JPH0388505A - Optional waveform generator - Google Patents

Abstract

PURPOSE:To improve the resolution of an output waveform over the resolution of a DA converter by storing errors between arithmetic results and DA input data which are less than minimum resolution and determining DA input data by an arithmetic control circuit so that the stored errors are less than a half as large as the value of a minimum bit at all times. CONSTITUTION:The arithmetic control circuit 10 finds waveform data according to a waveform definition expression. The arithmetic result is converted into data D consisting of a number N of bits for the DA converter 3 (provided that N<M). Then E=M-D is calculated and a next storage error is further found. Here, EX=E(X-1)+E, where EX is a current stored error and E(X-1) is a last one. When Ex is judged and EX<=-1/2LSB, +1 is added to D and EX to obtain new values. Here, LSB is the value of the minimum bit of the DA converter. When +1/2LSB<=EX, -1 is added to D and EX to obtain new values. Said operation is repeated to obtain high-resolution waveform data.

Description

DETAILED DESCRIPTION OF THE INVENTION <Field of Industrial Application> The present invention relates to increasing the resolution of digital-to-analog conversion in an arbitrary waveform generator.

<Prior Art> An example of a conventional digital arbitrary waveform generator is shown in FIG. In the figure, 1 is a waveform memory in which waveform data is stored, and 2 is a tarok address generator, which generates the address necessary to read out the waveform data from waveform memory 1, and also generates the readout waveform data (digital data). )
is a digital-to-analog converter (hereinafter referred to as a DA converter)
3 generates a conversion clock for analog conversion.

4 is an arithmetic/control circuit that calculates the waveform data to be given to the waveform memory 1 from the waveform definition equation, and also calculates the waveform data given to the waveform memory 1 by calculating the clock and
It controls the address generator 2 to appropriately generate addresses and clocks.

5 is a low-pass filter for removing high frequency components of the output of the DA converter 3 and smoothing the waveform.

In such a configuration, the waveform definition equation is calculated in advance in the calculation/control circuit 4 to obtain waveform data of the output waveform,
Store in waveform memory. Thereafter, the clock address generator 2 specifies an address, the waveform data is read from the waveform memory 1, and the DA converter 3 converts it into analog data. A continuous analog waveform can be obtained by repeating this data reading and analog conversion operation without interruption.

<Problems to be Solved by the Invention> By the way, in such a conventional arbitrary waveform generator, when a waveform definition formula is calculated in the calculation/control circuit 4 and converted into waveform data of an output waveform, the valid bits of the calculation value itself are Even though the number is large and the resolution is high, data below the resolution of the DA converter is rounded off and the value is stored in waveform memory 1.
is giving to

In this method, a quantization error occurs below the resolution of the DA converter, and for example, with an 8-bit DA converter, only 8-bit resolution can be obtained.

The specifications required for the DA converter of an arbitrary waveform generator are high speed and high resolution, but conventionally high-speed DA converters have low resolution, and high-resolution DA converters have low speed. There was a problem that both could not be satisfied with one DA converter.

Note that in order to output an arbitrary waveform with an arbitrary waveform generator, it is impossible to realize an ideal RI-A converter (an impulse response of an ideal low-pass filter) and an ideal low-pass filter. Usually an oversampling technique is used. In this method, the sampling frequency is set as high as possible and a somewhat rough low-pass filter is used.

However, using this method has the disadvantage that a DA converter with a high sampling rate has a small resolution and a large error in the voltage direction.

The purpose of the present invention has been made in view of the above points, and although it uses a high-speed, low-resolution DA converter, it is possible to output more than the resolution of the DA converter by using the waveform data creation method and band limitation. The object of the present invention is to provide an arbitrary waveform generator that can provide high resolution.

Means for Solving the Problems> In order to achieve such objects, the present invention calculates a waveform definition formula using an arithmetic/control circuit, stores the calculation result in a waveform memory as DA input data, and stores the calculation result in a waveform memory as DA input data. In an arbitrary waveform generator configured to generate an arbitrary waveform by sequentially reading waveform data from a waveform memory, converting it into analog data using a DA converter, and outputting it through a low-pass filter, the arithmetic and control circuit comprises: , when converting the operation result to DA input data, an error below the minimum resolution between the operation result and the DA input data is accumulated, and if the accumulated error exceeds half the value of the minimum bit of the DA converter, the DA input The present invention is characterized in that it includes a function of performing arithmetic processing by adding or subtracting data by one bit and applying feedback so that the accumulated error is always smaller than half of the value of the smallest bit.

