JPH0575271B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION "Field of Industrial Application" The present invention relates to a time interval analyzer capable of analyzing jitter in a pulse signal, frequency analysis of jitter, frequency of occurrence, etc.

"Background of the Invention" For example, in a magnetic disk device used as an external storage device for a personal computer, a digital audio device called a compact disk, or even a video device such as a laser disk, the signal played back is a coded pulse train signal. It is. If this pulse train signal contains jitter, errors are likely to occur during decoding. Particularly in the case of a computer's external storage device, errors cannot be tolerated.

Therefore, in order to determine the quality of these digital devices, it is necessary to measure whether the jitter of the reproduced pulse train signal is within a specified range.

``Prior art'' A method of measuring jitter in a pulse train signal is to project the pulse to be measured on a conventional oscilloscope as shown in Figure 11, and measure the fluctuation width W of the bright line that depicts this pulse (the horizontal width of the bright line). The amount of jitter is measured.

Another method is to input the pulse to be measured into a spectrum analyzer, and use the spectrum analyzer to analyze the frequency components of the pulse to be measured.
It is also possible to measure the jitter component.

``Problem to be solved by the invention'' In the method using an oscillograph, the measurer determines the amount of jitter by measuring the width W of the bright line projected on the tube surface of the oscillograph. It is difficult to quantify the amount. Particularly in digital audio equipment or digital video equipment, there are many types of pulse widths for reproduced pulses, so pulse waveforms with different pulse widths are displayed on the screen of an oscilloscope in an overlapping manner. As a result, the measurement of the bright line width W becomes increasingly difficult, and only uncertain measurement results are obtained.

On the other hand, when using a method using a spectrum analyzer, if an ideal spectrum analyzer with a good signal-to-noise ratio is used, the frequency components of jitter can be seen quantitatively.

However, since the signal level of the jitter component is minute compared to the frequency component of the pulse being measured, in order to display the jitter component as a spectrum,
A spectrum analyzer with a significantly superior signal-to-noise ratio must be prepared.

In other words, if a normal spectrum analyzer were used, the spectrum representing the jitter component would be hidden in the noise component and would be impossible to observe.

In addition, when there are many types of pulse widths of the pulse to be measured, there are fundamental wave spectra for each type of pulse width, and countless harmonic spectra are generated for each type of pulse width, resulting in a complicated image with just the spectra of those signals. , the spectrum representing jitter becomes increasingly difficult to distinguish.

An object of the present invention is to provide a time interval analyzer that can accurately measure the amount of jitter contained in a pulse.

"Means for Solving the Problems" This invention provides: A. A digital spectrum analyzer equipped with an AD converter in its input circuit; B.
A time-voltage converter that is provided on the front side of the AD converter and converts the pulse width of the pulse to be measured into a voltage signal proportional to the pulse width; A time interval analyzer is composed of a controller that detects a trailing edge and causes an AD converter to perform an AD conversion operation, and the following.

According to the configuration of the present invention, variations in the pulse width of the pulse to be measured and variations in the time between pulses can be extracted as a voltage signal by the time-voltage converter.

This voltage signal is synchronized with the trailing edge of the pulse to be measured.
By performing AD conversion, it is possible to obtain a digital signal that changes in response to fluctuations in the pulse width of the pulse to be measured.

In other words, this digital signal is a digital signal representing the amount of jitter that changes in accordance with fluctuations in the pulse width of the pulse to be measured. Therefore, by frequency-analyzing this digital signal using a spectrum analyzer, it is possible to know the frequency and amount of jitter.

Furthermore, by loading the AD-converted signal into memory and statistically processing the data loaded into the memory, it is possible to measure the history of pulse width fluctuations, frequency of occurrence, etc.

"Embodiment" FIG. 1 shows an embodiment of the present invention. 11 in the diagram
shows a conventionally known digital spectrum analyzer.

In this invention, the digital spectrum analyzer 11 includes a time-voltage converter 12, and an AD
A clock pulse synchronized with the trailing edge of the pulse to be measured is given to the converter 13, and the AD
It is characterized by a configuration in which a controller 14 is attached to perform the conversion operation.

