抖動(dòng)和眼圖的視覺分析

1、 抖動(dòng)和眼圖的視覺化分析 什么是抖動(dòng)？TIE 為信號(hào)相對于標(biāo)準(zhǔn)時(shí)鐘或者標(biāo)準(zhǔn)信號(hào)的定時(shí)誤差TIE 在高速數(shù)字系統(tǒng)中即為抖動(dòng)抖動(dòng)為實(shí)際數(shù)據(jù)與其理想位置的時(shí)間偏差0.0ns 0.990ns 2.000ns 2.980ns 4.000ns-0.010ns 0.000ns -0.020ns TIE 0.000ns 什么是眼圖？眼圖是怎么形成的？ Random Jitter(隨機(jī)抖動(dòng)隨機(jī)抖動(dòng)符合高斯型分布直方圖(估計(jì) pdf(數(shù)學(xué)模型 抖動(dòng)峰峰值=無窮大無界！內(nèi)部熱能現(xiàn)象Flicker Noise, Shot Noise熱能的原子與分子振動(dòng)分子的解體外部的宇宙射線 Deterministic Jitter

2、(確定性抖動(dòng) 確定性抖動(dòng)是非高斯分布并且有界 Periodic Jitter(周期性抖動(dòng) TIE 隨時(shí)間的變化是重復(fù)的、周期性的Periodic jitter和相位調(diào)制(PM是等效的系統(tǒng)時(shí)鐘（抖動(dòng)頻率在MHz 量級）開關(guān)電源（抖動(dòng)頻率在KHz 量級） Duty Cycle distortion(占空比失真上升時(shí)間和下降時(shí)間不對稱或者測試時(shí)參考電平選擇不當(dāng)0.0v Inter-Symbol Interference(碼間干擾抖動(dòng)DDJ 或PDJ 數(shù)據(jù)相關(guān)性抖動(dòng)或碼型相關(guān)性抖動(dòng), 和ISI 的術(shù)語是等價(jià)的.碼型是如何影響隨后的比特位的？由于傳輸鏈路的效應(yīng)、反射等 換個(gè)角度看抖動(dòng)，時(shí)域 看看我們有了

3、什么視角？ 7/7/201611抖動(dòng)視覺化時(shí)間趨勢圖直方圖告訴了我們分布，但是只有統(tǒng)計(jì)特性，缺少了時(shí)間信息時(shí)間趨勢圖可以直觀告訴我們波形里是否有特定頻率的調(diào)制下圖為5個(gè)周期 SSC 30khz 抖動(dòng)視覺化 Gaussian Random Noise Sinusoidal Jitter7/7/201612 7/7/201613抖動(dòng)視覺化頻譜圖從頻域上觀測抖動(dòng) 抖動(dòng)中決定性的頻率成分會(huì)在譜線上明顯超出噪底 哪個(gè)眼圖好？哪個(gè)直方圖好？視覺化眼圖和抖動(dòng)的問題？ 浴盆曲線誤碼率是關(guān)鍵vs. UI張開程度For a given position in the time theres a given pro

4、bability of error “BER ”, Bit Error Ratio1 UI 基于示波器分析的浴盆曲線Rj /Dj與Tj BERAssume bi-modal distribution (dual-Dirac, measure Tj at two BER 抖動(dòng)類型分析抖動(dòng)分離為誤碼產(chǎn)生的根本原因提供了更精確的定位和分析方法抖動(dòng)分析方法，參照T11 MJSQ，已經(jīng)被工業(yè)界廣泛接受Constituent Components of Jitter= Unbounded= Bounded 7/7/201618Jitter Visualization Bathtub PlotNote th

5、e eye closure of System I vs. System II due to the RJ-RJ is unbounded so the closure increases as BER level increasesSystem I has .053UI of RJ with no PJSystem II has .018UI of RJ and .14UI of PJ 5 and 10MhzSystem ISystem ISystem IISystem II Tektronix -Innovators of Jitter Analysis1998First Real-Tim

6、e Scope Based Jitter Analysis Software2002 Invented SW Based PLL Clock Recovery and the Spectral Approach for Jitter Separation2004Invented RT Eye rendering on a Real Time Scope2004-First vendor to support both modeled (Dual-Dirac and measured (Spectral jitter methods2005-Invented measurements with

7、Jitter and Noise reconciliation2011-First scope vendor with BUJ support2015RT Noise Analysis and Sampling BER and PDF Mask Testing 抖動(dòng)和眼圖的視覺化眼圖怎么切割的？時(shí)鐘決定！ TIE 抖動(dòng)需要參考時(shí)鐘參考時(shí)鐘提取的過程就是時(shí)鐘恢復(fù)參考時(shí)鐘有幾種確定的方式:Constant Clock with Minimum Mean Squared ErrorThis is the mathematically “ideal” clockBut, only applicable