Effect> In the present invention, in the arithmetic/control circuit, the arithmetic result and DA
If an error that is less than the minimum resolution between input data is accumulated and the accumulated error exceeds half of the value of the minimum bit of the DA converter, the DA input data is added or subtracted by 1 bit so that the accumulated error is always the minimum bit. The DA input data is determined so that it is less than half the value of .

This modulates the output of the DA converter, and the output waveform obtained by filtering this becomes an output waveform with a high resolution higher than the resolution of the DA converter.

Example The present invention will be described in detail below with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of an arbitrary waveform generator according to the present invention. In the figure, parts equivalent to those in FIG. 4 are designated by the same reference numerals, and their explanation will be omitted.

10 is an arithmetic/control circuit, which supplies and controls data to the waveform memory 1 and the clock/address generator 2 in the same way as in the conventional example;
When converting into input data to the A converter, the calculation result and DA
Accumulates errors below the minimum resolution between the input data and
When this amount reaches ±1/2LSB (LSB is the value of the minimum bit of the DA converter), the DA input data is added or subtracted by ILSB, and feedback is performed so that the accumulated error is always less than ±1/2LSB. The difference is that it is multiplied by

The operation in such a configuration will be explained next.

FIG. 2 shows the operation flow.

(1) The arithmetic/control circuit 10 obtains waveform data based on the waveform definition formula. Let M be the calculation result at this time.

(2) Convert the calculation result M into data with the number of bits N for the DA converter 3 (N<M). Conventionally, 1/2N
The value of M that is less than a bit is rounded off and stored directly in the waveform memory 1, but in the present invention, E=MD is determined, and then the next accumulated error is determined.

E -E(×-1) +E
In the case of 2LSB, the value obtained by adding 1 to D is used as a new value, and the value obtained by adding 1× to E is used as a new value Ex.

■If -1/2LSB<Ex<+1/2LSB, D
, Ex are both left unchanged.

(2) If +1/2LSB≦Ex, both D and Ex are each decremented by 1 to obtain new values.

(4) Write the value obtained in (3) above to the waveform memory.

By repeating the above operations, waveform data as shown in FIG. 3, for example, can be obtained.

Figure 3 is an example of using a 2-bit DA converter.
As shown in the figure (a), the calculation result E is 2.2°2.2.
2.2.2.2.2.3.2.3. , , , DA input data is shown, and FIG.

The same figure (c) shows the output data (data of the error feedback modulation method) from the calculation/control circuit 10 of the present invention for the same calculation result.For example, the third calculation result 2.2 is obtained. At this point, the accumulated error is the same, 6, so D is +1
and outputs 3 as D, and Ex becomes -1 and -0
, 4.

As a result of such processing, the A input data is output as shown in FIG.

Then, by passing the output of the DA converter 3 through the low-pass filter 5, the modulation frequency is removed, making it possible to generate an output below the minimum resolution of the DA converter.

Note that if the cutoff frequency of the low-pass filter is lowered, the resolution will logically increase by the lowered frequency.

<Effects of the Invention> As described in detail above, according to the present invention, it is possible to easily generate an output that is less than or equal to the minimum resolution of the DA converter.

Note that when the oversampling method was used in the conventional example, the error in the voltage direction became large, but if the oversampling method is applied to the present invention, a waveform with frequency components lower than the sampling frequency is generated. This has the advantage that voltage resolution is high and errors in arithmetic logic values and output waveforms are small.

[Brief explanation of drawings]

Fig. 1 is a block diagram showing an embodiment of an arbitrary waveform generator according to the present invention, Fig. 2 is a diagram showing an operation flow, and Fig. 3 is a calculated value and waveform data when using a 2-bit DA converter. FIG. 4 is a block diagram showing an example of a conventional arbitrary waveform generator. DESCRIPTION OF SYMBOLS 1... Waveform memory, 2... Clock/address generator, 3... DA converter, 10... Arithmetic/control circuit.

Claims (1)

[Claims] A waveform definition formula is computed by an arithmetic/control circuit, the computed results are stored in a waveform memory as DA input data, and the waveform data is sequentially read out from the waveform memory and converted into analog by a DA converter. In an arbitrary waveform generator configured to generate an arbitrary waveform by further outputting the waveform through a low-pass filter, the arithmetic/control circuit converts the arithmetic result into DA input data. If an error that is less than the minimum resolution between input data is accumulated and the accumulated error exceeds half of the value of the minimum bit of the DA converter, the DA input data is added or subtracted by 1 bit so that the accumulated error is always the minimum bit. An arbitrary waveform generator characterized in that it is configured to include a function of performing arithmetic processing to apply feedback so that the value of is smaller than half of the value of .