Design and analysis of analog filters by statistical methods

• Published in: 2018 IEEE International Instrumentation and Measurement Technology Conference (I2MTC) pp. 1 - 6
• Main Authors: Rocha, Fabiano Sanches, Biajo, Vitor Henrique Madureira, de Oliveira Mota, Hilton, Maia, Carlos Andrey, Vasconcelos, Flavio Henrique
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-05-2018
• Subjects: Band-pass filters; electronic filter; Gain; Low pass filters; Monte Carlo Method; Resistors; sensitivity; Uncertainty
• DOI: 10.1109/I2MTC.2018.8409856

Analog filters are widely used in measurement systems to dissociate signal from noise. However, when designing these or other electronics whose components will be assembled in an industrial process, one should consider that resistors, capacitors, operational amplifiers and other discrete items may not arises on the circuit boards with their nominal values taken from the color code or other plain sources. Keeping this in mind, it is important to consider a range of component possible values produced by the manufacturing process, in order to evaluate the behavior of some electronic devices. In this work, the intention was to investigate circuits such as low-pass filters employing the Monte Carlo method in order to figure out the influence of component values, taken now by their expected values, coverage interval and other statistics obtained from the probability density function, over some important circuit parameters like gain, sensitivity, etc. The results indicate that passband, stopband and cutoff are influenced differently due to capacitor and resistor tolerance.