TWI337458B - Converting non-equidistant signals into equidistant signals - Google Patents

Description

Nine, invention description:

[Course of the Ming Dynasty is a branch of the Xuan Xuan staff J This invention is about signal processing technology. C «tr 43⁄4. ]3

US 6,462,693 B1, filed by the same applicant, Agilent Technologies, discloses the conversion of a class of analog signals into a plurality of digital signals. The method includes the steps of comparing an amplitude value in the analog signal with a plurality of reference amplitude values to determine whether the analog value is greater or less than a reference value. The method further includes the step of generating a logical level evd in a digital signal corresponding to the decision in the comparing step. The method converts the sig-class § § into a time representation and then converts the time representation into a one-digit representation. The device includes a plurality of comparators, each comparator being coupled to receive the analog signal, separately receiving different values of the reference values, and generating the digital signal. The class is decremented from the digit representation. Us 6' 429' 799 ι, filed by the same applicant, Agilent Technologies, discloses a representation of converting a class of analog signals into a single digit. The hen method includes the steps of: generating a plurality of time-varying reference signals 'compacting an amplitude of the analog-to-number with an amplitude of each of the reference signals to determine whether the analog apostrophe amplitude is greater than or less than Or equal to the reference signal amplitude, and each time the analog signal amplitude is equal to the amplitude of the reference signals, a time stamp is generated. The apparatus includes a reference signal generator and a plurality of comparators, each of the comparators being coupled to receive the analog signal, respectively receiving a different one of the reference signals and generating a digital signal. The analog signal can be reconstructed from the digital representation. US 2004/0070529 A1, filed by the same applicant, Agilent Technologies, discloses pre-processing an analog signal and converting the pre-processed signal into a digital representation. The method includes the steps of: preprocessing the analog signal to generate a plurality of reference signals, comparing an amplitude of the preprocessed signal with an amplitude of the reference signals to determine whether the preprocessed signal amplitude is greater than , less than or equal to the reference signal amplitude, and a time stamp generated when the preprocessed signal amplitude and the reference signal amplitudes are equal. The device includes a pre-processor, a reference signal generator, and a plurality of comparators. One of the comparators receives the preprocessed signal from the preprocessor, receives a reference signal, respectively, and generates a digital

Is number. The preprocessed signal or the analog signal can be reconstructed from the digital representation. S·Sunter in IEEE International Test in 2004

