JPS62179224A - Digital/analog converter - Google Patents

Abstract

PURPOSE:To convert a digital data having a large bit length into an analog data with a low cost by combining a D/A conversion circuit having the K-bit of processing capability and the D/A conversion circuit of an L-bit of processing capability so as to apply D/A conversion to a digital data having a (K+L)-bit length. CONSTITUTION:A digital data represented by the floating point form desired to be subjected to D/A conversion is fed to a D/A converter 1 to obtain an analog data Y comprising a product Ym.Ye between a mantissa part D/A conversion output Ym and an exponential part D/A conversion output Ye by giving the mantissa part D/A conversion output Ym as a reference power supply Vref to an exponential part D/A conversion circuit 3. Thus, the digital data having a (K+L)-bit length is D/A-converted by the combination of two D/A conversion circuits 2 and 3 having the processing capability of K and L bits respectively. Then the digital data having a large bit length is converted into an analog data with a low cost.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a DA converter that converts digital data in floating point format from DA to digital using a plurality of DA converters.

[Prior Art] The format of data processed by computers includes unsigned integers, signed integers, and real numbers used to represent high-precision numbers. Real number type data has a mantissa and a mantissa. It is called a floating point format consisting of an exponent, and is distinguished from a signed integer format called a fixed point format. In floating point format, certain numerical data Y is divided into a mantissa M and an exponent E.
, using the base, it is expressed as Y=M-bF-. For binary numbers, b is 2.

In general, the number of bits constituting the mantissa and exponent parts and the data representation format of each part differ depending on the type of computer, but in the field of microcomputers, the format discussed by IEEE is being adopted as the standard format, and the bit length If is 32-bit single precision, the standard floating-point format is a 23-bit mantissa and 8-bit exponent and sign bit.

By the way, digital data Y in floating point format is
When expressed separately into K-bit mantissa vector component yM and L-bit exponent vector component yE, yM=[-1 for a, -2 for λ, aK-33690, ao
]yE=[bL-1・bL-2・bl-3...b
O]T. However, T is a transposition symbol.

Then, if the above digital data Y is converted into analog data and expressed, -(K-1) Y = (ax-12'+ a 2-1+ , ,
, +a02 1 ), and it can be seen that the analog data Y is represented by the product of the mantissa conversion term Ym and the exponent conversion term Ye. However, .

By the way, it goes without saying that the DA converter used to convert digital data Y into analog data Y requires a processing capacity that is determined by the bit length of the digital data. To convert digital data, a DA converter with a processing capacity of +L bits was simply used.

[Problems to be Solved by the Invention] The conventional DA converter described above has the following problems when performing DA conversion on digital data Y in floating point format consisting of a K-bit mantissa vector component yM and an L-bit exponent vector component yE. , a minimum of +L bits is required, and in the case of the microcomputer mentioned above, a 31-bit processing capacity is required, resulting in problems such as the complexity of the DA converter configuration and high manufacturing costs. There was a point.

[Means for Solving the Problems] This invention solves the above-mentioned problems, and converts the mantissa part into DA for digital data in floating point format consisting of a K-bit mantissa part and an L-bit exponent part. and an exponent 11D that converts the exponent part to DA.
A conversion circuit, and the mantissa part DA conversion circuit and the exponent part DA conversion circuit are configured in a series connection relationship in which one of the mantissa part DA conversion circuits uses the analog output of the other as a reference power supply. It is.

[Operation] With respect to digital data expressed in floating point format, the present invention converts analog data consisting of the product of a mantissa part DA conversion output and an exponent part DA conversion output, by converting the DA conversion output of one of them to the DA conversion output of the other. can be obtained by using it as a reference power supply or by subjecting it to A conversion.

[Example] Hereinafter, an example of the present invention will be described with reference to FIG. 1. FIG. 1 is a circuit diagram showing an embodiment of the DA converter of the present invention.

In FIG. 1, the DA converter l includes a mantissa DA conversion circuit 2 that converts the mantissa into DA, and a mantissa DA conversion circuit 2 that converts the mantissa into DA, and a mantissa DA converter circuit 2 that converts the mantissa into DA, regarding digital data Y in floating point format consisting of a K-bit mantissa and an L-bit exponent. The exponent part DA conversion circuit 3 performs DA conversion. In this embodiment, the mantissa part DA conversion circuit 2 and the exponent part DA conversion circuit 3 convert the analog output of the mantissa part DA conversion circuit 2 into the reference power supply V.
It is in a series connection relationship with rer.

That is, as is clear from the above formula, the analog data Y is the product Y of the mantissa conversion term Ym and the exponent conversion term Ye.
It is expressed as m-Ye, and the mantissa part and the exponent part mutually weight the gain.

Therefore, as shown in FIG.
is connected to the reference power supply terminal of the exponent part DA conversion circuit 3, so that the analog data Ym applied to the exponent part DA conversion circuit 3 as the mantissa part DA conversion output is multiplied by each term of the exponent part conversion term as a coefficient. and analog data Y is obtained.

In this way, the DA converter 1 converts digital data expressed in floating point format to be DA-converted into analog data Y consisting of the product Ym-Ye of the mantissa part DA conversion output Ym and the exponent part DA conversion output Ye. , the mantissa part DA
The conversion output Ym is the reference power supply V of the exponential part DA conversion output Ya.
Since the configuration is such that digital data having a bit length of +L bits as a whole is obtained by subjecting it to DA conversion with a processing capacity of K bits and DA conversion circuit 3 of L bits, , both of lower order 2
DA conversion can be performed by a combination of the DA conversion circuits 2.3, and thereby digital data having a large bit length can be converted to analog data at low cost.

It should be noted that, as in the DA converter 11 shown in FIG. 2, for the floating point color digital data Y consisting of a K-bit mantissa part and an L-bit exponent part, the connection mode is completely opposite to that of the above embodiment. It is also possible to perform DA conversion using the mantissa part DA conversion circuit I2 and the exponent part DA conversion circuit I3. In this case, the exponent part DA conversion output Ye becomes the mantissa part DA conversion output Ym.
Weighting is performed by DA conversion using the reference power source.

[Effects of the Invention] As explained above, the present invention converts analog data consisting of the product of the mantissa part DA conversion output and the exponent part DA conversion output with respect to digital data expressed in floating point format.
Since the configuration is such that the DA conversion output of either one is obtained by applying the DA conversion output of the other to DA conversion using the other DA conversion output as a reference power supply, digital data having a total bit length of +L bits can be processed with K bits of processing capacity. Two low-order DA converters, a DA converter circuit and an L-bit DA converter circuit.
DA conversion can be performed using a combination of conversion circuits, and this provides excellent effects such as converting digital data with a large bit length into analog data at low cost.

[Brief explanation of drawings]

FIG. 1 is a circuit diagram showing one embodiment of the DA converter of the present invention, and FIG. 2 is a circuit diagram showing another embodiment of the DA converter of the present invention. 1.11. ,,DA converter,2. +2. ,,mantissa DA
Conversion circuit, 3.13. ．． ．． Exponential part DA conversion circuit.

Claims (1)

[Claims] Regarding digital data in floating point format consisting of a K-bit mantissa part and an L-bit exponent part, it has a mantissa part DA conversion circuit that converts the mantissa part from DA to DA, and an exponent part DA conversion circuit that converts the exponent part from DA to DA. , these mantissa part DA converter circuit and exponent part DA converter circuit are DA converters in which either one uses the analog output of the other as a reference power supply and is connected in series.