Averaging and recording of digital deep-level transient spectroscopy transient signals
Journal Articles
Overview
Research
Identity
Additional Document Info
View All
Overview
abstract
A new approach to the signal processing in digital deep-level transient spectroscopy (DLTS) systems is introduced. The problems of signal recovery from noise and efficient data storage are addressed separately from the transient signal analysis. As a result of this approach, an improved digital averaging scheme for DLTS signal recovery from noise and transient data storage is proposed. We have shown that the combined action of two complementary digital averaging techniques can improve the DLTS digital signal processing. Pseudologarithmic time averaging is efficient in reducing the number of processed data points and improving the signal-to-noise ratio for the high frequency noise components and in the transient tail. We demonstrate that the normalized errors in the magnitude measurements introduced by this type of averaging remain below 1% for pure logarithmic and below 0.1% provided that the logarithmic averaging intervals are further divided into five or more equal parts. Continuous time averaging is well suited for improving overall signal-to-noise ratio, for continuous display and processing of data, and it is more efficient in using the computer resources. The described combination of hardware and software tools for the implementation of these techniques supplies continuously fresh data and does not require any synchronization with the main computer program. Compared to other digital DLTS systems, this new approach offers an improved short delay time resolution, improved signal-to-noise ratio, and more efficient data storage. At the same time it offers real-time observation of essentially noise-free transients and, like analog systems, a real-time display of the DLTS scan. The real-time display means that the displayed transient and DLTS scan can be updated on the computer screen after each pulse even for a very large number of averaged pulses, and the result can be predicted well before the acquisition process reaches this number. The proposed techniques allow one to combine the powerful transient analysis of the digital DLTS methods with the sensitivity and the convenience of the analog methods. The described averaging and data reduction techniques are intended primarily for DLTS data processing, but the same principles can be useful for many other physical experiments involving transient data analysis.