Conference Pages 95-104 “An Automated, Complete, Structural

Test Solution for SERDES' reveals the jitter of the 1 billion Hz serialization and reduction sequence (SERDES) that has become the main inter-wafer and interboard data transmission technology. Signal integrity is the primary factor in determining its bit error rate (typically less than 1〇7), so the main production test challenge is to test the jitter jitter and signal open eye (signaUye 〇pening). Out-of-chip jitter and rise/fall time measurements are limited by hardware complexity, access, bandwidth, and noise. Published in-wafer measurement techniques are limited by delay line jitter. This paper presents a new dithering technique that has been demonstrated to achieve RMS self-jittering less than lps and a new signal eye test with unbalanced bandwidth; neither test is used High speed circuit. The all-digital technique uses the receiver itself to demodulate the signal jitter to a low-speed bit stream that is analyzed by a single-clock domain synthesizable circuit. This is combined with the logical BIST and 1149.6 boundary scans to fully test a 1C. However, there is still a need for efficient signal processing. |[考务明内j SUMMARY OF THE INVENTION One object of the present invention is to enable efficient signal processing. This purpose is solved by a separate item of the patent application scope. Further embodiments are shown in the dependent claims of the scope of the patent application. In accordance with an exemplary embodiment of the present invention, an L-number processing device for processing a signal is provided, the signal processing device including a comparator unit for comparing the signal with a reference signal, and a a unit for generating a digital result signal of the comparison result, an evaluation unit for determining the correspondence between the digital result representations, and - for calculating a substantially evenly spaced output (based on, for example, a non-equidistant digital result signal) The output signal calculation unit of the signal. According to another exemplary embodiment, a measuring device is provided 'the metering device comprises a signal processing device having the above features, the signal processing device for processing a measurement related to the measurement performed by the measuring device (eg, repeating )signal. According to another exemplary embodiment, a signal processing method for processing a signal is provided. The method includes the steps of: comparing the signal with a reference signal 1337458, generating a digital result signal indicating the comparison result, and determining the The transition time of the equal-numbered result signals, and the calculation of substantially evenly spaced output signals (eg, based on the digital result signals). According to another exemplary embodiment, a computer readable medium is provided, 5 a computer program for processing a signal stored in the computer readable medium, the computer program being adapted to control or execute the above when executed by a processor method. In accordance with yet another exemplary embodiment, a program element for processing a signal is provided that, when executed by a processor, is adapted to control or perform the above method. Embodiments of the invention may be embodied or supported in part or in whole by one or more suitable software programs, which may be stored in any type of data carrier or may be provided by any type of data carrier, and It may or may be executed in any suitable data processing unit or by any suitable data processing unit. A software program or routine can be preferably used to process a signal. Signal processing in accordance with an embodiment of the present invention may be performed by a computer program (ie, software) or by one or more specific electronic optimization circuits (ie, hardware) or in a hybrid form (ie, using software components and hardware components) . In accordance with an exemplary embodiment, the 20 non-equidistant signals obtained in any signal processing scheme can be tuned to produce signals that are substantially equidistant (e.g., in the time domain). Thus, a uniform signal reconstruction scheme can be provided that allows for the provision of an output signal of a certain format that is slightly "normalized" (for a time series of one of the signals) and thus for many post-processing schemes It is appropriate. 8 1337458 According to an exemplary embodiment, a signal having a certain periodicity (so that it is repeated several times) is processed by low sampling. In this context, the phrase "low sampling" may refer to a sampling scheme in which a signal is sampled at a sampling rate less than twice the bandwidth of the signal. The repetitive signal is compared to a comparison or reference signal at these points in time after it has been determined at a number of time points on which the sample is to be performed by 5, for example substantially equidistant. The result of this comparison may be that the repetitive signal is greater than the reference signal, less than the reference signal, or equal to the reference signal. Therefore, a certain digital result signal can be generated to indicate a result of this comparison. Information about the repeated signal may be "encoded" at the 10th time point at which the digital result signal transitions from "0" to "Γ, or vice versa. In other words, regarding the need to be processed A message of the signal can be included at a point in time at which a transition of the logical value of the comparison occurs. The analog signal can be reconstructed from the digit representation. 15 Performing a so-called "coherent sampling" Coherent sampling may be advantageous, that is to say a defined time dependency between the operation of the individual components of the signal processing device, in particular the unit generating the reference signal and the decision signal. This homology can be obtained, for example, by providing a common clock signal to such elements, or by strictly joining or coupling the timing of such elements, e.g., using a lock 20 or the like. More specifically, a low-sampling analog signal can be performed using a homophone automatic test equipment (ATE) channel. For test electronics, particularly integrated electronic circuits that provide digital electronic output signals, a test or stimulus signal can be fed to an input to be tested by the DUT and from the device under test. The response signal can be evaluated by an automated test device, for example by comparison with the desired data. This automated test equipment may already include a specific test function, which is the test function or routine that the test equipment can perform. This test function can be incorporated into a test device in the form of an executable software code. When a test signal (possibly a repetitive signal) is transmitted within this electronic test set, it may also happen that the signal is converted between analog and digital formats. In order to perform this conversion, a low sampling feature in accordance with an exemplary embodiment can be implemented. It may be beneficial from the fact that such stimulation signals and/or response signals may have a specific 10 periodicity, i.e., may be a repetitive signal. The exemplary embodiment can transform the results of this data processing into a "standard" analog to digital converter (ADC) result format. For example, after the repeated signals are low-sampled at different time points in the repeated signal period, the data points are rearranged so that each new iteration 15 of the repeated signal is obtained between the data points. Inserting may be advantageous. However, by performing this rearrangement, the equidistant characteristics of the signals may be affected, so the output signals may be non-equidistant. Embodiments of the present invention can accept such non-equidistant signals and post-process them such that substantially equidistant output signal sequences are obtained, such as substantially at equal intervals in time or frequency. In particular, embodiments of the invention may be implemented in a analog to digital converter (ADC) where an analog signal is digitized. In other words, an analog waveform can be converted to a digital signal. In this context, a repeated signal can be sampled low. In this context, the phrase "analog" may refer to a signal having a continuous level (e.g., a modulated radio frequency signal). The word "digit" can be used to indicate a signal having a discrete level (e.g., a logical value of "1" or ‘‘〇,,). An exemplary field of application of embodiments of the present invention is "automatic test setup, ie, digitization of analog test signals in testing a device under test (DUT). Because this test is reversible, low sampling one The concept of a repetitive signal can be applied. - The wording "coherent sampling" can be specifically understood as "low sampling, the derivative form". "Low sampling" can mean that sampling is performed in a slower manner or at a slower sampling rate than is necessary to obtain all the necessary conditions when there is no repetition. In the "coherent sampling" environment, a defined ratio between the sampling period and the signal repetition can be ensured or adjusted. When performing this low sampling, it is preferred to use coherent sampling, and the information to be obtained is encoded into the transition time of the digital comparison signal. This signal is analyzed multiple times, for example twice. For each iteration, a new intermediate measurement point is obtained to continuously refine the measurement 'digital low sample, which can be performed, with the goal of measuring 15 transition time or time stamp with increased accuracy. For the coherent measurement of the transition time, the repetitive signal is sampled a plurality of times, and the -divide ratio between the signal frequency and the sampling rate can be selected. With this measurement, intermediate sample points can be generated, which increases the accuracy of the determined transition time. By repeating the waveform many times, and by selecting a sufficiently high sampling rate, the accuracy can be particularly increased. According to an exemplary embodiment, an analog-to-digital conversion is provided to implement a coherent interpolation for measuring the transition time of the reconstructed-analog waveform. In particular, such an ADC can be advantageously implemented in a measuring device such as a test device for testing a device under test (DUT). A control clock unit can provide a clock signal that can be used globally by the signal processing device to enable coherent time stamping. Coherent sampling increases the accuracy of sampling at a determined point in time. Thus, according to an exemplary embodiment, low sampling (or some digitization) of the inverse analog signal (including the modulated RF signal) can be performed. These signals may in particular have a frequency range between 1 00 MHz and 10 GHz. For example, as a modulation scheme, amplitude modulation, frequency modulation, or face modulation can be used. This low sampling can be included in the field of automatic test equipment (ATE), where mixed signals / The RF test is performed and the signal can be repeated here. Therefore, any technical field in which a repetitive signal appears may be a potential field of application of the exemplary embodiment. In many cases, the test criteria assume uniform access to spatial samples from a conventional ADC, or assume access to the signal spectrum. The wording "digit" means two levels that are independent of time. The wording "analog" means a level that is independent of time. The analog signal can include the modulated RF L number. The phrase "coherent" means an N/M frequency or period ratio, where n can be the number of sample points and M can be the inverse number of signals that need to be sampled' and where: N and Μ are mutually prime. According to an exemplary embodiment A signal processing device can be provided, which is evaluated by a repeated analog nickname. For this analog signal, a comparison with one-level or several known levels or any other waveform can be performed. The result of the 乂 乂 can be a digital signal. The digital comparison signal can then be sampled by homology to obtain a digital sample. Next, the precise transition can be determined by comparing the digital samples of the chopping. The level of the known waveform in time, the sample can be selected, can be repeated (four) and ==, multiple static levels, sine waves, other dynamic waveforms or pre-processed signals. May use the -digit ATE channel as Coherent sampler. In addition, jitter 5 can be selectively added to the signal to improve resolution/fine rhyme. When detecting transitions, 'before the transition', the signal can be determined before the transition. "Finally, the signal can be determined, and an "intermediate" time between the last point and the first point can be determined as a 'or-' count, which can be measured before and after the transition. An interval between them is executed. 1 (4) Another - Exemplary Embodiment '- Analog signal can be processed by the signal processing device. The analog signal can be compared to known levels. Then, the transition time of the comparison signal can be determined. Samples can be constructed. These samples can be converted to a common "ADC" format. In this context, an interpolation can be performed 'non-uniform fast Fourier transform _Fft' can be performed and/or 15 non-uniform discrete Fourier transform (NUDFT) can be performed. In the NUDFT scenario, a direct down-conversion of the (modulated)!^ signal may be performed. It is also possible to perform an offset correction by including the Dc item in the NUDFT. By including a known parasitic frequency in the :UDFT, parasitic suppression (10) (4) Qing (10) can be caused by the month b. In addition, 'noise suppression can be achieved. It is possible to use equally spaced frequencies in a single continuous frequency range (e.g., within a conventional cartridge). According to an exemplary embodiment, the RF reception $ using the digital ATE channel can be provided. In this context, the _ (modulated) narrowband (rf) signal can be used with / and can be compared to one (or more) static levels. A digital ATE channel 13 1337458 can be used for coherent sampling. The exact span time can be determined. A narrow spectrum can be reconstructed (NUDFT). This result can be downconverted by a simple frequency shift. Next, further exemplary embodiments of the present invention will be explained. 5 In the following, a further exemplary embodiment of the signal processing device will be explained. However, these embodiments are also applicable to measurement devices, methods, program components, and computer readable media. The signal processing device can be adapted to process a repeated analog signal. Such a ratio signal can represent any continuous value. The wording can have the following meaning repeatedly, the 10 signal has some periodicity and is repeated multiple times. Additionally, the signal processing device can be adapted to process a repetitive modulated radio frequency (RF) signal. In particular, this signal processing can be performed in the case of an RF receiver. The signal processing device can be adapted to be an analog to digital converter 15 (A D C ) for converting a reverse analog signal into a digital signal. Implementing this coherent sampling in this conversion allows the conversion to be performed with high precision. Recalculating the resulting signal so that the resulting signal at equidistant intervals simplifies the post processing of the converted signal. The decision unit can be adapted to determine the point in time for each of the multiple repetitions of 20. When the different periods of the repetitive signal are compared based on different iterations, the sampling points of the subsequent period may be displaced relative to each other. By taking this approach, the intermediate points of the sample can be obtained, so that the analysis is refined. The decision unit is adapted to determine each of the different time points for the plurality of iterations such that substantially evenly spaced time data is obtained. By selecting 14 1337458, the interval between the sampled point and the number of transitions is set to ϋ, and an effective method of this is also entered. (10) every time - repeatedly get the middle point