First, a commonly used digital spectrum analyzer will be explained. The digital spectrum analyzer 11 includes an amplifier 16, an analog filter 17 provided at the subsequent stage, an AD converter 13, a magnification change switch 18, and a magnification conversion circuit 1.
9, buffer memory 21, microcomputer 22, fast Fourier transform circuit 23, main memory 2
4 and a display 25.

In the example shown in the figure, a reciprocal converter 26 and a buffer memory 2 are connected between the AD converter 13 and the magnification change switch 18.
7 is provided. The reciprocal converter 26 converts time -
This is a converter that uses a voltage converter 12 and an AD converter 13 to perform reciprocal conversion of a signal that changes according to the jitter period to convert it into a frequency domain signal. The switch 28 is a switch for selecting whether or not to perform reciprocal conversion. When selecting contact A, data that changes with the period of jitter of the pulse to be measured is stored in the buffer memory 27.
give to When contact B is selected, a signal that changes with the jitter period is reciprocally converted into a frequency domain signal, and the signal is input to the buffer memory 27.

In this way, in this example, the buffer memory 27
The structure is such that it can capture either a signal that changes according to the jitter period or a signal that has been converted into a frequency domain signal.

The data taken into the buffer memory 27 is read out and transferred to the second buffer memory 21 after selecting whether or not to change the magnification using the magnification change switch 18.

The magnification change switch 18 is switched to a state in which the magnification is not changed by selecting contact A, and is switched to a state in which the magnification is changed by selecting contact B.

The magnification change circuit 19 is a digital frequency converter 1
9A, 19B and digital filters 19C, 1
9D and resampling switches 19E, 19F
It is composed of

Digital frequency converters 19A and 19B frequency-mix cos2πf ₀ and sin2πf ₀ to the jitter signal, respectively, and shift the center frequency of the jitter signal to an arbitrary frequency. Digital filter 19C, 19D
is provided to remove unnecessary wave components parasitic from a signal whose center frequency is set to a desired frequency.
The signal from which unnecessary wave components have been removed is resampled by resampling switches 19E and 19F, and is taken into buffer memory 21. The change in magnification is performed on signals in the frequency domain. The change in magnification is mainly determined by the resampling period (thinning rate) of the resampling switches 19E and 19F.

In other words, the capacity of the buffer memory 21 is, for example, 1024
If the data is taken into a memory that can store 1024 data and the data is taken in every period t over time T, then the same signal is
If it takes 2T to capture every 2t, the magnification will be 2.

By changing this magnification, the array of frequency spectra displayed on the display 25 can be enlarged and displayed on the frequency axis by an arbitrary magnification, and the resolution can be increased by that magnification. The improved resolution makes it possible to display spectra that were previously invisible.

As an example, there are models in use that can magnify images from twice that magnification to 256 times.

As explained earlier, the factor that determines the magnification is the resampling period (thinning rate) of the resampling switches 19E and 19F, and the center frequency for expansion is the digital frequency converter 19A and 19F.
This is the intermediate frequency set in B.

The data taken into the buffer memory 21 is sent to the fast Fourier transform circuit 23, subjected to Fourier transform, and stored in the main memory 24, and the frequency spectrum is displayed on the display 25.

From the above, the configuration and operation of the digital spectrum analyzer 11 will be understood.

In the present invention, a time-voltage converter 12 is provided on the input side of such a digital spectrum analyzer. An example of this time-voltage converter 12 will be explained using FIG. 2. The example shown in FIG. 2 illustrates a time-voltage converter with a structure as simple as possible in order to simplify the explanation.

This time-voltage converter 12 has a gate 12A that punches out a clock pulse P _d (FIG. 3D) shown in FIG. 3D by a specified pulse Pa shown in FIG. Clock pulse P _e
(Fig. 3E) is given, the clock pulse P _e 's H
Switch element 12 controlled to be turned on only during a logic period
B, a constant current circuit 12C that provides a constant current to the switch element 12B, a capacitor 12D that is charged in fixed amounts each time the switch element 12B is turned on, and a capacitor 12D connected to both ends of the capacitor 12D to charge the capacitor 12D. Charge reset signal
It can be constructed by a reset switch element 12E which is released by supplying P _c (FIG. 3C).

