Dynamic testing of an ADC for real application input

Abstract

This paper presents a new approach for determination of nonlinearity and Effective Number of Bits (ENOB) of an Analog to Digital Converter (ADC) for application input. Many times the input to ADC in an application is other than standard signals such as sine wave or triangular wave. Parameters of ADC determined using dynamic testing with standard signals are not useful if input signals are different. Different sinusoidal components of application input can be determined by spectrum analyzer. Application input can be generated by downloading data points created by computer in an Arbitrary Waveform Generator (AWG). Differential Nonlinearity (DNL) of an ADC is determined using deviation of actual histogram from reference or ideal histogram. Further estimation of Integral Nonlinearity (INL) is done from summation of DNL. ENOB is determined by estimating ideal rms error and actual rms error. These rms errors are computed by taking difference of sampled ADC input value available in computer and corresponding ADC output. Simulation results for 5 and 8 bit ADCs are reported and experimental results for an actual 8 bit ADC are also reported.. Comparison of simulation results by proposed method is done for standard sine wave, triangular wave and application inputs. It is expected that this work will initiate research toward ADC testing using application input.