8、 when post-processing a finite-length waveformBest for showing very-low-frequency effectsAlso shows very-low-frequency effects of scopes timebasePhase Locked Loop (e.g. Golden PLLTracks low-frequency jitter (e.g. clock driftModels “real world” clock recovery circuits very wellExplicit ClockThe clock

9、 is not recovered, but is directly probedExplicit Clock (SubrateThe clock is directly probed, but must be multiplied up by some integral factor7/7/201621 Importance of Clock RecoveryFrom spec, “The jitter measurement device shall comply with the JTF”.How do I verify JTF?JTF is difference between inp

10、ut clock (ref and input clock(unfilteredUse 1100b or 0011b pattern (proper 50% transition densityCheck 1 LF attenuation, 2 -3 dB corner frequency, and 3 slope7/7/201622JTF vs PLL Loop BandwidthConfiguring the correct PLL settings is key to correctmeasurementsMost standards have a reference/defined C

11、R setupFor example, USB 3.0 uses a Type II with JTF of 4.9MhzType I PLLType I PLL has 20dB of roll off per decadeJTF and PLL Loop Bandwidth are EqualType 2 PLLType II PLL has 40dB of roll off per decadeJTF and PLL Loop Bandwidth are not EqualFor example, USB 3.0 uses a Type 2 PLL with a JTF of 4.9Mh

12、z.The corresponding loop bandwidth is 10.126 MhzSetting the Loop Bandwidth as opposed to JTF will lead to incorrect jitter measurement results7/7/201623PLL Loop Bandwidth vs. Jitter Transfer Function (JTFA: Constant Clock Recovery B: PLL Clock Recovery Ratio of B/A7/7/201624 JTF Filtering Effects ba

13、sed on different PLL bandwidths 7/7/201627f 3dB = 30 kHzf 3dB = 300 kHzf 3dB = 3 MHz Jitter for Busy PeopleHints, Tips and Common Errors Using the Jitter Analysis ToolsIssues manifested in different layers of theprotocol stackCrosstalk, jitter, reflections, skewDisparity, encoding or CRC errorsWhere

14、 do I start debugging?Jitter and Eye Diagram ToolsOscilloscope-based for quick resultsFast jitter measurements withOne Button Jitter WizardCompare timing, jitter, eye, amplitude measurementsUser-definable clock recovery, filters, pass/fail limits, and reference levels More Hints for Successful Jitte

15、r AnalysisClock Recovery has a great deal of influence on jitter results. Think about what youre trying to accomplish.Constant-Clock is the most “unbiased”O(jiān)ften best if youre trying to see very-low-frequency effectsBut it can also show wander in the scopes timebasePLL recovery can model what a real

16、data receiver will seeIt can track and remove low-frequency effects, allowing you to “see through” to the jitter that really contributes to eye closureExplicit-Clock is appropriate if your design uses a forwarded clockMake sure your probes are deskewed Hints for looking at Spread-Spectrum ClockIf yo

17、u want to see the SSC effects, use TIE and PLL clock recovery with a bandwidth of at least 1 MHz. A Type-II (2nd -order PLL will track out the SSC more effectively than a Type-I PLL.want to observe the SSC profile:Use a Period measurement and turn on a 3rd -order low-pass filter(in DPOJET with a ban

18、dwidth of 200 kHzBecause Period trends accentuate high frequency noise, the low-frequency SSC trend will be obscured if you dont use a filter You cant use a Frequency measurement directly. The combination of filtering and the reciprocal operation (Freq = 1/Per cause distortion in the resulting waves

19、hape. (This is a mathematical fact, not a DPOJET defect.If you use a TIE measurement, youll see modulation that looks like a sine wave. This is normal. Its because TIE measures phase modulation, which is the integral of frequency. It turns out that the integral of a triangle wave looks very much lik

20、e a sine wave. 誤碼率與噪聲分析 Anatomy of a Serial Data LinkComplete LinkChannelAspirational goal: 0 errorsPractical Goal: Bit Error Rate < Target BERSince BER is the ultimate goal, why not measure it directly? Serial Data Link Integrity = Bit Error RateBit Error Ratio Testers (BERTs are the tools for m

21、easuring BER directlyWhy not use ONLY BERTs for Serial Data Link Analysis?Difficult to model/emulate equalizerMeasurements could take a very long timeInstruments are very expensive and not all that flexibleDoes not analyze the root causes of the impairments of the linksAlternative approach: use a sc

22、ope and advanced analysis toolsEasily move from Compliance to DebugBetter equipped to identify root causes of eye closureEqualizer can easily be modeledMore cost effectiveFaster throughput Why Measure Jitter and Noise?Link Model: Transmitter + Channel + ReceiverTransmitter generates a stream of symb

