眼图基本原理上课讲义.ppt
《眼图基本原理上课讲义.ppt》由会员分享,可在线阅读,更多相关《眼图基本原理上课讲义.ppt(70页珍藏版)》请在淘文阁 - 分享文档赚钱的网站上搜索。
1、眼图基本原理高速串行总线和传统并行总线对比2022/11/162V1.0 Confidential高速串行总线新兴标准和数据速率FibreChannel 4X 4.25 Gb/s 4.25Gb/s(1X)FBDIMM 4.8 Gb/s 500 Gb/s(14 lanes x 8 channels)2022/11/163V1.0 Confidential眼图定义2022/11/164V1.0 Confidential一般信号与眼图的区别2022/11/165V1.0 Confidential眼图特征定义 2022/11/166V1.0 Confidential眼图特征定义2022/11/167V
2、1.0 Confidential眼图形成示意2022/11/168V1.0 Confidential眼图在示波器上的形成2022/11/169V1.0 Confidential眼图的CLK采样示波器的CLK通常可能是用户提供的时钟,恢复时钟,或者与数据信号本身同步的码同步信号.实时示波器通过一次触发完成所有数据的采样,不需附加的同步信号和触发信号.通常通过软件PLL方法恢复时钟.Tektronix RT-Eye软件即使用此方法.TDS8200 ET Scope(20GHz 80E03)TDS6154C RT w.DSP2022/11/1610V1.0 Confidential眼图与CLK时钟速
3、率提高,眼图质量下降125M CLK250M CLK500M CLK2022/11/1611V1.0 Confidential示波器上的眼示波器上的眼图图提供串行总线数据的丰富信息:通过比较发送端和接收端的眼图质量可以分析出信号传输的问题 眼高,眼底,眼宽 上升时间 下降时间 模板测试2022/11/1612V1.0 Confidential眼图与系统速率水平时域上关联因素 Transition(rise)time Aperture timeTiming uncertainty,skew and jitter垂直域 幅度损耗和噪声Unimpared bit edge2022/11/1613V1
4、.0 Confidential眼图和传输质量Larger eye opening means greater tolerance to amplitude and timing variationsLarger eye opening translates to simpler system design and better reliabilityWider transition region indicates greater jitter and higher bit error rateBest error rate obtained when decision point is ne
5、ar center of the eye2022/11/1614V1.0 ConfidentialP ValuesPTopmean,the mean value of PTopPTopsigma,the standard dev of PTopPBasemean,mean value within aperturePBasesigma,std dev of PBase in aperturePcrossmean,vertical mean of crossing pt.T ValuesTcrossmean,horiz mean of crossing pt.Tcrosssigma,horiz
6、std dev of crossing pt.Tcrosspk-pk,horiz pk-pk of crossing pt.The Eye Aperture defaults to the center 20%of the interval from TCross1 to TCross2 but is settable from 0 to 100%眼图测量特征量TCross1TCross2PCross1PCross2Eye AperturePBasePTop2022/11/1615V1.0 Confidential4 Jitter RMS=TCross1sigma4 Jitter Pk-Pk=
7、TCross1pk-pk4 Extinction Ratio=PTopmean/PBasemean (ER also in dB and%)4 Quality Factor=(PTopmean-PBasemean)/(PTopsigma+PBasesigma)4 Eye Height=(PTopmean-3*PTopsigma)-(PBasemean+3*PBasesigma)4 Eye Width=(TCross2mean-3*TCross2sigma)-(TCross1mean+3*TCross1sigma)4 Crossing Percent,Duty Cycle Distortion,
8、Noisepk-pk,NoiseRMS,SNR眼图测量特征量眼图测量特征量TCross1TCross2PCross1PCross2Eye AperturePBasePTop2022/11/1616V1.0 Confidential眼图的测量近端近端远端远端采样点 串扰 随机抖动 确定性抖动码间干扰(ISI)占空比失真(DCD)正弦抖动(SJ)无关联的 幅度损耗2022/11/1617V1.0 Confidential串行串行总线总线:BasicblocksasicblocksTransmittersInterconnects(paths)ReceiversTx+-path+-+-path+-+
