主板信號量測.ppt_第1頁
主板信號量測.ppt_第2頁
主板信號量測.ppt_第3頁
主板信號量測.ppt_第4頁
主板信號量測.ppt_第5頁
已閱讀5頁,還剩35頁未讀 繼續(xù)免費閱讀

下載本文檔

版權說明:本文檔由用戶提供并上傳,收益歸屬內(nèi)容提供方,若內(nèi)容存在侵權,請進行舉報或認領

文檔簡介

1、Timing Analysis & Signal Simulation on PC Systems,By Pierre Lee at ARIMA 4/16/2002,Hold,Common Clock Data Transfer,D0 D1 D2,D0 D1 D2,Clock Driver,Driving,Receiving,Tco,Flight Time,Setup,1,2,3,4,Defining Tco,Din,Clock,OutputBuffer,InternalLogic,RL = 50 W,Clock rises t = 0,Vmeas,Tco,Load for Tco measu

2、rement (from databook),Tco = time from clock rise to Vmeas into test load,Components of Tco,Din,Clock,OutputBuffer,InternalLogic,RL = 50 W,Clock rises t = 0,Vmeas,Tco,Internal delay = from clock rise to the point where the output begins to switch,External (buffer) delay = howlong the buffer takes to

3、 drive thereference load to Vmeas,For Signal Integrity Purposes .,Its all in the Math .,Driving,Receiving,Tco,Flight Time,1,2,3,4,?,t = 0,Clock Jitter,Clock Driver,Clock Jitter occurs when the clock period varies from one period to the next Usually caused by PLL instability in the clock driver Jitte

4、r increases / decreases the clock period, decreasing the effective clock cycle,Clock Skew,D0 D1 D2,Clock Driver,D0 D1 D2,t = 0,t = 1,t = 2,Occurs when different devices see the clock transition at different times Increases / decreases the apparent clock cycle. Depending on which devices are driving

5、/ receiving Reduces the effective clock cycle,Bus Clock Cycle Budgeting,For each Driver Receiver path: Tflightmax Receiver(Hold) - Driver(Tcomin) + Skew + Crosstalk,Common Clock Bus Example,Intel Pentium-Pro reference design Processor/Chipset Bus (GTL+, 66 MHz) Intel GTL+ Design Guidelines,Defining

6、Device Timing,Timings taken from “AC (dynamic) Specifications” sections of Intel datasheets Most datasheets available via WWW Important parameters Clock Data Valid for GTL+ Bus Setup / Hold requirements for GTL+ signals PLL Jitter (if specd),Pentium Pro,440FX (timings from 440LX),Determining Flight

7、Times,Tflightmax = 4.50 ns Tflightmin = 0.45 ns,Signal Wave Propagation,Flight time t1 = L/c,c 6.5 in/ns Tflightmax = 4.50 ns Tflightmin = 0.45 ns 29in L 3in,Risk,Ohms law ?,i = (va-vb) / Rtrace,i when Rtrace 0,Model of transmission line,Signal Wave Propagation,Flight time t1 = L/c,Impedance Change,

8、Layer change Reference plane crossing Trace splitting Series component Connector Pull high & pull down,Impedance Change,Layer change,Reference plane crossing,Trace splitting,Series component,Connector,Pull high,Schematics vs. Layout,Microstrip Section,Zo = f( w, h, t, er),Zo = Characteristic impedan

9、ce (W) w = Width of trace (mils) t = Thickness of trace (mils) h = Thickness of dielectric layer (mils) er = Dielectric constant of the dielectric layer,Impedance Verification by TDR,Impedance Verification by TDR,Impedance Measured by HP 54754A,Flip-Chip Package,IBIS ModelI/O Buffer Information Spec

10、.,Simulation by 3com,Low Voltage CMOS,Un-terminated,Terminated,Low Voltage CMOS,Un-terminated,Terminated,Clock gen. Waveform,IDE Waveform,Crosstalk,When traces are close together, a change in current flow in one trace will cause current to flow in an adjacent trace Aggressor traces induce currents i

11、n adjacent victim traces Changes in current flow only occur during the aggressors rising and falling edges,Crosstalk,No Crosstalk,Impacts of Crosstalk,Crosstalk causes switching noise to appear on victim traces that would otherwise be quiet Crosstalk degrades the system noise budget when it occurs b

12、etween signals that are not part of the same signal group Since crosstalk only occurs when drivers are switching, the magnitude of crosstalk between bus bits is usually not important,Aggressor,Victim,Crosstalk-induced noise,Electrical Model for Crosstalk,Crosstalk - Impact on Bus Timing,Crosstalk be

13、tween adjacent bus bits affects edge speed (and therefore flight time) Denser routing makes better use of board space, but at the expense of larger variations in flight time Pre-layout crosstalk analysis helps the designer make the best tradeoff between routing density and signal integrity,Even Mode

14、,Reference,Odd Mode,D0 D1 D2,D0 D1 D2,D0 D1 D2,Example of Crosstalk,Risk in mass production,PCBs statistical quality Driving strength variation in chipset Passive components quality Temperature variation,Monte Carlo Analysis,A risk analysis instead of analytical method Used in complicate systems or too many parameters Suitable for mass production models,Example of Monte Carlo,Monte Carlo for Clock Signal,Conclusion,Impedance matching is important

溫馨提示

  • 1. 本站所有資源如無特殊說明,都需要本地電腦安裝OFFICE2007和PDF閱讀器。圖紙軟件為CAD,CAXA,PROE,UG,SolidWorks等.壓縮文件請下載最新的WinRAR軟件解壓。
  • 2. 本站的文檔不包含任何第三方提供的附件圖紙等,如果需要附件,請聯(lián)系上傳者。文件的所有權益歸上傳用戶所有。
  • 3. 本站RAR壓縮包中若帶圖紙,網(wǎng)頁內(nèi)容里面會有圖紙預覽,若沒有圖紙預覽就沒有圖紙。
  • 4. 未經(jīng)權益所有人同意不得將文件中的內(nèi)容挪作商業(yè)或盈利用途。
  • 5. 人人文庫網(wǎng)僅提供信息存儲空間,僅對用戶上傳內(nèi)容的表現(xiàn)方式做保護處理,對用戶上傳分享的文檔內(nèi)容本身不做任何修改或編輯,并不能對任何下載內(nèi)容負責。
  • 6. 下載文件中如有侵權或不適當內(nèi)容,請與我們聯(lián)系,我們立即糾正。
  • 7. 本站不保證下載資源的準確性、安全性和完整性, 同時也不承擔用戶因使用這些下載資源對自己和他人造成任何形式的傷害或損失。

評論

0/150

提交評論