This circuit is a simple step wave generation circuit. By this step wave generation circuit, the pulse width T _p of the pulse P _a is
can be converted into a voltage signal V _s (FIG. 3F).

The controller 14 detects the trailing edge of the pulse P _a to be measured,
A pulse P _b (FIG. 3B) synchronized with this trailing edge is generated. This pulse P _b is applied to the AD converter 13 and the buffer memory 27 to cause the AD converter 13 to perform an AD conversion operation and to cause the buffer memory 27 to perform a data writing operation.

When the spectrum analyzer 11 is operated as an oscilloscope, the changeover switches 18 and 28 may be switched to contact A, respectively.

In this way, the AD conversion output of the AD converter 13 can be written in the buffer memory 27 as it is, that is, in the form of a periodic signal, and can also be transferred to the buffer memory 21 and displayed on the display 25.

As an example, if it is assumed that the input pulse P _a has one type of pulse width and the one type of pulse width is fluctuating due to a problem in the reproduction system, the screen of the display 25 For example, if the voltage waveform corresponding to the fluctuation of the pulse width of the input pulse P _a is
Drawn as shown in the figure.

This voltage waveform shows how the pulse width T _p changes over time. By measuring the period T _n of this displayed waveform, it is possible to know the fluctuation period of the pulse width, that is, the period and frequency of the jitter.

Furthermore, by measuring the amplitude W _n of this waveform, the maximum and minimum values of jitter can be determined.

If we invert the polarity of the input pulse P _a and input it as a negative polarity pulse, then the time corresponding to the time interval between the pulses P _a is converted to voltage, and this is AD converted, so this AD By displaying the converted signal on the display 25, variations in the time interval between pulses can be displayed.

By the way, for example, like a compact disk, T
= 231.385 ns as a unit and has 9 types of pulse widths from 3T to 11T, and when a pulse signal in which signals of these 9 types of pulse widths are irregularly generated is input, the screen of the oscilloscope will appear as shown in Figure 5. Nine bright lines LN ₁ to LN ₉ are drawn as shown in . In other words, when viewed microscopically, these nine bright lines LN ₁ to LN ₉ are dots each displayed by one pulse, as shown in Figure 6.
It is drawn as a line by the set of D ₁ , D ₂ , D ₃ ..., and since the dots move at high speed, there are almost no bright lines between the lines LN ₁ to _{LN 9} . It is never displayed.

The nine bright lines LN ₁ 〜 displayed in this way
LN ₉ displays the pulse width variation for each pulse width. In other words, it is possible to know whether the pulse width varies depending on whether the width (vertical width) of each bright line LN ₁ to LN ₉ is thick or thin.

By the way, if you draw all nine lines LN ₁ to LN ₉ , you can tell whether the pulse width varies depending on whether each line is thick or thin, but the rules for that variation are If there is a frequency or regularity, it is not possible to measure the period until it is displayed.

For this reason, the spectrum analyzer 11 is operated as an original spectrum analyzer. For this purpose, the changeover switches 18 and 28 can be turned to the contact point B.

By switching the changeover switch 28 to contact B, the periodic signal is reciprocally converted by the reciprocal converter 26 and becomes a frequency domain signal. Pulse the signal at this frequency
_Pb (FIG. 3B) is read into the buffer memory 27.

The signal taken into the buffer memory 27 is read out and transferred to the second buffer memory 21 via the magnification changing circuit 19. The magnification in the frequency axis direction is selected by the magnification changing circuit 19, and the resolution in the frequency axis direction is determined.

The data whose magnification has been changed by the magnification changing circuit 19 is once taken into the second buffer memory 21, and then Fourier transformed by the fast Fourier transformer 23 and subjected to frequency analysis. The frequency analysis results are stored in the main memory 24.
are sequentially taken in and displayed on the display 25.