9、-path+-+-+-RcvDifferential TransmitterDual-coax differential cableTwisted-pair differential cableCoupled-pair traces on circuit board(backplanes)Differential Receiver2022/11/1618V1.0 Confidential高速串行总线信号线上不仅包含数据信息还包含时钟信息pathTx+-+-+-+-RcvDifferential serial data is sent without any clock signal acros
10、s the interconnect to the receiverCDRDATADATACLOCKA clock is“recovered”from the incoming data at the receiver through a clock and data recovery circuit(CDR).2022/11/1619V1.0 Confidential高速串行总线数据传输过程中干扰因素增多path+-+-path+-+-Tx+-path+-+-+-RcvTx+-path+-+-+-Rcvpath+-+-Tx+-path+-+-+-Rcvpath+-+-Transition p
11、oints often involve combinations of solder joints,circuit board vias,and connectors:these all can have substantial effect on the total link performance.2022/11/1620V1.0 Confidential高速串行总线损耗速率越高互连距离越长,损耗越大Reference Maxim Note HFDN-27.0(Rev.0,09/03)Tx+-+-Rcvpath+-+-Clean,open,logical 1&0 at launch fro
12、m transmitterSmeared edges at end of long interconnect.Logical 1&0 can be hard to distinguish at end of long interconnects;(this is often called a“closed eye”)Fast,sharp,edges at transmitter launch2022/11/1621V1.0 Confidential高速串行总线串扰 path+-+-Rcv+-+-Rcv+-+-Rcv+-+-RcvSerial data can be a single diffe
13、rential signalTx+-Tx+-Tx+-Tx+-Tx+-+-+-+-Rcvbut generally there are multiple“lanes”of serial data running side by side;these can CROSSTALK with each other.2022/11/1622V1.0 Confidential高速串行总线精度 设计者要更加精确的掌控设计Tx+-path+-+-6.25Gb/s at Tx launch into backplane6.25Gb/s at 17in(43cm)of backplane6.25Gb/s afte
14、r 34in(86cm)of backplaneSmall differences in levels being measured2022/11/1623V1.0 Confidential如何获得张开的眼图走线长度 短走线并非始终能够满足.短走线意味低损耗.走线宽度 宽走线可以降低趋肤效应.减小板材的介电常数 即降低介电损耗(Dielectric Loss),但将增加成本.信号预加重和均衡处理 通过对跳变位预加重(Pre-Emphasis)处理补偿线路上因信号跳变产生的针对高频分量的损耗.其它接收端均衡处理等.2022/11/1624V1.0 Confidential眼图与BER8究竟这个眼
15、睛有多开究竟这个眼睛有多开?(这却与你观察多久有关这却与你观察多久有关)?2022/11/1625V1.0 Confidential眼图与BERpdf 1pdf 2cdf 20.5UI12ABC0UI1UIEye open for BER 1.0E-12-0.5UI1.5UIcdf 1110眼睛眼睛睁开度睁开度 vs 误误码率码率2022/11/1626V1.0 Confidential眼图与BER利用眼图分析抖动得到浴盆曲线(Bath Tub),从而进行误码率的估算。2022/11/1627V1.0 Confidential泰克提供的眼图测量方案从数据恢复时钟(CDR),眼图模板测试硬件 C
16、DR(PLL)软件 CDR(PLL+其它)测量绘图抖动:趋势,频谱,直方图,浴盆曲线2022/11/1628V1.0 Confidential硬件时钟恢复TDS6000B 系列系列,TDS7704B 和和 CSA7404B 显示器显示器锗化硅触发锗化硅触发 系统系统放大器放大器/衰减器衰减器数据数据采集采集系统系统Golden PLLFbd/166720 dB/dec来自被来自被 选中的选中的通道通道恢复恢复时钟时钟Ch2 Ch3Ch4从1Mbit/s到 3.125Gbit/s的所有电接口和光接口速率独有的 锗化硅触发锗化硅触发系统系统2022/11/1629V1.0 Confidential
17、硬件时钟恢复产生的眼图TDS6804B/P7380SMADSP增强功能XAUI 超过 3.125Gb/s眼图测量和统计2022/11/1630V1.0 Confidential新型新型测试方法方法RT-Eye方法生成眼图新兴标准要求新型工具TDS RT-Eye(选项RTE)基于泰克在软件领域的专家经验,创建了一个交钥匙式的一致性测试和分析解决方案2022/11/1631V1.0 Confidential新型新型测试方法方法RT-Eye方法生成眼图软件时钟恢复眼图全系列标准专用参数测量,包括幅度、定时和抖动多种格线图,进行补充数据分析标准专用模块作为插件提供用户可定制的波形模板、测量极限和报告功
