Measurement Journal Issue 6

Another interesting point is the concentration of spectrum energy in the frequency range below 6 MHz. Increasing the memory size reveals that this jitter is low-frequency RJ and not low-frequency PJ. Additional information can be gleaned from the BER bathtub window. As shown in Figure 6, the EZJIT Plus software ﬁrst analyzes the RJ (rms) value from the jitter spectrum. It then uses that value to extrapolate the bathtub curve to achieve the desired BER level. Making accurate TJ measurements Measurements showed that the SATA DUT generated a large RJ component in the lower frequency range, but there were doubts about the accuracy of the jitter software. The best way to judge the accuracy is to compare a direct jitter measurement with the estimated TJ from the EZJIT Plus software. A BERT is the only instrument that can measure TJ directly. To ensure valid comparisons between different instruments (e.g., a BERT and a scope), it is important to use either the CDR function or the observed jitter transfer function (OJTF) of the CDR. Because the SATA device cannot provide a reference clock to both the scope and the BERT, it is necessary to use similar settings for the tunable CDR in the BERT and the CDR function in the EZJIT Plus software. In this case, the BERT was an Agilent J-BERT N4903A highperformance serial BERT. In the EZJIT Plus software, the CDR function was set to the second order with 1.55-MHz bandwidth and 0.707 damping factor. To create comparable measurements, the hardware parameters inside the J-BERT were adjusted as closely as possible to the EZJIT Plus settings (Figure 7). Poor ﬁ t with bathtub curve Good ﬁ t with bathtub curve Figure 6. With increased memory size, the larger RJ (rms) value enables a smoother ﬁ t between measured and extrapolated sections of the BER bathtub curve In the BER bathtub diagrams, the blue curves at the top show the measured sections of the histogram database. The light gray curve is the extrapolation section. As mentioned earlier, the RJ (rms) value is calculated from the jitter spectrum. If the waveform parametrics cannot be modeled, or measured well enough, then the extrapolation section will not be properly aligned with the measured section of the curve. In contrast, a correct RJ (rms) value extracted from the jitter spectrum will provide an accurate ﬁt with the measured histogram tail. Figure 7. Hardware CDR setup parameters within the J-BERT The results are shown in Figures 8 and 9. To minimize the long wait associated with measurements made at the 10-12 BER level, both the J-BERT and EZJIT Plus measurements were made with a 10-9 BER level. Comparing the results, the J-BERT measured a TJ of 251 ps while the EZJIT Plus software calculated a TJ of 259.67 ps. 37 Agilent Measurement Journal

Table of Contents Feed for the Digital Edition of Measurement Journal Issue 6

Measurement Journal Issue 6 Using a Few Words to Convey a Deeply-Held Spirit Charting a Unique Path to Technology Innovation Contents Announcing First Commercial GC/MS Metabolites Library Assessing Nonlinear Behavior Testing Combines Form and Function Testing Multiplay Networks Higher Sensitivity LC/MS Easing Military Radio Test Stepping Up Calibration Accuracy Increasing Spectrum Analyzer Bandwidth 3GPP Tool Supports TS 36 Standards The Olympic Story Spotlights Agilent's Continuing Contribution to Technology Innovation Ensuring a Level Playing Field in Competitive Sports GPS: Coming of Age Resolving Design Issues in HSPA Mobile Devices Achieving Accurate Results with Oscilloscope-Based Jitter-Analysis Software RF Handheld Testers Guarantee Traffic Stability Under Olympic-Sized Stress Conditions Utilizing LAN-Based Instrumentation to Measure Total Harmonic Distortion in Remote Facilities IMT-Advanced: 4G Wireless Takes Shape in an Olympic Year EPO Doping: A Speed Engine, Turbo-Charged by Chemistry

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.