Agilent Measurement Journal - Issue 4 - 2008 - (Page 62)

Tip: Resolution is often taken as an indication of measurement accuracy; however, overall accuracy can be one of the most difﬁcult speciﬁcations to determine from the speciﬁcations presented on a data sheet. In fact, a resolution value quoted as “bits” does not include any indication of the noise or distortion levels that effectively diminish a device’s ability to distinguish between discrete levels. A better evaluation of system performance with respect to measurement accuracy is provided by the “effective bits” or “effective number of bits” (ENOB). Tip: ENOB speciﬁcations should be compared carefully because the value can be highly dependent on amplitude and frequency. An ENOB value should be quoted with the signal input level used, sampling rate and bandwidth over which the speciﬁcation is measured. From the example shown in Figure 3, if quoted only at 10 MHz, the 14-bit digitizer shows more effective bits than the 12-bit digitizer quoted at 50 MHz. This is clearly misleading because the 12-bit system shows better performance from about 25 MHz and beyond. Determining the effective number of bits When assessing a system in terms of effective bits, all error sources are included. Evaluation of system performance can be made without considering the individual error sources. Theoretically, a 16-bit digitizer will be able to measure one part in 65,536; however, operating under real-world conditions, a good 16-bit board has at most 13.5 bits of accuracy. What’s more, a poorly designed 14-bit board may perform less accurately than a well-designed 12-bit system. This is especially true when reviewing effective bits as a function of input signal frequency. Applying SINAD The signal-to-noise ratio (SNR) is the ratio, expressed in decibels, of the root mean square (RMS) value of the input signal at the output to the RMS value of the sum of all other spectral components. In contrast, the SINAD ratio, also expressed in decibels, is the ratio of the RMS value of the input signal at the output to the RMS value of all of the other spectral components, including harmonics. Because SINAD includes both noise and distortion, it can be used to directly calculate the effective number of bits. Tip: Compare SINAD speciﬁcations carefully because the value can be highly dependent on amplitude and frequency. As with ENOB, the value should be quoted with the signal input level used, sampling rate and bandwidth over which the speciﬁcation is measured. 10.50 10.00 9.50 9.00 8.50 8.00 7.50 7.00 6.50 6.00 0 10 20 30 40 50 60 70 Input frequency (MHz) 80 90 100 Effective number of bits Factoring in sources of sampling error Analogous to bit resolution, sampling rate, quoted in samples per second (Sa/s), is often regarded as a measure of temporal accuracy. As with resolution, however, the given sampling rate of a digitizer does not indicate the levels of temporal noise added to the signal. Sampling rate and all device-timing functions of a digitizer are driven by clock circuitry, either internal or external, that is assumed to be stable. Commonly based around a crystal oscillator, timing circuits are susceptible to frequency drift and spurious signals. 12-bit digitizer 14-bit digitizer Figure 3. This diagram uses ENOB to compare commercial 14-bit and 12-bit digitizers that have 100 MSa/s sampling rates, 2 to 100 MHz input frequency range and input signal amplitude at 97.5 percent of 1V full scale. The ENOB value can be calculated from the signal-to-noiseand-distortion (SINAD) ratio: ENOB SINAD – 1.76 6.02 62 Agilent Measurement Journal

Table of Contents Feed for the Digital Edition of Agilent Measurement Journal - Issue 4 - 2008

Agilent Measurement Journal - Issue 4 - 2008 Delivering Confidence through Compliance with Standards Contents Emerging Innovations Trying Early Device Implementations at the IEEE 1588 PlugFest Achieving Greater Confidence in Measurement Accuracy through Consistency in Calibration Services Overcoming the Challenges of RFID Component Testing 3GPP LTE: Introducing Single- Carrier FDMA Ensuring Reliable Operation and Performance in Converged IP Networks Testing Storage Area Networks and Devices at 8.5-Gb/s Fibre Channel Rates Choosing an Appropriate Calibration Method for Vector Network Analysis Making Traceable EVM Measurements with Digital Oscilloscopes Exploring Terahertz Measurement, Imaging and Spectroscopy: The Electromagnetic Spectrum’s Final Frontier Interpreting Quoted Specifications when Selecting Digitizers

Agilent Measurement Journal - Issue 4 - 2008

For optimal viewing of this digital publication, please enable JavaScript and then refresh the page. If you would like to try to load the digital publication without using Flash Player detection, please click here.