23、olsReceiver uses a slicer to make a decision on the transmitted symbolThe Bit Decision is made at a certain time (t of the symbol interval and a comparison of the sliced data to a threshold (v is performedJitter impairs the time slicing positionNoise impairs the decision threshold?Jitter combined wi

24、th Noise Analysis is a better predictor of BER performance! A Quick Look at Jitter and Noise DualityJitter analysis evaluates a waveform in the horizontal dimension based on when the waveform crosses ahorizontal reference line.Jitter decomposition is based on spectral analysis of Time Interval Error

25、 vs. timeIndividual jitter components can be separated (i.e. PJ, RJ, DDJ, etc.TJ can then be estimated at a target BER level38Noise evaluates along a vertical dimension on the basis ofcrossings of a vertical reference line at some percentage of the unit interval (usually 50%.Noise decomposition is b

26、ased on spectral analysis of voltage error vs. timeIndividual noise components can be separated (i.e. PN, RN, DDN, etc.TN can then be estimated at a target BER level 抖動(dòng)和噪聲的解析Jitter and Noise Decomposition provide deep insight into BER Full Jitter Analysis vs. Mask Testingstatistical eye closure at a

27、ny other voltage.Conventional mask testing considers both time and voltage , but cannot extrapolate eye closure at low BER. Can we combine the best of both? 41Statistical Jitter + Noise AnalysisBy jointly analyzing Jitter and Noise, behavior at allpoints in the eye can be extrapolated at low BERThe

28、methodology is analogous to current jitter analysis, but is performed across both dimensions of the eyeJitter and noise are separated into components (Random, Periodic, Data-Dependent,The components are reassembled into a model that allows accurate extrapolation. 42Timing-Induced JitterSince jitter

29、is defined as a shift in an edges time relative to its expected position, it iseasy to think of jitter as being causedby horizontal (chronological displacement.Note that the displaced edge (green has not moved vertically in this example. 43Noise-Induced JitterConsider a burst of voltage noise (right

30、 that displaces a waveform vertically.In this case, the displaced edge (green has not moved horizontally.The jitter as measured at the chosen reference voltage is identical in these cases!So, why should we care? Two fundamentally different effects have caused the same amount of jitter, and44Noise-to

31、-Jitter (AM-to-PM Conversion Since waveform transitions are never instantaneous, the slope (slew rate of the edge acts as a gain constant that controls how effectively noise is converted to“observed jitter”. 45Horizontal and Vertical Components of Random JitterWe can think of RJ as being composed of

32、 two components.Horizontally induced: RJ(hVertically induced: RJ(v Since these two components are uncorrelated with each other, they add in theRSS sense:RJ =RJ(h2+RJ(v2Similarly, PJ can be decomposed into PJ(h and PJ(v based on root cause 46Horizontal and Vertical Components of Random NoiseWe measur

33、e noise at a reference point in the bit interval (usually 50%If slew rate isnt zero, jitter (horizontal displacement causes observed noise So as with RJ, RN can be decomposed into components:Horizontally induced: RN(hVertically induced: RN(vSimilarly, PN can be decomposed into PN(h and PN(v based on

34、 root cause Noise to Jitter and Jitter to Noise ConversionConsider: an “ideal” edge in a patternactually has two impairments:Jitter(h (see the blue traceand Noise(note that both of Jitter and Noise result in jitter on edgeThe Combined response (bottomright includes the jittercaused bynoiseNon-impair

35、ed bit edgeWe can separate the noisecontribution of jitter for diagnosticpurposes by breaking RJ intoRJ(v and RJ(hDPOJET and 80SJNB are the only tool that will show you this separation, and thus give youan important troubleshooting hint: e.g. is it crosstalk causing trouble, or the clocks? 48Theory:

36、 Construction of the BER EyeConsider a very simple pattern: 7 bit repeatingOverlay multiple segments of the 7-bit pattern. Each one has noise and jitter, so although the bit pattern is clear, they follow many slightly different paths: Average many pattern repeats together. Everything that is uncorre

37、lated with the pattern averages out. What remains is called the correlated waveform.This waveform fully characterizes DDJ, DCD, DDN, ISI all data dependent effects 49Theory: Construction of the BER Eye Part 2The correlated waveform can be snipped into individual bits and overlaid to form an eye diag

38、ram, using the recovered clock as the alignment reference. This forms the correlated eye: 50Theory: Construction of the BER Eye Part 3Spectral jitter separation is used to find PDFs of the random and periodic jitter.The RJ and PJ PDFs are convolved to find the uncorrelated jitter PDF (red A similar