Browse Prior Art Database

Analogue-to-Digital Converter with Improved Accuracy

IP.com Disclosure Number: IPCOM000117436D
Original Publication Date: 1996-Feb-01
Included in the Prior Art Database: 2005-Mar-31
Document File: 4 page(s) / 89K

Publishing Venue

IBM

Related People

Martin, FJ: AUTHOR [+2]

Abstract

The conversion of an analogue electronic signal from a physical transducer (e.g., strain gauge, thermocouple) or from a voltage source to an input acceptable to a digital system (e.g., microcontroller) requires an Analogue to Digital Conversion (ADC) stage. Analogue-to-Digital Converters are available with a number of levels of resolution (e.g., 8 bit, 14 bit). The cost of the ADCs increases markedly with increased speed or resolution. For a given application, the use of a higher resolution device would in many cases be advantageous but would be prohibitive on cost.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Analogue-to-Digital Converter with Improved Accuracy

      The conversion of an analogue electronic signal from a physical
transducer (e.g., strain gauge, thermocouple) or from a voltage
source to an input acceptable to a digital system (e.g.,
microcontroller) requires  an Analogue to Digital Conversion (ADC)
stage.  Analogue-to-Digital Converters are available with a number of
levels of resolution (e.g., 8  bit, 14 bit).  The cost of the ADCs
increases markedly with increased speed or resolution.  For a given
application, the use of a higher resolution device would in many
cases be advantageous but would be prohibitive on cost.

      The method described here enables a low cost ADC (e.g., 8 bit)
to provide a higher resolution digital signal when used to measure a
"DC" voltage input.  This sort of input is typical of many of the
measurements required.  The term DC is used to imply that the signal
remains constant for the duration of the signal measurements.

      A known cyclic perturbation (noise signal) is added to the
input signal.  The combined signal is asynchronously sampled by the
ADC.  If the perturbing waveform has the correct amplitude,the
resultant set of data from the ADC will show a statistical
distribution which will be biased depending on the input level.
Statistical averaging of this data set can be used to determine the
input DC value to a greater resolution than the intrinsic resolution
of the ADC.

      There is clearly a trade off of temporal resolution and a
processing overhead by use of this approach.  This is offset by the
low temporal resolution required by many applications, the de facto
presence of some processing capability in most digital compatible
systems, and the perturbing waveform often being available at no cost
by allowing noise present in other parts of the system (e.g., from
DC-DC converters) to "leak" into the required input.

      Assume a noise signal R is a randomly varying triangular
waveform with amplitude A, that is (refer to original document for
mathematical formula).

      The probability of R being sample as any value between -A and
+A is therefore uniform.  Let t...