An example of the display is shown in FIG. The spectra SP ₁ , SP ₂ , SP ₃ , SP ₄ . . . shown in FIG. 7 indicate frequency components of jitter. SP ₁ represents the fundamental wave component of jitter. SP ₂ , SP ₃ , SP ₄ to this fundamental wave
It can be seen that the signals of the frequencies indicated by ... are superimposed. The frequency of the signal represented by each spectrum SP ₁ , SP _{2 .} . . can be read from the scale 31 on the horizontal axis on the display surface. The amount of the fundamental wave and the signal superimposed on it is indicated by the scale 32 on the vertical axis on the display screen.
It can be read by

By performing frequency analysis using the fast Fourier transformer 23 in this manner, it is possible to frequency analyze and measure jitter included in the pulse to be measured P _a . Furthermore, even if there are many types of pulse widths of the pulse P _a to be measured, the frequency components of jitter can be displayed as a spectrum.

Furthermore, in order to obtain the jitter superimposed on a pulse of one pulse width, when reading data from the buffer memory 27, only the data of a certain frequency component is read using the window function of the microcomputer 22, and this data is subjected to fast Fourier transform. Then frequency analysis can be performed. In this way, only the jitter superimposed on a single period pulse can be displayed as a spectrum.

On the other hand, by providing the buffer memory 27, it is possible to reproduce how the jitter changes by reading the data taken into the buffer memory 27. In other words, you can see the history of pulse width fluctuations. Further, by using the statistical processing function of the buffer memory 27 and the microcomputer 22, it is also possible to display, using a certain pulse width as a reference, the frequency or distribution of pulses with pulse widths that occur before and after the reference pulse width.

FIG. 8 shows another embodiment of the invention. In this example, a time measuring means 33 is provided to measure the pulse width of the pulse to be measured, and data is stored in the buffer memory 27 only when a pulse falling between the upper and lower time limits set in the time measuring means 33 is input. This shows a case where the configuration is configured to allow import.

By providing the time measuring means 33 in this way, even when pulses to be measured with multiple pulse widths are given to the input, a pulse with one pulse width is extracted from the pulses with multiple types of pulse widths and stored in the buffer memory. 27 can be written to. As a result, when measuring jitter superimposed on a pulse with a single pulse width, the buffer memory 27 stores only data showing fluctuations in that pulse width, and does not store data on other pulse widths.
can make maximum use of its capacity.

"Operations and Effects of the Invention" As explained above, according to the present invention, the pulse to be measured is time-voltage converted to convert the pulse width fluctuation or pulse interval fluctuation of the pulse to be measured into voltage fluctuation, and this voltage fluctuation signal is Since the configuration is such that the signal is AD-converted and input to the spectrum analyzer, a signal that does not contain the frequency component of the pulse under test, that is, only has a jitter component, can be input to the spectrum analyzer. As a result, when operated as a spectrum analyzer, only the spectrum showing the frequency components of jitter can be displayed on the display, and other signals with high levels will not be displayed. Therefore, the displayed spectrum is displayed sufficiently large compared to the noise component, and the level, frequency, etc. can be reliably measured without being overwhelmed by the noise component.

However, if you operate the spectrum analyzer as an oscilloscope, you can see the jitter waveform. Therefore, the maximum and minimum values of jitter can be measured from the amplitude of this waveform.

As described above, according to the present invention, various measurements related to pulse train signals can be carried out, and the effect is extremely large when it is used for evaluating digital regenerators and determining whether or not they are good or bad.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a connection diagram for explaining an example of a time-to-voltage converter used in this invention, and FIG. 3 is a time-to-voltage converter.
4 and 5 are front views showing examples of analysis results of the time interval analyzer according to the present invention, and FIG. 6 is a waveform diagram for explaining the timing of the voltage converter and AD conversion operation. FIG. 7 is a front view for explaining the main analysis results of this invention, and FIG. 8 is for explaining another embodiment of this invention. FIG. 9 is a diagram for explaining a conventional jitter measurement method. 11: Spectrum analyzer, 12: Time-
Voltage converter, 13: AD converter, 14: Controller,
27: Battle memory.

Claims (1)

[Scope of Claims] 1. A. A digital spectrum analyzer equipped with an AD converter in its input circuit; B. A digital spectrum analyzer provided in the preceding stage of the AD converter, which converts the pulse width of the pulse to be measured into a voltage proportional to the pulse width. A time interval analyzer comprising: a time-voltage converter; and a controller that detects a trailing edge of a pulse to be measured that is input to the time-voltage converter and causes the AD converter to perform an AD conversion operation.