... is one of a group consisting of a single-signal that is constant over time, a complex number of signals per change Γ =, a basis - a predetermined waveform in time: ... and a sinusoidal signal. Iron: Any other waveform of the cow is possible, such as any trigonometric function, - mineral tooth function, a step function, and so on. 10 15

20, used to test and a number of repeated signals on the same correlation, determine the transition time. Therefore, different repetitive samples (4) can be combined to increase the accuracy, and special attention is obtained to obtain the ## between the two materials '. This refines the data format conversion. The number of time points can be greater than the number of times. The ratio between the number of time points and the number of times of the reverse may be - non-residual value. The latter approach ensures that additional feedback provides additional information about the signal. Advantageously, the signal processing means may comprise an arranging unit adapted to align the digital result signals associated with the different replies by means of a one-year chronology at a corresponding point in time in the repetitive signal period. After rearranging or reordering the signals after low sampling, several points in the case of a digital result signal transition can be refined. The generating unit may be adapted to generate a digital result signal having a first logical value (eg, "1") (in this case, each of the repeated signals is greater than the reference signal) and is adapted to generate a second logical value ( For example, "〇") the digital result signal (in this case the respective repeated signals are smaller than the reference signal). As a result of 15 1337458, the transition time can be calculated as a point in time at which the transition from the first logical value to the second logical value occurs. For this purpose, the evaluation unit can determine the transition time based on an analysis of the last time point in a sequence in which the digital result signal has a first logical value. Additionally, a first time point may be determined as a transition time in a sequence in which the digital result signal has the second logic value. It is also possible to calculate an average time between the last time point of the first logical value and the first time point of the second logical value. In addition, by performing a statistical analysis of the transition interval, it is possible to refine the accuracy of the transition time in a further step. For example, due to effects such as jitter, noise, or other signal distortion, it may occur that a sequence of interaction values of the first logical value and the second logical value occurs around a transition time. In this scenario, it is more reasonable to statistically analyze the data points in this area and calculate a (most) possible transition time. This can include interpolation, statistical averaging, fitting a mathematical 15 probability function to a determined digit value (some thermometer code), such as a cumulative probability density function, and so on. The evaluation unit can be adapted to determine the transition times in accordance with an operational mode that is selected based on the degree of signal jitter or the presence of signal noise. When signal jitter/noise is large, it is relatively possible that signal distortion occurs and the transition 20 interval contains some distorted measurement points. When the signal jitter is small, a numerically easier solution is preferred, that is, using the last time point of the first logic value or the first time point of the second logic value or an average thereof is an actual Transition time. The signal processing device can include an output signal calculation unit, the output 16 1337458 signal calculation unit being adapted to calculate substantially evenly spaced output signals. By this reordering, equidistance can be achieved. However, it may be desirable for post processing of the signals that are substantially evenly spaced, particularly evenly spaced over time or evenly spaced within frequency. 5 in order to obtain such a uniformly spaced signal, the output signal calculation unit may be adapted to use a mathematical function formed by using a Sine interpolation (ie, a ratio of the sine of the argument to the argument, wherein The argument can be time), a polynomial interpolation (such as a Lagrange interpolation, the same strip interpolation, or a linear interpolation) to determine the substantially evenly spaced output signal. It is also possible to implement a fractional delay filter. The output signal calculation unit is applicable to perform non-uniform fast Fourier transform (NUFFT) with a fast multi-pole transform (FFT), a non-uniform fast Fourier transform (NUFFT), and a fast multi-pole method (FMM), and a At least 15 of the groups consisting of non-uniform discrete Fourier transforms (NUDFT) are used to calculate the output signal. In particular, the signal processing device can be adapted to process the repetitive signal using coherent sampling. Coherent sampling indicates that the time characteristics of the sampling are well established. Coherent sampling can be enabled by implementing a clock generating unit adapted to generate a 20 common clock signal for a plurality of elements of the signal processing device. For example, the clock generating unit can provide a common clock signal to a device under test, a reference signal generator for generating the reference signal, and a unit for evaluating or processing the comparison signal. Therefore, the clock generating unit can generate a common clock signal to some or all of the components of the signal processing device. 17 1337458 Additionally, a plurality of clock generating units can be provided, each adapted to generate an additional clock signal to the designated component of one of the signal processing devices, wherein the clock generating units can be frequency locked. For example, three lock frequency generators having the same output frequency can be provided to generate a separate clock signal for the device under test, a reference signal generator, and a unit for evaluating the comparison signal. In addition, three independent clock generators with sufficient accuracy to ensure the desired output frequency can be implemented. The signal processing device further includes a jitter increasing unit adapted to selectively increase the sway to the repetitive signal and/or the comparison signal and/or the coherent sampling clock. By doing this, the signal can be smeared out in a defined manner to increase the accuracy of the signal conversion. The signal processing device can include a pair of DUTs providing an environment for an automated test equipment unit (ATE) such that the DUT is a source for providing the repetitive signal, such as an Agilent 93000 test device. Thus, the signal processing device can be implemented in particular as an ADC for testing the ATE of a device under test (e.g., a wafer of a mobile phone). The automatic test equipment can provide an environment to the DUT such that the DUT is a source for providing the repetitive signal and can be implemented for testing a device under test. In the following, further exemplary embodiments of the measuring device will be explained. However, these embodiments are also applicable to the signal processing apparatus, the method, the program element, and the computer readable medium. A signal generating unit of the measuring device is adapted to generate an stimulating signal 18 1337458 (as a basis for the repetitive signal or the repetitive signal), the stimuli signal being applied to a device under test to test the device under test. It is also possible to generate a repetitive signal based on a (repetitive or non-repetitive) stimulation signal of the measuring device. The stimulus signal can be any signal pattern applied to the pins of a DUT (e.g., a wafer under test) 5, and a response signal can be detected on the other pins of the DUT. By comparing these response signals with the desired signals, it can be determined whether the device under test is acceptable or needs to be rejected. When the test sequence is applied to the device under test, the signals are usually repeated several times. Thus, a signal processing device in accordance with an exemplary embodiment can be advantageously implemented in such a measurement device, particularly in the case of an ADC used in such a measurement device. The measuring device can be further adapted to receive a response k number (as the repetitive signal) from a device