18、能,支持专有标准和新兴标准生成报告2022/11/1632V1.0 Confidential新型新型测试方法方法RT-Eye方法生成眼图低抖动本底噪声在TDS6604 或 TDS7704B上为700 fs RMS单触发事件,而不是ET方法中的多触发事件支持不同的时钟恢复模型锁相环(PLL)相位内插重复取样(恒定时钟,连续位)数据相关分析把跳变位与非跳变位分开码型长度检测,进行抖动分析(Rj/Dj分离)恢复高达 10Gb/s 的时钟高达 5Gb/s 的眼图高达 10Gb/s 的抖动分析0.1 UI的抖动本底噪声2022/11/1633V1.0 ConfidentialRT-Eye眼图示例TDS6
19、804B和下一代标准2.5Gb/s5Gb/s6.25Gb/s2022/11/1634V1.0 Confidential新型新型测试方法方法RT-Eye方法生成眼图分析结果2022/11/1635V1.0 Confidential新型新型测试方法方法RT-Eye方法生成眼图抖动一致性测量T TX-EYE-MEDIAN-to-MAX-JITTERTTX-EYE-MEDIAN-MAX-JITTER2022/11/1636V1.0 Confidential新型新型测试方法方法RT-Eye方法生成眼图抖动一致性测量抖动眼图张开+总抖动=1UI2022/11/1637V1.0 Confidential采样
20、示波器眼图生成When a clock signal is used to trigger the equivalent-time 8200 scope the sampled DATA signals generally create EYE PATTERNS(between clock triggers the sampled DATA could be either a logical 1 or 0)DatavoltagetimePrecision variable delayA clock trigger can be user-supplied or recovered from t
21、he data to trigger the equivalent time samplerClockEye patterns are the common result of clock-triggering in Equivalent time sampling:vectors are not drawn since adjacent samples can jump from logical 1 to 0 frequentlyMask test2022/11/1638V1.0 ConfidentialCSA8200 with 20GHz 80E03采样示波器比较DRT示波器有更好的信号保
22、真度A fast 15ps rising edge stimulus into an ET system and RT system;Averaging On.aberrations are listed,step.DSP ONtR 10-90=22psaberrPkPk=0%,6.1%tR 10-90=30.3psaberrPkPk=9.5%,7.2%TDS6154C,DSP On.Averaging on2022/11/1639V1.0 ConfidentialCSA8200 with 20GHz 80E03采样示波器比较DRT示波器有更好的信号保真度 A fast 15ps rising
23、 edge stimulus into an ET system and RT system.aberrations are listed,step.TDS6154C,DSP On.DSP ONtR 10-90=21psaberrPkPk=0%,5.5%tR 10-90=22.5psaberrPkPk=10.4%,11.3%Averaging off2022/11/1640V1.0 ConfidentialTDS8200 ET Scope(20GHz 80E03)采样示波器与DRT示波器眼图对比TDS6154C RT w.DSPAn Advantest 12Gb BERT(D3186)was
24、used to differentially drive two channels and also supplied a reference clock.LEFT:CHA-CHB on an TDS8200 with 80E03 20GHz sampler and an 82A04 Phase Reference moduleRIGHT:CH2-CH3 on a TDS6154A(TDS6154C)realtime scope;eye generated using RT-Eye.2022/11/1641V1.0 ConfidentialTDS8200 ET 80E04采样示波器与DRT示波
25、器眼图对比TDS6154C RTO,DSP OnSlow Eye Diagram(4GFC,cca.80ps edges,600mV full swing)(TDS6154 comparison is a composition)Mask Margin=80%Mask Margin=47%Mask_MarginCSA is about 1.7x Mask_MarginTDS2022/11/1642V1.0 Confidential应用对比选择TDS6154CDebugTriggeringProbingSignal AcquisitionSingle shot capableNon-repeti
- 配套讲稿:
如PPT文件的首页显示word图标,表示该PPT已包含配套word讲稿。双击word图标可打开word文档。
- 特殊限制:
部分文档作品中含有的国旗、国徽等图片,仅作为作品整体效果示例展示,禁止商用。设计者仅对作品中独创性部分享有著作权。
- 关 键 词:
- 基本原理 上课 讲义
限制150内