Entwurf und Implementierung von digitalen Fehlerkorrekturverfahren für Mixed-Signal-Frontend-Schaltungen

Various non-ideal effects as well as environmental influences can affect the transfer characteristics of analog microelectronic circuits. In order to reduce the consequences of these effects, additional circuitry or different compensation techniques are usually used. Unfortunately, most of these methods have several drawbacks. They significantly increase the design time and costs of the circuit and lead to a higher consumption of energy and chip area. For this reason, an innovative method of error detection and correction is developed in this work, which is based on digital calibration and particularly well suited for the application in mixed-signal systems. With this approach, the error correction is no longer performed by complex, analog compensation techniques but by additional digital signal processing. First of all, a suitable test signal is fed into the analog circuit to make different errors visible. Subsequently, the test signal is evaluated in the digital part of the system by adaptive filters and appropriate algorithms to determine the kind and size of present errors. This information is used afterwards to eliminate any errors from the test and the useful signal. With this procedure, the disadvantages of analog compensation methods are removed with only a low overhead on digital circuitry. Furthermore, it is not only possible to correct static as well as time-variant errors but the circuit also gains the ability to monitor itself. As a result, the whole system becomes more reliable and robust. In order to demonstrate the performance of the introduced approach, it is used to correct different gain- and offset-errors in a generic sensor interface circuit. For this purpose, a simulation-based and a practical verification are performed to analyze various application and error cases. While a simulation model of the interface-circuit is used for the first task, a printed circuit board with electronic components is used for the second one. With both forms of verification, the successful application of the developed method can be shown.