under test in response to applying a stimulus signal to the device under test to test the device under test. In this scenario, the device under test can generate a repeat signal that can then be evaluated by a signal processor of the measuring device. More particularly, the measuring device can include an analog to digital converter (ADC), a sensor device (eg, a parameter for sensing a dut), a test device or device for testing Test device (such as a device of the Agilem 93(10) 20 series), a device for chemical, biological and/or pharmaceutical analysis, a fluid separation system suitable for separating a fluid mixture, a capillary electrophoresis device, a liquid chromatography At least one of a group consisting of a device, a gas chromatography device, an electronic measuring device, and a mass spectrometer. More generally, embodiments can be used in many fields of electronics and measurement applications, such as 19 in a life sciences system, or in any 7-bay domain that can be used for analog or digital electronics, in such fields. Accurate signal conversion or signal processing is a question 'especially when a repeated signal is used. BRIEF DESCRIPTION OF THE DRAWINGS Other objects and many additional advantages of the embodiments of the present invention will be apparent from the description and appended claims Features that are substantially or functionally equivalent or similar are denoted by the same reference symbols. Fig. 1 shows a signal processing apparatus in accordance with an exemplary embodiment of the present invention. Figure 2 shows the schema in the case of a coherent time stamp. Figure 3 illustrates a signal processing apparatus in accordance with an exemplary embodiment. Figure 4 shows a diagram illustrating the function of a signal processing device when there is jitter. Figure 5 shows a diagram illustrating an operational mode of a data processing device when there is jitter. Figure 6 shows a diagram during an operation of a signal processing device for reconstructing an analog waveform in accordance with an exemplary embodiment. Fig. 7 is a view showing a data processing apparatus according to an exemplary embodiment of the present invention. Figure 8 illustrates a uniform interpolation in the case of a signal processing device in accordance with an exemplary embodiment. Figure 9 shows a measuring device in accordance with an exemplary embodiment. The illustrations in the drawings are schematic. (1) 7458 [Embodiment of the present invention] A detailed description of the preferred embodiment will be explained hereinafter with reference to Fig. 1 a signal processing apparatus 100 according to an exemplary embodiment of the present invention. 5 A repetitive signal source 101 is adapted to generate a periodic repetitive signal 102. The inverted signal 102 is a periodic signal that is repeated several times. This repetitive signal 102 is used in the case of studying a device under test. The repeated signal 102 is provided to an input of the decision unit 103. The decision unit 103 is adapted to determine a number of time points for the low sampling of the repeated signal 102. The time points at which the repetitive signal 10 10 2 is sampled may be supplied to the flip-flop [3] in the form of a control signal 104, the flip-flop 130 being located at an output of a comparator unit 1 〇5. A comparator signal 120 can be provided at the output of the comparator unit 105 to indicate a result of the comparison. The inverted signal 102 is provided to a first signal input of the comparator unit 105. A reference signal 106 generated by a reference signal generator unit 15 107 is supplied to a first k-number input of the comparator unit 105. As can be seen in FIG. 1, a clock generating unit 1A8 generates clock signals 109a, 10b, which are separately supplied to the inverted signal source 10 and the reference signal generator 107. . In this case, it can be ensured that the repetitive signal source 101 and the reference signal generator 107 20 are synchronized to enable coherent sampling. The comparator unit 105 is adapted to compare the repeated signal 1〇2 with the reference signal 106. The result of this comparison is provided at the output of one of the comparator units 105 coupled to one of the inputs of the flip-flop 130. The comparison signal 120 provided at the output of the comparator unit 105 is provided to the positive and negative 21 1337 458 finder 130' and takes into account the control signal 1 indicating a number of specific comparison times, the flip flop 130 producing a digital result The signal is 丨1〇. The digital result signal (4) is supplied to the rearrangement unit lu, and the rearrangement unit lu is adapted to be used according to the time point corresponding to the time point defined by the time control signal 1 〇4 within the inverse 贱K) 2 Representative, and the digital result signals that are inversely related to the repeated signal (10) are rearranged (see Figure 2). After the components of the signal 110 have been reordered, they are provided to φ an evaluation unit 112 for determining the transition time of the digital result signal 1 "0". These transition times are assumed to include information from signal processing 10 of the system. The evaluation unit 112 evaluates the transition time signal 114' based on the aligned signals 丨13. The transition time signals 114 are provided to an output signal generating unit 115. The output signal calculation unit 115 calculates an output signal 116 that is substantially evenly spaced over time and is provided at an output of the signal processing device 100. % The repeated signal] 02 is a repeated analog signal, and more particularly a repeated modulated RF signal. The signal processing device 1 is adapted for analog to digital signal conversion such that a digital signal 116 indicative of the analog waveform of the signal 丨〇2 is provided. 20 As explained in detail below with reference to FIG. 2, the decision unit 1-3 determines a different time point for each of the plurality of repetitions of the signal 102, and the comparison is performed by the comparator unit 〇5 at the time points. . Therefore, the signals generated by the cells 1 〇 3 are substantially evenly spaced apart in time. The reference signal 106 is a fixed signal (c〇nstam) in the embodiment of Fig. 1. At 22 1337458, the repeated "greater than or equal to the reference signal (4) (four) (four) (at a particular point in time), the digital result signal 110 In another case where the repeated signal is less than the reference signal 106, the logical value of the digital signature number 110. 5 The output signal calculation unit 115 can perform - interpolation so that substantially The output signal u 6 is evenly spaced in time. For this purpose, a plurality of signal conversion algorithms can be applied by the output signal generating unit 115. The clock generating unit 108 generates a first for the repeated signal source 1〇1. The clock signal 109a' generates a second clock signal 10 109b to the reference signal generator 1 。 7. As another option for providing two individual clock signals 109a, 10%, it is also possible that the clock generating unit 1 〇 8 The repetitive signal source 101 and the reference signal generator 107 generate a common clock signal. Thus, the embodiment of Fig. 1 enables a measurement of the transition time of the analog waveform reconstruction. The figures 2 〇〇 and 25 图 shown in the figure, which will be further clarified by 15. The time is depicted along a transverse coordinate 201 of the drawing 200. Alonging the ordinate of the drawing 200 2 〇 2, one The signal value is depicted. A first graphic displays the inverse analog signal 102. A second graphic displays the fixed reference signal 106. The time point 2〇3 of the comparison signal 丨2〇 similar to the step function is shown in FIG. Further shown. Along a coordinate 210 of the graph 250, a modified time axis 210 (time modulus) / m = i / 3) is displayed. An ordinate 211 ' along the pattern 250 The signal values are depicted. The rearranged data points 2〇3 are also displayed in the graph 250. In addition, the comparison signal 120 is displayed. In addition, the overtravel time 251 can be obtained from the graph of Fig. 2. In the formula 250, r=〇.7 is the value of the reference signal 1 〇6, n=32 is the number of data points 203. M=3 is the number of times the repeated signal 102 is repeated. Therefore, In Fig. 2, the signal 1〇2 to be sampled has a sinusoidal shape, but may have any alternative shape, especially week. The shape of the period. Advantageously, N and Μ are either prime or mutually prime.

