JPS62136129A - Test method for analog-digital converter - Google Patents

Abstract

PURPOSE:To evaluate simply a conversion error at a high speed by providing a special relation between an input signal frequency fs and a sampling clock frequency fCLK, interleaving an output code of an A/D converter to be tested at an interval of (N-1) codes and comparing two consecutive data strings obtained in this way. CONSTITUTION:The relation of fs=(fCLK/N)+DELTAf (N:1,2,3 and fs>>DELTAf) is held between a sampling clock frequency fCLK and a frequency fs of a continuous period function wave inputted to an A/D converter 3 to be tested, the output codes obtained by the conversion of the A/D converter 3 are interleaved at an interval of (N-1) codes and two consecutive data strings obtained are inputted to a subtractor 7. An error discriminator 8 detects the conversion error by observing the difference of the two conswecutive data from the subtractor 7. The data string being the result of interleaving the output codes at an interval of (N-1) codes becomes a frequency function wave data by one period with a number of fCLK/NDELTAf and beat frequency DELTAf. The value DELTAf is selected so that the difference of two consecutive data at the maximum gradient of the periodic function wave is sufficiently smaller than the allowable calculated conversion error.

Description

[Detailed description of the invention] Industrial applications The present invention relates to a method for testing a MU/D converter.

This method is particularly effective for testing the dynamic characteristics of A/D converters with high conversion speeds, and is also effective for simplifying test equipment and speeding up testing.

Conventional technology What is the conventional dynamic characteristics test method for A/D converters?

As shown in FIG. 2, the sine wave signal generator 9 generates a sine wave.
The sampling clock from the sampling clock generator 2 is input to the analog input terminal 3 of the A/D converter 3 under test.
1. A digital output code inputted to the sampling clock input terminal 32 and outputted from the digital output terminal 33 is stored in the buffer memory /Q. An ideal sine wave closest to this data is calculated from the obtained data string by a computer 11, and dynamic characteristics are evaluated from the error between this ideal sine wave and the actually converted data.

Problems to be Solved by the Invention Although it is possible to evaluate the conversion accuracy of an A/D converter from the error between the ideal sine wave obtained by calculation and the actually converted data using such a test method, it is not possible to use it for calculation. The number of sample data to be analyzed is small, it is difficult to fully evaluate the test subject/D converter against all conditions between the input signal and the sampling clock, and calculations by a computer require a considerable amount of time. In addition, the test equipment is complicated because it requires a high-precision sine wave generator, a high-speed buffer memory, and a computing computer.

In view of the above points, the present invention evaluates the conversion accuracy of an A/D converter by processing a large amount of converted digital output codes in real time. The present invention provides a test method that can speed up testing without requiring a high-speed buffer memory or a computer for calculation.

Means for Solving the Problems In order to solve the above problems, the present invention provides
Continuous periodic function wave (fs:>Δf) input to the converter
), extracting every other (N-1) code from the output code obtained by conversion by the A/D converter, and comparing two consecutive pieces of data in the obtained data string. be.

Effect: The data string obtained at this time consists of one period at the input signal frequency. Here, by selecting Δf so that the difference between two consecutive data at the maximum slope part of the periodic function wave is a value sufficiently smaller than the calculation-allowable conversion error (approximately 0.1 LSB or less), Two consecutive data columns
The purpose is to compare the data and find the difference to evaluate the dynamic characteristics of the human/D converter. In addition, in the present invention, it is possible to evaluate the MU/D converter by simply comparing two consecutive data strings obtained every (N-1) times, so it can be realized with simple hardware logic, and in real time. A large amount of output data can be processed in a short amount of time. Furthermore, the waveform quality of a single-handed signal is not required.

Embodiment FIG. 1 shows an embodiment of the present invention. In FIG. 1, 1 is a periodic function wave generator, and 2 is a sampling clock generator.

3 is an A/D converter under test, which has analog signal input terminals 31. It has a sampling clock input terminal 32, and the converted output code is outputted from a digital output terminal 33. 4 is an X frequency divider which generates a clock for extracting every (N-1) output code, and is supplied to latch 5 and latch B 6. Latch person 5 input niha,
Since the output of the system/D converter 3 and the output of the latch operator 5 are connected to the input of the latch B60, data one clock before the latch input is always set in the latch B6. Therefore, the subtracter 7 is always input with two consecutive data strings obtained by extracting every (N-1) output code from the A/D converter 3, and by observing the difference, the error determiner 8 Conversion errors can be detected.

+Δf (N: 1.2.3...) (fs>Δf)
This becomes frequency function wave data for one period of frequency Δf.

The value of H depends on fs and f depending on the performance of the human/D converter 3.

Select the optimal value from the relationship □, and set the value of Δf so that the difference between two consecutive data of the maximum slope part of the 5 periodic function waves is sufficiently smaller than the allowable conversion error (o, about 1LsB or less). choose. For example, if a sine wave with the full dynamic range is input to a 1 m pit Mu/D converter, the difference between two consecutive data points at the maximum slope will be ±1 LsB unless there is a conversion error. The conversion error of the MU/D converter 3 can be detected by observing the difference.

For the input signal in this embodiment, only the slope of the maximum slope part is important and is not affected by the waveform quality. Therefore, there is no problem even if the input exceeds the input dynamic range of the M/D converter under test.

Effects of the Invention As described above, according to the present invention, it is possible to easily and quickly evaluate the conversion error when an AC signal is input to an IA/D converter, and it is extremely useful in practice.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram of a dynamic characteristic testing device for a human/D converter according to an embodiment of the present invention, and FIG. 2 is a block diagram of a conventional testing device. 1... Periodic function generator, 2... Sampling clock generator, 3... A/D converter under test, 31... Analog signal input terminal , 32...
... Sampling clock input terminal 133 ... Digital signal output terminal. 4...y frequency divider, 6...latchim, 6
...Latch B, 7...Subtractor, 8...
...Error judger, 9...Sine wave signal generator, 10...Buffer memory, 11...
Computer. Name of agent: Patent attorney Toshio Nakao and 1 other person 2nd
figure

Claims (1)

[Claims] Input signal frequency f_S of continuous periodic function input to the A/D converter under test and sampling clock frequency f_C_
f_S=(f_C_L_K/N)+Δf between L_K
(N: 1, 2, 3...) (f_S≫Δf), and every (N-1) digital output codes converted by the A/D converter under test are 1. A test method for an A/D converter, which comprises comparing two consecutive pieces of data obtained by extracting data.