IEEE MULTIDISCIPLINARY ENGINEERING EDUCATION MAGAZINE, VOL. 2, NO. 1, MARCH 2007 1 Analyzing Sensitivity in Electric Circuits Emmanuel A. Gonzalez, Graduate Student Member, IEEE, Martin Christian G. Leonor, Leonard U. Ambata, and Cornelio S. Francisco Abstract—This lecture note presents a well-known sensitivity analysis technique using Taylor Series expansion that can be used by students, engineers, and scientists in analyzing electric circuits. The technique presented, which is also available in [1], is discussed in detail, having a voltage-divider circuit as an example. Index Terms—Sensitivity, Taylor series expansion. S I. INTRODUCTION ensitivity analysis is very important especially in the fields of electrical and electronics engineering due to the fact that circuits practically have tolerance values. Some examples are high-performance synchronous circuit [9], power electronic circuits [10], and even transmission lines [11]. In fact, there are a number of texts dealing with circuit sensitivity analysis tools [2]-[6]. In reality, there is no such thing as a “perfect” component, and each element in an electric circuit will always have tolerance values, specifically minimum and maximum values. For example, if we consider a simple 100-ohm resistor, it will always have a tolerance value of, for example, five percent (5%). Thus, the minimum and maximum values of this passive electric component are 95 and 105 ohms, respectively, without also considering temperature effects. Even the most precise electric components available in the market have their own tolerances. Thus, there is a need to understand the basics of sensitivity and its application to electric circuit analysis. In this lecture note, the authors present a well-known sensitivity analysis tool that is used by engineers and scientists in analyzing the sensitivity of an electric circuit [1]. In this note, we are restricted to memoryless circuits, i.e. circuits that do not store any information at all, e.g. voltage and current. The basic analysis technique used in this note is rather straightforward and does not require too much calculus, unlike other techniques available in some science and engineering literature. A brief application of the Taylor series expansion is presented and is used as the key component in sensitivity analysis. II. WHAT IS SENSITIVITY ANALYSIS Sensitivity, generally, is the ability of an entity to be susceptible to stimulation. If we are talking about electric circuits, the sensitivity of that circuit is its ability to react with changes in certain parameters. It is a measure of how the system reacts to any stimulation, either internal or external. For example, sensitivity can be the measure of how a voltagedivider circuit composed of a constant dc voltage source and two resistors in series reacts in terms of the output voltage, with respect to perturbations in one resistor. When we say perturbation, we mean that the resistor changes from within its minimum and maximum levels. Sensitivity Analysis, on the other hand, is the study of the sensitivity in a system. If we are applying mathematical and scientific concepts in analyzing the variations in the output voltage of a voltage-divider circuit with perturbations in one resistor, then we can say that we are performing a sensitivity analysis on the circuit. Generally, sensitivity can be described as the limiting ratio of the fractional change in a measured value to the fractional change in value of a certain parameter. Thus, if the change in a measured value is too small even if the change in a certain parameter is large, then we can say that the system is not sensitive or has a low sensitivity. On the other hand, if the change in a measured value is too large even if the change in a certain parameter is small, then we say that the system is too