US 5,500,644 AGranted

Procedure and device for self-calibrating analog-to-digital conversion

AssigneeThomson-CSF

InventorsJean Denjean, Jean-Pierre Mann

Life of this patent

Estimating the term…

Loading patent…

Abstract

The invention concerns rapid, precise analog-to-digital converters (ADC). To achieve rapid conversion, the invention uses a fast though imprecise ADC, whose numeric output value is used to address a memory location in a first memory block containing correction values. Calibration measurements are inserted between conversion measurements, the time intervals being sufficient to establish a stable reference voltage at the output of a generator. The switching between conversion and calibration measurements is made using a switch. During calibration, the differences between the conversion results and the real values are written into a second memory block at addresses corresponding to the numeric output values of the ADC. Means of control and calculation enable sequencing and switching of the roles of the calibration and correction memory blocks at the end of a calibration.

Claims13

Dependent claims

Process for conversion of an analog value X, which can range between limits X.sub.1 and X.sub.2, into a numeric value N ranging between limits N.sub.1 and N.sub.2.spsb.n and comprising a sequence of bits, each bit having a weight depending on its position in the sequence, the values N.sub.1 and N.sub.2.spsb.n corresponding to the values X.sub.1 and X.sub.2 and 2.sup.n -1 intermediate values N.sub.1 between N.sub.1 and N.sub.2.spsb.n each corresponding to an intermediate value Xi, the pth numeric value after N.sub.1 corresponding to the value:

##EQU3## the process being implemented using a measuring circuit to which is input an instantaneous value X.sub.j of the analog value X and which outputs a corresponding numeric value N.sub.j, the process being of the type in which a previous calibration is used to assign a correction value h.sub.j to each output numeric value N.sub.j, wherein the calibration measurements are repeated throughout the conversion process without interrupting the conversion measurements, and wherein two distinct sets of correction values are created, a first and a second, the first created with the correction values h.sub.j1 of a previous calibration and the second created with the correction values h.sub.j2 of a current calibration, a current calibration being completed when all the correction values h.sub.jr that created the second set have been settled and wherein the first and second sets are swapped each time when a calibration is completed.

Device for conversion of an analog value X, which can range between limits X.sub.1 and X.sub.2, into a numeric value N ranging between limits N.sub.1 and N.sub.2.spsb.n, the values N.sub.1 and N.sub.2.spsb.n comprising a succession of bits each bit having a weight depending on its position in the succession, the values N1 and N.sub.2.spsb.n corresponding to the values X.sub.1 and X.sub.2 and 2.sup.n -1 intermediate values N.sub.i between N.sub.1 and N.sub.2.spsb.n each corresponding to an intermediate value X.sub.i, a pth numeric value after N.sub.1 corresponding to the value:

##EQU4## the device including an analog-to-digital converter receiving an instantaneous input value X.sub.j and generating at its output a numeric value N.sub.j, the device including a first block of memory locations, an adder receiving the value output by the converter and also the correction h.sub.j obtained from a memory location in the first block whose address is determined by the numeric order of magnitude N.sub.j, a value N.sub.jk output by the device being equal to the sum of the values N.sub.j and h.sub.j, wherein the device includes upstream of the converter a controlled analog switch with at least two positions, a measuring position and a calibration position, the switching from one position to the other being controlled by a sequencer, and when the switch is in measuring position, the converter receives the instantaneous value X.sub.j to convert and outputs the numeric value N.sub.j which is applied to the input of the adder, the other input of the adder receiving the value h.sub.j from the first memory block, the adder generating an output equal to N.sub.jk =N.sub.j + h.sub.j, and when the switch is in calibration position the converter receives an instantaneous reference value X.sub.jr from a digital-to-analog converter and generates an output value N.sub.jr which is applied to the input of a microprocessor, this microprocessor updating the memory location having the address N.sub.jr in a second block of memory locations, the sequencer controlling the inversion of the first and second memory blocks once the calibration phase is complete, the second memory block then becoming the correction memory block and the first block becoming that receiving the calibration values.

Device for conversion of an analog value X, which can range between limits X.sub.1 and X.sub.2, into a n-bit numeric value N ranging between limits N.sub.1 and N.sub.2.spsb.n in steps of P.sub.p equal to:

(x.sub.2 x.sub.1)/2.sup.n,
the device including means of conversion handling a first group M.sub.1 of n.sub.1 bits, representing the most significant bits, the step P.sub.1 of this first group being equal to:
(X.sub.2 -X.sub.1)/2.sup.n.sbsp.1.sup.,
a second group M.sub.2 of n.sub.2 bits, representing the bits of significance immediately inferior, the step p.sub.2 of this second group being equal to:
(X.sub.2 -X.sub.1)/2.sup.(n.sbsp.1.sup.+n.sbsp.2.sup.-n'.sbsp.2.sup.),
where n'.sub.2 represents the number of the more significant bits whose weight is equal to that of the least significant bits of the first group, and so on until the last group M.sub.p representing the least significant bits whose step p.sub.p is equal to the measuring step of the device:
(X.sub.2 -X.sub.1)/2.sup.(n.sbsp.1.sup.+n.sbsp.2.sup.+n.sbsp.p.sup.-n'.sbsp.2.sup.- n'.sbsp.3.sup.-n'.sbsp.p.sup.),
the device including first means which receive the input value X at an instant where its instantaneous value is X.sub.j, and convert the value X.sub.j into a numeric value N.sub.j1 expressed by the bits of the first group, the device including, if p is greater than 1, first means of converting the numeric value N.sub.j1 into an analog value X.sub.j1, means of subtracting the analog values X.sub.j and X.sub.j1, and means of converting the difference X.sub.j -X.sub.j1 into a numeric value N.sub.j2 expressed by the bits of the second group, and means of converting the numeric value N.sub.j2 into an analog value X.sub.j2, means of calculating the difference (X.sub.j -X.sub.j1)-X.sub.j2, and means of converting this difference (X.sub.j -X.sub.j1)-X.sub.j2 into a numeric value N.sub.j3 expressed by the bits of the third group, and so on until the means of converting the difference (X.sub.j -X.sub.j1 -X.sub.j2 -X.sub.jp -1)-X.sub. jp into a numeric value N.sub.p expressed by the bits of the last group, and, if p is greater than 1, means receiving the numeric values N.sub.j1 to N.sub.jp and returning a numeric value N.sub.j which is the sum of the weights of each of the bits of each of the p groups of bits, wherein the device also includes a switch, means of control and calculation, two memory blocks for calibration and correction, means of controlling the switching of the switch, the switch having a first and second input and an output, the switch being controlled on each of its inputs by the means of control and calculation, on the first input the switch being connected to the means of conversion of the input value X, these means of control and calculation adding to the output value of the means of conversion all the p correction values stored in the storage locations of the first memory block, each correction coming from a sub-block assigned to a group of bits (M.sub.1, M.sub.2 . . . M.sub.p), the address of the storage location in the sub-block being determined by the output value of the group of bits, and on the second input the switch transmitting to the means of conversion a reference voltage provided by a voltage generator controlled by the means of control and calculation, these means of control and calculation being programmed to write in the storage locations of the second memory block a set of p correction values, each value being assigned to a sub-block of the calibration block and each sub-block receiving at the address equal to the numeric output value of one of the groups of bits (M.sub.1, M.sub.2 . . . M.sub.p) a value equal to the difference between the value represented by this numeric output value and the numeric value which would have been output had the measurement been exact, the means of switching inverting the roles of the calibration and correction memory blocks when it receives a command from the means of control and calculation.

5 more elements

Device for conversion of an analog value X, which can range between limited X.sub.1 and X.sub.2, into a numeric value N ranging between limits N.sub.1 and N.sub.2.spsb.n, the values N.sub.1 and N.sub.2.spsb.n comprising a succession of bits each bit having a weight depending on its position in the succession, the values N.sub.1 and N.sub.2.spsb.n corresponding to the values X.sub.1 and X.sub.2 and 2.sup.n -1 intermediate values N.sub.i between N.sub.1 and N.sub.2.spsb.n each corresponding to an intermediate value X.sub.i, a pth numeric value after N.sub.1 corresponding to the value:

##EQU5## the device including an analog-to-digital converter receiving an instantaneous input value X.sub.j and generating at its output a numeric value N.sub.j, the device including a first block of memory locations, an adder receiving the value output by the converter and also the correction h.sub.j obtained from a memory location in the first block whose address is determined by the numeric order of magnitude N.sub.j, a value N.sub.jk output by the device being equal to the sum of the values N.sub.j and h.sub.j, wherein the device includes upstream of the converter a controlled analog switch with at least two positions, a measuring position and a calibration position, the switching from one position to the other being controlled by a sequencer, and when the switch is in measuring position, the converter receives the instantaneous value X.sub.j to convert and outputs the numeric value N.sub.j which is applied to the input of the adder, the other input of the adder receiving the value h.sub.j from the first memory block, the adder generating an output equal to N.sub.jk =N.sub.j +h.sub.j, and when the switch is in calibration position the converter receives an instantaneous reference value X.sub.jr from a digital-to-analog converter and generates an output value N.sub.jr which is applied to the input of a microprocessor, this microprocessor updating the memory location having the address N.sub.jr in a second block of memory locations, the sequencer controlling the inversion of the first and second memory blocks once the calibration phase is complete, the second memory block then becoming the correction memory block and the first block becoming that receiving the calibration values;
wherein the means of conversion includes a sample and hold circuit, an analog-to-digital converter for the most significant bits, an analog-to-digital converter for the least significant bits, a digital-to-analog converter and a subtractor, the sample and hold circuit receiving an input voltage of the means of conversion and feeding the high bit converter and an input of the substractor, a second input of the subtractor being fed by the digital-to-analog converter receiving the numeric output of the high bit converter, one output of the subtractor feeding the low bit analog-to-digital converter.

Dependencies are inferred from the claim text.