锁相环相位噪声和抖动的关系

hughhuang

最近在研究数字锁相环的噪声，因为应用背景是serdes，所以不像RF比较关心相躁一样，而是比较关心jitter。那么问题就来了，jitter分位RJ和DJ，而DJ又包括BUJ和DDJ，我怎么能将理论计算出来的相躁曲线、实际测得的相躁曲线和这些jitter种类分别对应上呢？比如按噪声源分类，包括PFD的噪声、divider的噪声、DCO的相位噪声和量化噪声等，这些噪声经过系统之后对应到最终输出的相位噪声曲线这个我会计算，但是怎么对应到最终的输出jitter呢？毕竟RJ是无穷大的，热噪声我积分应该积到多少频率呢？还有DCO的量化噪声应该属于DJ没错，但是应该属于BUJ还是DDJ呢？不明白的地方太多了，请各位前辈指教，或者推荐一些讲解详尽的资料也好。万谢！

hughhuang

还有一个问题，如果我对DCO的量化噪声进行积分得到量化噪声的rms jitter，那我如何转化为p2p jitter呢？毕竟DJ不是正态分布的，肯定不能用p2p jitter=14*rms jitter啊。请前辈们指点。万谢！

hughhuang

哪位前辈有试过将相躁转化为Jitter么？比如将DCO量化噪声转化为jitter？我看一些资料上提供的方法就是积分，然后得到rms jitter，但是对于DJ而言，rms jitter怎么转化为p2p jitter呢？毕竟对于RJ可以取正负7个sigma，然后p2p=14*rms，但DJ又不服从正态分布。我们的应用背景是serdes，所以比较关心jitter，而不是PN。很纠结，特此请教各位前辈。万谢！

CuiBing

好可怜，没有回答。

龙爱小说

哎，有同样的疑惑，不知楼主现在是否已经知晓答案了，能不能分享一下心得呀？

Thomas_Bai

楼主现在已是大神了

gongjiang

You can calculate the RJ and DJ from the phase noise plot (sum of PFD, divider and DCO). RJ means random jitter which is due to noise. The rms RJ integration bandwidth is related to your applications, which f_LOW typically is the packet duration (say 200us in BLE) and f_HIGH is the modulation bandwidth. Even you integrated it from 1Hz (1s, that's a lot for most applications) to infinite frequency, the RMS integrated jitter is still bounded (not infinite). However, when flicker noise is presented in PFD or DCO and type I loop is used, the RJ is not bounded any more as you increase the measurement time. The DJ is typically related to the spurious tones of your PLL_Out. You can convert spur to DJ by the equation DJ=2/pi*T_out*10^(spur/20).  DCO quantization error will also introduce phase noise at the DCO output (just like a DAC), which is RJ instead of DJ consider its input is random. On the other hand, the DCO quantization error can also introduce spurious tone, which in this case is DJ. I am sorry that I never deal with BUJ and DDJ. You can refer to Bogdan's famous textbook for further details.

Jason_Zhangs

相位噪声曲线通过积分得到的RMS值似乎只是一个sigma的值，量化噪声实在不知道该怎么转换成jitter
求助！

实在很忙

相位噪声和jitter是可以换算的，一个是频率特征另一个是时域特征，本质都是一样的

糯米zz

同求同求