JPH01174119A - Evaluation device for ad converter - Google Patents

Abstract

PURPOSE:To attain dynamic characteristic measurement by counting a sampling number of times of an output digital code of an AD converter by a digital code and obtaining a linear error based thereupon. CONSTITUTION:A triangle wave outputted from a triangle wave oscillator 11 is subject to digital processing by an AD converter ADC 12 being a measurement object and outputted in a form of digital code(Dc) corresponding to the level of an input signal. The Dc from the ADC 12 is sent to a fetch section 13, fetched at a specific sampling frequency and one address is assigned to each Dc by a RAM 13b and the sampling number is counted in this address. After the Dc is fetched for a prescribed period, a CPU 14 sends an address designation signal to the RAM 13b and the count of Dc is read from each address and written on a disk device 15 with a file name given thereto. Moreover, the operation is applied to the count to obtain a linearity error and a differentiation linear error and the result is outputted to an output device 16.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention provides an AD converter that can directly and highly accurately measure the nonlinearity error and differential linearity error of an AD converter with digital output. The present invention relates to a converter evaluation device.

[Prior Art] As is well known, conversion errors of an AD converter include a quantization error, a nonlinearity error, an offset error, a gain error, and a total error that is a combination of these errors.

The nonlinearity error described above indicates how much the relationship between the input analog voltage and the output digital code deviates from an ideal straight line. (See Iwao Sagara, A/D-D/A conversion technology of the microcomputer age, Nikkan Kogyo Shimbun, first edition, p. 16-28).

FIG. 5 shows an example of a conventional AD converter evaluation method. After the sine wave with constant amplitude and frequency input to the AD converter (ADC) I is converted into a digital code,
It is inversely converted by a DA converter (DAC) 2 and becomes an analog quantity. This analog quantity corresponds to the input sine wave, and is input to the distortion meter 3 to measure the distortion factor. This predicts the total ADCI error.

[Problems to be Solved by the Invention] By the way, the conventional evaluation method for AD converters described above has the following drawbacks.

■Since the output of ADCI is inversely converted by DAC2, D
It will be influenced by AC2.

(2) Since the input waveform to ADCI has constant amplitude and frequency, it is not possible to measure dynamic characteristics that vary over time.

■It is only possible to predict the total error.

This invention was made against this background, and
It is an object of the present invention to provide an evaluation device for an AD converter that can directly and dynamically measure the nonlinearity error of a D converter without using a DA converter.

[Means for Solving the Problems] In order to solve the above problems, the present invention provides an oscillator that generates a repetitive waveform whose amplitude changes linearly, such as a triangular wave, and an AD converter for the output of this oscillator for a fixed measurement period. a counting means for sampling the output digital code output from the AD converter when input to the converter and counting the number of samplings for each code; and a linearity error and/or differential linearity error of the AD converter based on the counting results. It is characterized by comprising a calculation means for performing calculations and an output means for outputting the calculation results.

[Function] According to the above configuration, the input terminal to the AD converter changes linearly and is converted into a digital code corresponding to the input voltage. In this case, the width of the input voltage converted into one digital code is proportional to the number of times this digital code is sampled. For example, if the number of samplings of digital code A is 150 and the number of samplings of digital code B is 100, the width of the input voltage converted to digital code A is
This is 1.5 times the width of the input voltage converted to . Therefore, if you input a linearly changing waveform to an AD converter, sample its output dental code, and find the number of samplings for each digital code, you can calculate the width of the input voltage converted to each digital code and the converted voltage. It is possible to find the threshold value.

Consider the case of a triangular wave. The probability density function of the triangular wave is P(V
), the full-scale voltage is A, and considering that the triangular wave voltage rises from O to A and then falls from A to O, P(V) = 1/2A... (1) becomes To establish.

Therefore, the voltage of the sampled triangular wave is Va=V
The probability P (V a, V b) between b is calculated by integrating the probability density function between Va=Vb.
, Vb)-(Vb-Va)/2 A---(2)
becomes. Here, the total number of samples is Nt, and P (V
a, V b) = H/N t--(3),
Regarding the voltage vb, Vb=2A −H/Nt
+Va--(4).

Here, the output digital code i (this i is the number of the digital code, for example, 0, 1.2 from the smallest value)
Let H(i) be the number of samplings (referred to as n), that is, the number of counts, and let Vi be the threshold voltage of the output digital code i, then from equation (4) above, Vi = 2 A-H(i)/ N t+ V i-1
...(5). This threshold voltage Vi is
This means that the digital code i is output when the input voltage to the AD converter reaches Vl.

Here, the cumulative histogram CH(iX-H(0)+H(
1)+H(2)+...10(i)), and adding both sides of equation (5), we get Vi=2A-CH(i)/Nt...(6) , threshold voltage Vi is obtained.

The linearity error and the differential linearity error can be determined using this threshold voltage Vi, the details of which will be described later.

[Example] Hereinafter, the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.

In the figure, reference numeral 11 denotes a triangular wave oscillator, which oscillates a triangular wave of a constant frequency (audio frequency) whose amplitude varies linearly, for example, between θ and 5×. The triangular wave output from the oscillator 11 is supplied to the input terminal of the ADC 12 to be measured, and is converted from an analog signal to a digital signal. This digital signal is output in the form of a digital code (for example, 2° S combination) that corresponds to the level of the human input signal.