In accordance with the described embodiment, the transition 251 of the repeated digital comparison signal can be accurately evaluated or estimated by coherent digital sampling using a digital channel. The time interval between two measured transients 2〇3 can be expressed as Ts, and the duration of one of the repeated signals 102 can be expressed as D. Preferably, the equation NxTs = MxTR is maintained, where N and Μ are mutually prime, and KxN samples are known. Then according to the arrangement scheme iscrt=m〇(l(ixM,N) samples can be rearranged

sequence. The transition time 251 is between the last sample time before the transition and the first sample time after the transition. According to the coherent sampling scheme of the embodiments of Figs. 1 and 2, an arbitrary fine time resolution OVN) can be obtained, especially when N is sufficiently large. A substantially linear time measurement may result (for the TS hypothesis - fixed frequency sampling clock). 20 — The level accuracy of the sample is determined only by (four) transit time _ quantity _determination' is not determined by the number of static comparison levels. In the following, according to the signal processing device 3 shown in Fig. 3, the same feature will be explained in accordance with the present invention - an exemplary embodiment. Figure 3 shows the device to be tested. To be installed, please generate a signal to be tested 24 1337458 (where t is time), and the signal to be tested S(1) is supplied to a first signal input of comparison II 302. A reference signal r(1) is supplied to a second signal input of the comparator 302 and is generated by a reference signal generator 303. At an output of the comparator 302, a comparison signal c(t) is provided and the comparison 5 signal c(t) is provided to a flip-flop 304. The output of the flip flop 304 is transferred to an input of a memory 305. An output of the memory 305 is coupled to an input of a transition time operation unit 306 for determining the desired transition time.

In addition, the data processing device 3A includes a common clock generator unit 10 307 that produces a common clock for different components of the system 300. The clock generator unit 307 is coupled to a divider DN unit 308' for providing a clock signal to the flip-flop 3〇4. Further, the clock generator unit 3〇7 is consuming to the divider Dref 3〇9 to generate a clock signal to be supplied to the reference signal generator 3〇3. In addition, μ the clock generator 307 is consumed by a divider ·3ι〇, which is round

3H) Next _ to the device to be tested, so that a clock signal is generated for the side to be tested. 20 According to the embodiment of Fig. 3, the homology can be ensured, for example via the common clock reference. The signal to be tested is controlled by the time-receiving rule. ^ After the age-receiving test, it has a period of μ and is controlled by the New Zealand device (4). A static comparison level does not require a clock. In addition, the one-digit period Ts can be controlled by the divider single N3G8. ^ ; The actual example of Figure 3 is optional, there may be three with the same output frequency __ handsome, the fiber accuracy can be