sensitive or has a high sensitivity. Sometimes, we want to have a system that has a low sensitivity, such as electronic amplifiers and oscillators in the electronic systems perspective, and mechanical shock absorbers in mechanical systems perspective. Furthermore, a high sensitive system is needed especially if we are dealing with measurements and data acquisition. Examples are sensors that require higher sensitivities such as chemical sensors and vibration sensors. III. TAYLOR SERIES EXPANSION Before we proceed to the discussion on sensitivity analysis, a review on Taylor series expansion must be discussed first. Taylor series expansion is an expansion technique used in order to approximate a certain function. The Taylor series expansion of f ( x ) about x = x 0 is defined as The authors are with the Department of Electronics and Communications Engineering, College of Engineering, De La Salle University - Manila ({gonzaleze, leonorm, ambatal, franciscoc} @dlsu.edu.ph). Publisher Identification Number 1558-7908-122006_02 1558-7908 © 2007 IEEE Education Society Student Activities Committee (EdSocSAC) http://www.ieee.org/edsocsac IEEE MULTIDISCIPLINARY ENGINEERING EDUCATION MAGAZINE, VOL. 2, NO. 1, MARCH 2007 IV. SENSITIVITY ANALYSIS OF A SIMPLE VOLTAGE DIVIDER CIRCUIT R1 = 50 ± 5% Ω + vi = 5 V + R2 = 100 ± 2% Ω − vo − Fig. 1. An example of a voltage-divider circuit. f (x ) = f (x 0 ) + f ′ (x 0 ) (x − x 0 ) + + f ′′ ( x 0 ) ( x − x 0 ) 2 (1) 2! f ′′′ ( x 0 ) ( x − x 0 ) 3 3! + ". We now present a simple sensitivity analysis technique for a memoryless electric circuit, which is also available in [1]. Consider a voltage divider circuit shown in Fig. 1. The circuit is composed of a constant dc voltage source and two resistors in series. The output is the voltage across the 100-ohm resistor, while the input is a 5-volt dc voltage source. The value of the series resistance is 50 ohms with a tolerance of 5%, while the 100-ohm resistor has a tolerance of 2%. Thus, if the 50-ohm resistor has a 5% tolerance, then its value can vary from 47.5 to 52.5 ohms. On other hand, if the 100-ohm resistor has a 2% tolerance, then its value can vary from 98 to 102 ohms. What will happen to the output voltage if the resistances in the circuit change? Or in other words, how accurate will the output voltage be with changes in resistor values? To answer these questions, we need to determine first the change of values in the resistors. Let the difference between the actual resistor value and its ideal value be the change in resistance value, i.e. ∆R Ractual − Rideal . Thus, we can have ∆R1 R1,actual − R1,ideal and ∆R2 R2,actual − R2,ideal . If we The ideas involved in Taylor series for functions of one variable can be generalized. For example, the Taylor series of a two-variable function f ( x , y ) about x = x 0 and y = y 0 assume that ∆R1 and ∆R2 are small enough compared to their ideal values, then the Taylor series expansion v 0 ( R1 + ∆R1, R2 + ∆R2 ) = can be expressed as f (x, y ) = f (x 0, y0 ) + + ∂f ( x 0 , y 0 ) ∂f ( x 0 , y 0 ) ∂x +2 ∂x 2 2 ∂ f (a, b ) ∂x ∂y (x ∂ f (x 0, y0 ) ∂yx + (x − x0 ) + 2 (2) − x 0 ) (y − y 0 ) of 2 ∂R1 2 1 ∂ v0 ∂R2 ∆R2 2 ∆R1 + 2 ∂R1∂R2 ∆R1∆R2 (4) (y − y ) + ". 2 ∂f ( x 0 , y 0 ) ∂f ( x 0 , y 0 ) ∂x (y − y ) . (x 2 ∂R2 2 ∆R2 + " 2 v 0 ( R1 + ∆R1, R2 + ∆R2 ) can be approximated by neglecting second- and higher-order terms, thus having the expression 0 2 f (x, y ) ≈ f (x 0 , y0 ) + ∂y ∂ v0 2 v 0 ( R1 + ∆R1 , R2 + ∆R2 ) ≈ However, if one wants to make a linear approximation of (2), then the second- and higher-ordered terms can be eliminated, having the expression + 1 ∂ v0 ∂R1 ∂v 0 ∆R1 + 2 2 + − x0 ) (y − y ) 1 ∂ f (x 0 , y0 ) 2 ∂v 0 v 0 ( R1 , R2 ) + 0 ∂y 2 + (x 2 v 0 ( R1 , R2 ) + ∂v 0 ∂R1 ∆R1 + ∂v 0 ∂R2 ∆R2 . (5) Consequently, the deviation of the output voltage from its