The digital code output from ADCI 2 is sent to import section 13 . This acquisition section 13 has a sampling section 13a and a RAM 13b. For example, the sampling section 132L has a frequency of 48kHz (or 96kHz).
The output digital code of the ADC 12 is taken in at a sampling frequency of The number of samplings is counted. In other words, each time sampling is performed, the contents of the address corresponding to the digital code at that time are incremented. Since this operation is all performed by hardware, it is possible to capture data in real time.

Note that the ratio between the frequency of the above-mentioned triangular wave and the sampling frequency is set to be an irrational number.

This means that when measuring over multiple periods of a triangular wave,
This is to ensure that the sampling timing is slightly shifted.

After capturing the digital code for a certain measurement period (an integral multiple of the period of the triangular wave), the CPU 14 stores the data in the RAM 13.
An address designation signal is sent to b, the count number of each digital code is sequentially read from each address, a file name is attached, and this is written to the disk device 15.

In addition, the calculation described later is performed on this count number to obtain a linearity error and a differential linearity error, and the results are sent to the output device 1.
Output to 6. This output device 16 consists of a CI (T display device) or a printer.

It should be noted that the above-mentioned import section 13 corresponds to the counting means and the CPU 14 corresponds to the calculation means in the claims.

Next, the operation of the CPU 14 in this embodiment will be explained with reference to the flowchart in FIG.

The operating principle of this invention has already been explained in the "Operation" section, so it will be omitted here.

Step S1: In order to read the measurement results stored in the disk device 15, the name of the file storing the desired measurement results is read.

Step S2 The contents of this file, that is, the number of sampling times H(i) of each digital code are read from the disk device 15. Here, the value i indicates the number of the digital code and takes a value from 0 to n, where 0 corresponds to input voltage 0 and n corresponds to full-scale voltage A.

Step S3 For each digital coat l, the sum (cumulative histogram) of sampling times H(i) from 0 to i (cumulative histogram) CH(
Take i). That is, c H(i)=H(0)+H(1)+H(2)+
...... Find +H(i).

FIG. 3 shows an example of a histogram of the sampling number H(i). In the figure, the horizontal axis shows the digital code number 1, and the vertical axis shows the sampling number H(i
) is expressed as a percentage. However, this percentage is
Assuming that there is no error in 2, the normal number of samplings is 000, when it is twice the normal number it is 100%, when it is three times the normal number it is 200%, and so on.

Step S4: Calculate the threshold voltage V (i) of the digital code i from the cumulative histogram CH (i).

It has already been explained that this threshold voltage V(i) is given by V(i)=2A-CH(i)/Nt, where the total number of samples is Nt.

Note that the threshold voltage V C+) corresponds to the minimum input voltage converted into the i-th digital code i.

Step S5: Obtain linearity error LN(i) from the threshold voltage V(i). That is, LN(i)-(V(i)V(0))/ILSB-(i+
1) Calculate the linearity error LN(i) using (7). Here, the first term on the right side corresponds to the input voltage difference determined from the measured values expressed in LSB units, and the second term corresponds to the theoretical value expressed in LSB units.

Step S6 Differential linearity error D N (i
). That is, D N(i)=(V (i+1) −V (i))/
The differential linearity error DN(i) is calculated by l L S B -1 (8). Here, the first term on the right side is the differential value of the input voltage obtained from the measured value expressed in LSB units, and the second term is the theoretical value expressed in LSB units.
It corresponds to what is expressed in units.

The error thus calculated is output to the fourth output device 16 by the output device 16.
The output will be in the format shown in the figure. Here, Fig. 4 (a
) is a diagram showing an example of linearity error L N (i), and FIG. 4(b) is a diagram showing an example of differential linearity error D N (i), and the horizontal axis shows digital code number 11. On the vertical axis, the linearity error LN(i) and the differential linearity error D N (i) are shown in units of ILSB.

As can be seen from these figures, both errors are 0.00
Since it is output in a manner that swings up and down around the center, it is possible to visually grasp the error corresponding to each digital code.

[Effects of the Invention] As explained above, in this invention, the number of samplings of the output digital code from the AD converter is counted for each digital code, and the linearity error and differential linearity error are calculated based on this. Therefore, the following effects can be obtained.

(2) Since the output of the AD converter is directly measured without passing through the DA converter, the influence of the DA converter can be removed.

■Since measurement is performed by inputting waveforms that change over time, dynamic characteristic measurements are possible.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.
FIG. 2 is a flowchart for explaining the operation of the embodiment, FIG. 3 is a graph showing an example of a histogram of the number of samplings of each digital code in the embodiment, and FIG.
Figures (a) and (b) are graphs showing examples of linearity errors and differential linearity errors, respectively, and Figure 5 is a block diagram for explaining the conventional evaluation method. +1...Triangle wave oscillator, 12...ADC (
AD converter), 13... Intake unit (counting means), 14... CPU (calculating means), 15...
. . . Disk device, 16 . . . Output device.

Claims (1)

[Claims] An oscillator that generates a repetitive waveform whose amplitude changes linearly, such as a triangular wave, and the output of this oscillator are measured over a fixed measurement period AD.
A counting means for sampling the output digital code output from the AD converter when input to the converter and counting the number of sampling times for each code;
An evaluation device for an AD converter, comprising a calculation means for calculating a linearity error and/or a differential linearity error of a D converter, and an output means for outputting a calculation result.