I 25 1337458 is supplied with a .-σ _4 independent clock generator to ensure the desired output frequency. It is important to think that the blocks 304 to 306 of the system can be replaced by any desired time to digital converter unit. The conversion from non-equidistant samples to equidistant samples can be performed according to any signal processing scheme, and in particular can be applied to any digitization of signals with time warfare. In other words, generating equidistant samples from non-equidistant samples is not limited to repeated signals. Figure 4 shows the pattern of Fig. 4〇〇, 430, 450 〇10 in the case of the second time in the case of measuring the transition time of a defective signal (ie, a signal interfered by noise and/or jitter). In the case of jitter or jitter, by looking at all the samples near the transition and selecting an appropriate "average, transition point, one of the time stamps is better evaluated, the town is obtained. That is, jitter can be allowed between samples. Interpolation. Also • 5J* can artificially increase jitter if not sufficiently noisy (such as adding a small PRBS signal to the sampling clock, or connecting a noisy diode to a 15 input). The depiction in the figure includes a noise level of 0.02 rms. In the following 'refer to Figure 5, the transition time for measuring a defective signal (ie, a signal with noise or jitter) will be explained in detail. Figure 5 shows again A pattern 500 similar to the pattern 400 of FIG. 4 and a pattern 550 similar to the pattern 450 of FIG. 4. 2 A simple way to determine a transition time without averaging is to take the first "1" In the scenario of Figure 5, measurement point 203 #12 It is selected as the transition time, so the household is 12/256. The total number of samples is 256. Alternatively, the last "〇, can be selected, which results in the evaluation of the transition time at measurement point 203 #22 0 26 1337458 Achieve some average. For example, an intermediate value between the first "1" and the last "〇" can be selected. This will result in a transition time t = (22 + 12) / 2 / 256 = 17 / 256. In addition, even more averages may be included, which results in a still very computational solution and appropriate results. According to this embodiment, the value "0" can be counted. This can result in a transition time t = 16/2 5 6 . A better transition time assessment (but with more calculations) involves finding the minimum deviation deviation (EiSEJ time "T". S samples around the transition, i = l, ..., S, each with sample time Ti, data value mountain (“0” or “1”). 10 The following equation describes the case of a rising transition: i=\Mi=\Ji<T ί—=1, ί/,=0,Γ2,, Ε2= t (r_〇2+ ί>,-Γ)2 /=l,c/( =\,tt<T /=1.^=0,Γ£/( In the following, refer to Figure 6, re-construction of analog waveform It will be explained in detail. 15 A first graph 600 shows the time dependence of the signals s(t), r(1) and c(t) for a first resolution value of 0.005 according to Fig. 3. A second pattern 650 The time dependence of these signals s(t), r(1), and c(t) is shown for a second resolution value of 0.0001. The exact transition time Tk is determined as described above. The sampled level 20 is known. The reference waveform is given at time Tk, Vk = r(Tk), or exactly the known static level. The sample can be a (time, value) pair (Tk, Vk). The circles in Figures 600, 650 are shown for 0.005. One resolution (Figure 27 1337458 600) and 0.00 A sample (measured) of a resolution (FIG. 650) of 01 (Tk, Vi〇e hereinafter referred to FIG. 7) will be explained in accordance with a signal processing apparatus 700 according to an exemplary embodiment of the present invention. The left-hand side of the figure is substantially equal to Figure 3, so only five additional elements are explained below. The signal processing device 700 comprises a uniform interpolation unit 701, a fast-speed Fourier transform unit 702, and a non-uniform fast Fourier transform unit. 703, an inverse fast Fourier transform unit 〇4, a non-uniform discrete Fourier transform unit 705, a down conversion unit (frequency shift) 7〇6, and an inverse fast Fourier transform unit 707. The transition time unit An output of 306 is provided to a first output 708 as a non-uniform time sample signal. Furthermore, the output of the transition time operation unit 306 is provided to an input of the uniform interpolation unit 〇1. The uniform interpolation unit An output of 701 provides a uniform time sample signal 15 709. Further, the output of the uniform interpolation unit 701 is coupled to an input of the fft unit 702, wherein an input of the FFT unit 702 A uniform spectral signal 710 is provided. In addition, the output of the transition time arithmetic unit 〇6 is provided to an input of the NUFFT unit 703. At an output 20 of the nuFFT unit 703, a uniform spectral signal 7U is In addition, the output of the NUFFT unit 703 is coupled to an input of the inverse FFT unit 〇 4. At an output of the inverse Fft unit tl 704, a uniform time sample 712 is provided. Further, the round-out signal of the transition time arithmetic unit 306 is supplied to the non-uniform hook. At an output of the non-uniform 1) 1: butyl unit 7 〇 5, a signal η] is provided by 28 'the apostrophe 713 indicating a spectrum L at the selected user-defined frequency. In addition, the non-uniform sentence _ The output of the unit period is coupled to the input of the down conversion unit 706. A base frequency number 714 is provided at an output written by the down conversion unit. Furthermore, the output of the down conversion unit 706 is consumed by an input of the inverse FFT unit 7〇7. At the output of the counter scary unit 707, a uniform fundamental frequency sample time signal 715 is provided. In a practical application, usually only a small number of outputs are predicted, such as one or two. The processing according to the embodiment of Figure 7 allows the transition time to be converted to a 10-like ADC sample format. For some applications, non-uniform samples (T, Vi) are fine, such as for eye diagrams. However, many applications require data samples that are evenly spaced or that are spectrally represented at the frequency of the uniform interval. To achieve this solution, it is possible to use uniform interpolation to convert the non-uniformly spaced time samples into samples with uniform time intervals (as is common in the well-known ADC field). For this purpose, a standard FFT can be applied. It is also possible to use non-uniform fast Fourier transform (NUFFT) to calculate a spectrum at a frequency that is evenly spaced from non-uniformly spaced time samples. Preferably, the NUFFT can be combined with the Fast Multipole Method (FMM) for fast time calculation. Using standard inverse FFT, time samples can be calculated. It is also possible to use non-uniform DFT (non-uniform discrete Fourier transform, not so "fast,") to convert non-uniformly spaced time samples into a spectrum at the frequency defined by the user. This is for narrow frequencies. (e.g., rF) 29 1337458 Signals are useful. Figure 8 shows a first pattern 800 and a second pattern 850, which are similar to the pattern of Figure 6.

Figure 8 illustrates a uniform interpolation. It is possible to use the reconstruction algorithm to get the time data for the interval between the sentences. For this purpose, a Sine(1) interpolation is possible, a polynomial interpolation (Lagrange interpolation, spline interpolation, linear interpolation) is possible, and/or a fractional delay filter can be implemented. The pattern 850 illustrates a non-uniform sample from the transition time of a comparison signal by a circle. The dotted line shows the true DUT waveform. The star display in Figure 850 10 uses a uniform interpolation of spline interpolation (e.g., MATLAB interpl (..., 'spline')). In the following, the NUFFT feature will be explained in detail. For this purpose, the term will be defined as follows: JC,. £/?, / = 1..1, time sample Λ =0-1)./. Already two [..#, frequency point A =X(/;)eC, spectrum x.=x((i) = fjXkeJ2:l(k-lw· k = \ W = ) = (θ2πα'-1 ) / ιΛ ), non-uniform Fourier matrix wik = (wi)k] >= eJ2nf〇1' 15 /

w 丨 w,2 w2 w22 N-\

N-\ , matrix N-\

XX wu X,

WNN N J

x = WX

X = w_1 .x, non-uniform FFT 30 1337458 It is possible to use NUFFT to calculate the spectrum at a frequency that is evenly spaced from time samples that are not evenly spaced. A uniform frequency spacing may cause the Fourier matrix to be a Vandermonde matrix (which is always reversible for different Xis, where the row is the power of a row of vectors)' which allows for accelerated conversion compared to any frequency. The NUDFT for narrowband signals (RF) will be explained in detail below. For this purpose, the term will be defined below: X; €仏_ = 1.._#, time sample force foot moxibustion = 1 ... household, attention frequency A = ) e C, attention to frequency spectrum