nominal value is then − x0 ) (3) 0 A good text containing information in Taylor series expansion can be found in [6] and [7]. v 0 ( R1 + ∆R1 , R2 + ∆R2 ) − v 0 ( R1 , R2 ) ≈ ∂v 0 ∂R1 ∆R1 + 1558-7908 © 2007 IEEE Education Society Student Activities Committee (EdSocSAC) http://www.ieee.org/edsocsac ∂v 0 ∂R2 ∆R2 . (6) IEEE MULTIDISCIPLINARY ENGINEERING EDUCATION MAGAZINE, VOL. 2, NO. 1, MARCH 2007 Thus, if we consider the circuit given in Figure 1, then substituting the values will result to v 0 ( R1 , R2 ) = = ∂v 0 ∂R1 R2 R1 + R2 10 3 = − 100 vi = 50 + 100 ( 5) 1 analysis by eliminating second- and higher-order terms, thus, having a linear approximation of the sensitivity. Furthermore, a voltage-divider circuit is used as an example to appreciate the said analysis technique. ACKNOWLEDGMENT V; (R 3 R2 + R2 ) 2 vi = − 100 ( 50 + 100 ) 2 (5) The authors would like give thanks to the anonymous reviewers for their helpful and insightful comments and suggestions about the contents of this lecture note. ≈ −0.0222 V / Ω ; and ∂v 0 ∂R2 (R 1 R1 + R2 ) 2 vi = 50 ( 50 + 100 ) 2 (5) ≈ 0.0111 V / Ω Furthermore, we can now obtain a linear approximation of the output voltage v 0 ( R1 , R2 ) in the worst case scenario, having the largest values of ∆R1 and ∆R2 with equal signs. Hence, ∆v 0 max can be obtained by using Eq. (5), having ∆v 0 max = ∂v 0 ∂R1 ∆R1,max + ∂v 0 ∂R2 ∆R2,max which is approximately equal to 0.0777 volts. With this value, it can be assumed that if the values of both resistors are at their maxima, the output voltage, i.e. voltage across the 100ohm resistor, will increase from 3.3333 V to approximately 3.4110 V. V. CONCLUSION In this lecture note, a simple sensitivity analysis technique used in the science and engineering fields is presented for its application to memoryless electric circuits. It was shown that a simple Taylor series expansion can be used in sensitivity REFERENCES [1] G. C. Temes and J. W. LaPatra, Introduction to Circuit Synthesis and Design, McGraw-Hill, Inc. 1977.W.-K. Chen, Linear Networks and Systems (Book style). Belmont, CA: Wadsworth, 1993, pp. 123–135. [2] S. W. Director and R. A. Rohrer, “The generalized adjoint network and network sensitivities,” IEEE Trans. Circuit Theory, Aug. 1969, pp. 318323. [3] J. W. Bandler and R. E. Seviora, “Computation of sensitivities for noncommensurate networks,” IEEE Trans. Circuit Theory, Jan. 1971, pp. 174-178. [4] L. L. Hayes and G. C. Temes, “An efficient numerical method for the computation of zero and pole sensitivities,” Int. J. Electron, Jan 1972, pp. 21-31. [5] R. L. Boylestad and L. Nashelsky, Electronics Devices and Circuit Theory, 7th ed., New Jersey: Prentice-Hall, Inc., 1999. [6] N. S. Nise, Control Systems Engineering, 3rd Ed., John Wiley & Sons, Inc., 2000. [7] M. R. Spiegel, Schaum’s Outline of Theories and Problems of Advanced Mathematics for Engineering and Scientists, McGraw-Hill, Inc., 1971. [8] E. Kreysig, Advanced Engineering Mathematics, 8th Ed., Singapore: John Wiley & Sons (Asia) Pte. Ltd., 2003. [9] X. Liu, M. C. Papaefthymiou, and E. G. Friedman, “Minimizing sensitivity to delay variations in high-performance synchronous circuits,” Proc. 1999 Design, Automation and Test in Europe Conference and Exhibit, Mar. 19-12, 1999 [10] N. Fermia and G. Spagnulo, “Genetic optimization of interval arithmetic-based worst case circuit tolerance analysis,” IEEE Trans. Circuits and Systems I: Fundamental Theory and Applications, vol. 46, no. 12, Dec. 1999, pp. 1441-1456. [11] J.-F. Mao and E. S. Kuh, “Fast simulation and sensitivity analysis of lossy transmission lines by the method of characteristics,” IEEE Trans. Circuits and Systems I: Fundamental Theory and Applications, vol. 44, iss. 5, May 1997. 1558-7908 © 2007 IEEE Education Society Student Activities Committee (EdSocSAC) http://www.ieee.org/edsocsac