#TITLE_ALTERNATIVE#

• Main Author: RUSDIAN PUTRA (NIM : 13106115); Pembimbing, : Dr. Ir. Zainal Abidin, ISNAINDA
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/16117
• Institution: Institut Teknologi Bandung

RMS value is an important number because it is often used in various types of measurements, such as machine vibration measurement. The measurement was conducted to determine the operating conditions of the machine by comparing the RMS value obtained from the measurements with that of ISO standards. In many cases, the RMS values obtained from measurements are differ from the real RMS value, causing <br /> <br /> <br /> uncertainty in determining the operating condition of the machine. Therefore, a research must be done to pinpoint the cause of RMS error during measurement. <br /> <br /> <br /> In this research, an analysis of RMS error is perform for continuous and discrete signals. The analysis of continuous signals is conducted by deriving the equation of the RMS error. Based on the obtained equation, three variables related to RMS error are obtained, namely m (non negatif integer numbers), q (rational numbers from 0 to 1), and &#981; <br /> <br /> <br /> (signal phase). The influence of these variables are analyzed by drawing the RMS error as a function of these variables in three-dimensional graphics. In addition, analysis of discrete signals is conducted by varying the four parameters, namely N (number of samples), r (ratio between signal frequency and frequency sampling), Nr (the ratio between length of time record and signal period), and &#981; (signal phase). <br /> <br /> <br /> Based on this analysis, it can be concluded that the RMS error occurred when the time record is not an integer multiple of the signal period. For continuous signals, in order to obtain RMS error less than 1%, the measurement should be done with time record of at least eight times the signal period. For discrete signal, in order to obtain RMS error less than 1%, the measurement should be done with time record of at least nine times the signal period. As an example, for 100 samples, the RMS error of discrete signals will be less than 1% only when the ranges of r is from 0,0829 to 0,4171.