P

Xi = x(/,.) = Y^X.e^'· k=\ W =(心)=(e W'), non-uniform Fourier matrix

~丨1 · • ··· ww^ • l 乂, Λ · • . . . WNP, XP J

x = W.X

x = W+ ·χ, non-uniform DFT W+ sCWWfW·, definition of virtual inverse matrix w =(w*,), the definition of transpose may be only for p frequencies in a (narrow) frequency range of interest [fL, fH] Fk 10 calculates the spectrum to avoid calculations for millions of frequencies from DC to RF. Alternatively, it is possible to include DC (f = 〇) in the NUDFT to tolerate the offset error. It may be chosen to include known parasitic frequencies to avoid aliasing into these frequencies of interest. 15 It is also possible to directly perform down-conversion to a fundamental frequency in the frequency domain by a simple frequency shift. No hardware down conversion is required in this scenario. Tian N &gt; B day guard, usually N»p, can improve the evaluation of X on average. w variable 31 1337458 into a rectangle. Using the Moore-Penrose virtual inverse matrix w+, the spectrum χ can be calculated. Next, another scenario will be described: '€心=1...% time sample Λ· =(&lt;+moxime-1)./. 6/^ = 1_..^/-<, focus on the frequency At =/(/*)eC, focus on the spectrum on the frequency kd xi ~ χ(0 = J^XkejU{h^k-])f^ w = ( w,,) = (&lt;?/2 is called "-" from), non-uniform Fourier matrix w =

W, Bu w Bu +, w, tt+: W2 VV*, +1 Vl; *i+; • · . • · · • ♦Al ^·+, ^+: , Λ - - - 2 12 2w w :: : w' w2 w·

T tH:w V Si Λ, N &gt;

χ = W.X

x = VT1 · χ, non-uniform FFT The Fourier matrix is a Vandermonde matrix and a pair of angles when the frequency is chosen to be equally spaced within a continuous range of frequencies of interest (typically for RF signals) The product of the line matrix, which simplifies the inversion. Figure 9 shows an exemplary real measurement device 32 900 in accordance with the present invention. 10 1337458 The measuring device 900 includes a control computer (e.g., a workstation, a pc or a laptop) 901 that controls the entire test of the measuring device. A test control unit 902 generates signals and/or receives signals from 〇1; 903, 904, including a repetitive signal 102 that is correlated with measurements performed by the measuring device 9A. A signal processing device 1 having the features described in Fig. 1 is implemented within or separated from the test control unit 902. The test control unit 902 includes a plurality of pins 9〇5 that are connected to the standby devices 9G3, 9G4 (e.g., memory, wafers for mobile phones, etc.). The stimulus signal is provided to the DUTs 903, 904 during the test so that a particular test pattern is executed and the response signal is evaluated by the test control unit 902 and the control computer 901. It is to be noted that the word "comprising" does not exclude other elements or features, and "a" does not exclude the plural. In addition, the elements 15 described in association with the different embodiments can be combined. [Simplified illustration] A signal processing apparatus according to an exemplary embodiment of the present invention is shown. Fig. 2 is a diagram showing a case in the case of a coherent time stamp. Fig. 3 is a diagram illustrating a signal processing apparatus according to an exemplary embodiment. The figure shows a diagram illustrating the function of the signal processing device when there is jitter. Figure 5 shows a diagram illustrating the operation mode of a data processing device when there is jitter. Figure 6 shows an exemplary Embodiments, a diagram of a signal processing device for an operation of analog waveform reconstruction. Figure 7 illustrates a data processing device in accordance with an exemplary embodiment of the present invention. Figure 8 illustrates an exemplary Embodiment 1 A uniform interpolation in the case of a signal processing device. Fig. 9 shows a measuring device according to an exemplary embodiment. 100] signal processing device 115...output signal generating unit 101···reverse signal source 116...output signal 102...periodical repeat signal 120...comparator signal 103···determination unit 130...reactor 104...control Signal 200: Fig. 105... Comparator unit 201: abscissa 106... reference signal 202... ordinate 107···reference signal generator unit 203... data point 108···clock generation unit 210...horizontal coordinates 109a, 109b... Clock signal 211... ordinate 110... digital result signal 250... pattern 111... rearrangement unit 251... transition time 112... evaluation unit 300... signal processing device 113···arranged signal 301...device to be tested 114.·· Transition time signal 302... comparator 34 1337458

303···Reference signal generator 304...Factor 305...Memory 306...Transition time operation unit 307··· Clock generator unit 308.. Divider DN unit 309.. . divider Dref 310...divider DM 400...FIG. 430...Pattern 450...Pattern 500...Pattern 550...Pattern 600...Pattern 650...Pattern 700...Signal Processing Device 7〇. Uniform Interpolation Unit 702...Fast Fourier Transform Unit 703 ...non-uniform fast Fourier transform unit 704...inverse fast Fourier transform unit 705.. non-uniform discrete Fourier transform unit 706··'down-conversion unit 707··, inverse fast Fourier transform unit 708...output 709···even time Sample signal 710.. Uniform spectrum signal 7U... Uniform spectrum signal 712···Uniform time sample 713...Signal 714...Base frequency spectrum signal 715···Uniform base frequency sample Time letter 800...Figure 850... 900...Measurement device 901·..Control computer 902...Test control unit 903.,·Device under test 904...Device under test 905...Pin 35

Claims (1)

