Nowadays nearly all measurements of every quantity are carried out through digital measuring systems. These are usually composed by one or more sensors, signal conditioning circuits, analog to digital converters (ADCs) and computers or microprocessors, where digital signal processing is performed to extract the relevant information from the ADC output codes.
The accuracy of these measurement instruments is strongly dependent on the quality of calibration techniques and signal processing algorithms. When the systematic errors are time invariant, they can be characterized during a calibration process and mathematically removed during the measurement procedure.
One of the key components in modern measuring systems is the Analog-to-Digital Converter. Their performance has a determinant impact on the performance of digital instruments, and on the quality of the measurement systems. Traditionally, digital instruments are designed by considering ADCs as ideal components affected only by sampling and quantization errors. Unfortunately, actual errors in ADCs transfer function, strongly affect the overall accuracy and bandwidth of the systems where they are included.
To implement powerful calibration techniques the behavior of ADCs has to be well described. Dynamic characterization of ADCs is usually performed by the code density histogram, DFT or sine-fitting tests, in which the participants in this project have ample knowledge and internationally published work.
In this project, we propose to apply our experience on ADC testing and use signal processing techniques to built high accuracy digital measurement systems. We will demonstrate these techniques by implementing an impedance meter and by performing error correction in the results of an electrocardiograph. Simultaneously high frequency ADC testing techniques will be improved.
The impedance measurement system will be based on the use of ADCs included in general-purpose digitizers and sine-fitting techniques. The system will perform from DC to tens of MHz. Calibration techniques will remove the influence of systematic errors. They will be based on a complete characterization of the digitizing channels and implementation of the corresponding error correction. The unavoidable influence of noise in any experimental setup will be reduced by the use of sine-fitting techniques. This will allow the construction of high accuracy, but not expensive instruments. Preliminary results already obtained by using a medium cost 12 bit PC data acquisition boards show that the technique is very powerful and can lead to measurement accuracy comparable with that obtained by sophisticated very high cost dedicated impedance measurement equipment. The first results for impedances with magnitudes and phases near 300ohm and 45º shown standard deviations close to 0.01ohm and 0.01º.
Similar techniques will be used to remove the power-line interference in electrocardiogram results. Sine-fitting or frequency domain techniques (IpDFT) will be used.
|Start Date: 01-02-2004|
|End Date: 01-02-2007|
|Team: António Manuel da Cruz Serra, Pedro Miguel Pinto Ramos, Ricardo Queirós, Francisco André Corrêa Alegria, Manuel Fonseca da Silva, Raúl Daniel Lavado Carneiro Martins|
|Groups: Instrumentation and Measurements – Lx|
|Local Coordinator: António Manuel da Cruz Serra|