DC offset.

7. 3. The calibration device according to claim 1, wherein the AD conversion device includes one or more AD converter cores, each of the AD converter cores includes a plurality of sub-AD converters, and the correction information calculation unit calculates at least one of the frequency characteristics of the mismatch characteristics between the AD converter cores and the frequency characteristics of the mismatch characteristics between the sub-AD converters.

8. 3. The calibration device according to claim 1, wherein the calibration signal generator adjusts the crest ratio of the sample signal by adjusting the phase of each of the carrier signals.

9. A method for calibrating an AD conversion device in which a plurality of AD converters are operated in a time interleaved manner, comprising: a calibration signal generating step of generating a calibration signal to be input to the AD conversion device; a storage step of storing a sample signal obtained by AD converting the calibration signal at a predetermined sampling frequency in the AD conversion device in a sample signal storage unit; a correction information calculation step of calculating, from the sample signals stored in the sample signal storage unit, frequency characteristics of mismatch characteristics that characterize mismatches between the plurality of AD converters, and calculating correction information for correcting mismatches between the plurality of AD converters from the frequency characteristics of the mismatch characteristics; Including, In the calibration signal generating step, a multi-carrier signal is generated as the calibration signal by combining a plurality of carrier signals having different frequencies within a frequency range of not more than half the sampling frequency, and each of the carrier signals has a frequency different from any of the aliases contained in the sample signal.

10. A method for calibrating an AD conversion device in which a plurality of AD converters are operated in a time interleaved manner, comprising: a calibration signal generating step of generating a calibration signal to be input to the AD conversion device; a storage step of storing a sample signal obtained by AD converting the calibration signal at a predetermined sampling frequency in the AD conversion device in a sample signal storage unit; a correction information calculation step of calculating, from the sample signals stored in the sample signal storage unit, frequency characteristics of mismatch characteristics that characterize mismatches between the plurality of AD converters, and calculating correction information for correcting mismatches between the plurality of AD converters from the frequency characteristics of the mismatch characteristics; Including, In the calibration signal generating step, a multi-carrier signal is generated as the calibration signal by combining a plurality of carrier signals having different frequencies within a frequency range of more than half the sampling frequency and less than the sampling frequency, and each of the carrier signals has a frequency different from any of the aliases contained in the sample signal.

Images