大家一起讨论抖动

zbb8637

太专业了！不太理解。先收藏一下吧！

asus0929

謝謝分享...學習了

stupid

Q: Jitter Spectrum on DCA-J versus Jitter Spectrum with EZJIT+

A: When measuring Jitter Spectrum using the 83496B with the 86100CU-400 PLL and Jitter Spectrum software the measurement is made by first demodulating the jitter signal from the incoming clock or data signal. A sampler inside the clock recovery circuit then samples the jitter signal typically at 52MSa/s and an fft transforms to the frequency domain.

EZJit Plus uses a very different technique. The incoming clock or data signal is sampled at full speed (typically 40GSa/s). Software clock recovery then allows the jitter to be extracted in the form of a Time Interval Error signal. An fft can then be performed on the TIE to get the Jitter Spectrum.

* E8 {, c5 a6 W& DThere are many differences in the acquisition hardware and measurement techniques that will affect the measured outcome.
( o' l# Z/ x5 b0 J- ?
6 E2 O" {4 l! M) D6 {  r% mA couple of important points:
: ~7 c3 \0 ?: `  p# G3 q
2 t* ^0 n6 j3 K# K: h1. EZJit Plus is used on a wideband oscilloscope that captures the entire frequency content of the signal. This high bandwidth also means that there is more noise that transfers through to the jitter spectrum than with the DCA-J method. Thus the DCA-J method typically has a lower noise floor than the EZJit Plus method.

5 Z5 b# @) K. r6 s
2. Since EZJit Plus captures the entire signal, the Jitter Spectrum can be viewed up to Data Rate/2 whereas with the DCA-J the Jitter Spectrum is limited to 26MHz offset.

虽然DCA-J做出来的抖动频谱只能到26MHz，但是对于大部分应用还是足够啦，因为PLL的带宽通常都不会超过10MHz，而且

stupid

Q: What is the differenc between SJ and PJ?

. _/ _+ k- S! F$ y4 J! yA: The SJ generation uses an IQ modulator, PJ the voltage controlled delay line. Use SJ for low frequency jitter (up to a few MHz) with high amplitudes (max is a function of the frequency, up to a thousand ui at low freqs). Use PJ for high frequency jitter (up to 300MHz) with low amplitudes (max 200ps). PJ offers more waveforms (sinusoidal, rectangular and triangular), SJ is always sinusoidal.
1 n3 Q$ F3 D/ }8 B& C) }4 b
# H1 L- s* [) I9 rSinusoidal jitter is something you'll rarely see out in the wild - the usual periodic jitter (for example from power supply noise crosstalk and other interference) has a much more complicated waveform.

If you need to test the jitter tolerance behavior of a Phase Locked Loop (PLL) or Clock/Data Recovery (CDR) however, you want to use a simple test signal with a well defined jitter frequency content. Sinusoidal jitter is perfect for that, and as a bonus it's fairly easy to generate. The use of "artificial" SJ for these t&m application goes back to the SONET standard.

( H6 Y! p' \& D, p' cIn real world sinusoidal jitter may be caused by cross talk from a clock signal.

jeddite

有本书名字就叫jitter分析和测量吧，楼主把它也贴上去吧

stupid

如果你说的是Mike Peng Li的那本书的话，网上有好多pdf版，我比较老土，一直看的是印刷版。看这本书有两个地方特别值得注意
, j7 F; Q" t" T* N$ U1. 假定电路系统是LTS，也就是线性的
+ S+ s+ i- l5 \9 N$ {! N4 L7 k5 T2. BUJ依然是一个伤感情的话题
7 R# Y- g' I% u& \1 s当然这本书对jitter的总结非常到位。

ggm_1029

非常的不错，分析的非常的全面！

hzz137

好料！！！！！！1

kom_77

十分的好，print出来后仔细学习.谢谢高手！

zgl846

总结的不错哦，顶一个 bjbdf466zgl

zgl846

不错不错，楼主辛苦了，顶一个 bjbdf466zgl

zgl846

不错不错，加油哦！   bjbdf466zgl

jasonchenp

Mark

Gionee

leojl_liu

太复杂