^^7458 Extravagant patent application scope replacement - this p 09 30·|--, one.b revision ten, patent application scope ·· - L - a signal processing device for processing signals, the signal processing device includes : - a comparator unit 'for comparing the signal with a _ reference signal, a generating unit 'secret generating a plurality of digital result signals indicating a comparison result, An evaluation unit, which is a transition time of the result (4), is an output signal calculation unit adapted to calculate a plurality of output signals substantially spaced apart. 2. If you apply for a patent range! The signal processing device of the item is adapted to process a repetitive signal. 3. The signal processing apparatus of claim 3, comprising: a determining unit for determining a plurality of low sampling signals point. 4. The signal processing device of claim 3, wherein the comparison unit is adapted to compare the signal with the reference signal at the plurality of time points. 5. The signal processing apparatus of claim 3, wherein the output signal calculation unit is adapted to calculate substantially evenly spaced output signals based on the non-uniformly spaced digital result signals. 6. A signal processing device as claimed in claim i, which is adapted to process an analog signal. 36. The signal processing device of claim 1, wherein the signal processing device is adapted to process the modulated RF signal. 8. The signal processing device of claim 1, which is adapted to be used as an analog to digital converter to convert an analog signal into a digital signal. 9. The signal processing apparatus of claim 3, wherein the decision is adapted to determine a plurality of time points that are different for at least a portion of the plurality of repetitions. 10. The signal processing device of claim 3, wherein the determining unit is adapted to determine a plurality of different time points for at least a portion of the plurality of repetitions to optimize the precision of each additional repeated signal processing . U. The signal processing device of claim 1, wherein the reference signal is one of a group consisting of: a single signal fixed in time, and a plurality of signals fixed in time. a signal that changes in time according to a predetermined waveform, and a sinusoidal signal. 12. The signal processing apparatus of claim 1, wherein the evaluation unit is adapted to take into account signals at time points that are inversely related to the plurality of repetitions to determine the transition times. The signal processing device of claim 3, wherein the number of the time points is greater than the number of times of repetition. 14. The signal processing device of claim 3, comprising: a rearrangement unit adapted to rearrange and according to a one-year table corresponding to a plurality of time points within one cycle of the repeated signal The different 37 1337458 are repeatedly related to the digital result signals. 15. The signal processing apparatus of claim </ RTI> wherein the generating unit is adapted to generate a digital result signal having a first logical value when the respective one of the signals is greater than the reference signal, and in each When the signal is smaller than the reference signal, a digital result signal having a second logic value is generated. 16. The signal processing apparatus of claim 5, wherein the evaluation unit φ element is adapted to determine the transition time based on an analysis of at least one of the group consisting of: digit results The signal has a last time point in a sequence of the first logic value and a digit result k number of the first time point in a sequence of the second logic value. π. The signal processing device of claim 15, wherein the evaluation form 70 is adapted to determine the transition time based on a statistical analysis of a plurality of time points forming an interval, the interval including the first logical value The % bit result signal and includes a digital result signal having the second logic value. 18. The signal processing device of claim 15, wherein the evaluation unit is adapted to: determine based on fitting a mathematical probability function, in particular a cumulative probability density function, to a plurality of time points forming an interval The excess time, the interval comprising a digital result signal having the first logical value and comprising a digital result signal having the second logical value, and determining the transition times from the result of the fitting. 19' The signal processing device of claim </ RTI> wherein the evaluation form 38 1337458 is adapted to determine the transition time based on the presence of at least one of the group consisting of signal jitter and signal noise. . 20. The signal processing device of claim 1, wherein the output signal calculation unit is adapted to determine the substantially evenly spaced output signals by using at least one of the group consisting of: Sine interpolation, polynomial interpolation, Lagrange interpolation, spline interpolation, linear interpolation, and fractional delay 遽-wave. 21. The signal processing device of claim 1, wherein the output signal calculation unit is adapted to calculate the output signals that are substantially evenly spaced in time or substantially evenly spaced in frequency. 22. The signal processing apparatus of claim 1, wherein the output signal calculation unit is adapted to calculate the output signal by performing at least one of the group consisting of: Fast Fourier Transform (Fast Fourier) Transformation ), Non-Uniform Fast Fourier Transformation, Non-uniform Fast Fourier Transform with Fast Multipole Method, and Non-Unif〇rm Discrete Fourier Transformation . 23. The signal processing device of claim 2, wherein the signal processing device is adapted to process the repetitive signal using coherent sampling. 24. The signal processing device of claim 1, comprising: a clock generating unit adapted to generate a common clock signal for a plurality of components of the signal processing device. 39. The method of claim 1, wherein the signal processing device comprises: a plurality of clock generating units each adapted to generate an individual clock signal for a specified component of the signal processing device, wherein the plurality of clock signals are generated The clock generating units are frequency locked to each other. 26. The signal processing device of claim 1, comprising: a jitter increasing unit adapted to selectively increase jitter to the signal. 27. The signal processing device of claim 1, comprising: an automatic test instrument unit as a source for providing the signal. 28. The signal processing device of claim 1, wherein the signal processing device is adapted to be an automatic test instrument for testing a device under test. 29. A measuring device comprising: a signal processing device as claimed in claim 1 for processing a signal associated with a measurement performed by the measuring device. 30. The measuring device of claim 29, wherein the measuring device is adapted to generate a stimulation signal as the signal, the stimulation signal being applied to a device under test to test the device to be tested. 31. The measuring device 1 of claim 29, wherein the measuring device 1 is adapted to receive a response signal from the device under test in response to applying a stimulus signal to the device under test to test the device under test. 32. The measuring device of claim 29, comprising at least one of: an analog to digital converter, a sensor device, a test device for testing a device or substance to be tested, for chemistry, biology A device for learning and/or pharmaceutical analysis, a fluid separation system suitable for separating a mixture of fluids, a capillary electrophoresis device, a liquid chromatography device, a gas chromatography device, an electronic measurement device, and a mass spectrometer. 33. A signal processing method for processing a signal, the method comprising the steps of: comparing the signal with a reference signal, generating a plurality of digital result signals indicating a comparison result, determining a transition time of the digital result signals, and calculating substantially Evenly spaced output signals. From a computer readable medium storing a computer program for signal processing, the computer program is adapted to (4) or perform a method of processing a signal when executed by a processing H, the method comprising the steps of: The signals are compared to generate a plurality of digital result signals indicating non-comparison results, and the transition time of the digital result signals is determined, and the alpha ten is calculated as an evenly spaced output signal. 35. A signal processing U component, the program component is executed by the processor The method is suitable for controlling or executing a method for processing a signal, the method comprising the steps of: comparing the signal with a reference signal, generating a plurality of digital result signals indicating a comparison result, determining a transition time of the digital result signals / Calculate the output _ that is substantially equally spaced.