zhouzhiping849的个人空间 https://blog.eetop.cn/164433 [收藏] [复制] [分享] [RSS]

空间首页 动态 记录 日志 相册 主题 分享 留言板 个人资料

日志

3 types of CDC problems

已有 1405 次阅读| 2010-7-15 13:59

CDC:cross-clock domain crossing。

(1) Meta-stability

-- why is bad?

    The output of DFF2 is unpredictable.

-- Solution:

    Add an additional flop in the destination domain.

(2) Data hold problem (data loss)

-- why is bad?

     Possible that the signal in the faster clock domain of DFF1 is not held long enough to be captured in the much slower clock domain of DFF2.

-- Solution:

    Inmplement pulse extenders.

(3) Re-convergence (correlation)

-- why is bad?

      Due to branch latencies incorrect data can be processed at the re-convergent point.

-- Solution:

     Implement a gray encoder before the clock crossing and not between the clock crossing. (gray code ensures only one bit on the converging bus changes within one clock cycle)

 (4) Design intent (handshake, FIFO)

Sometimes more sophisticated synchronization schemes are employed.  
--  Handshake sync. Scheme: good for portability, IP, reuse.
-- FIFO sync. scheme: good for high volume data transfer.
 
(5)  Reset Synchronization.
    Asynchronous reset must be de-asserted synchronously.
(Because asynchronous de-assertion of resets may put design into unintended state)

点赞

评论 (0 个评论)

facelist

您需要登录后才可以评论 登录 | 注册

  • 关注TA
  • 加好友
  • 联系TA
  • 0

    周排名
  • 0

    月排名
  • 0

    总排名
  • 0

    关注
  • 1

    粉丝
  • 0

    好友
  • 11

    获赞
  • 19

    评论
  • 779

    访问数
关闭

站长推荐 上一条 /1 下一条

小黑屋| 关于我们| 联系我们| 在线咨询| 隐私声明| EETOP 创芯网
( 京ICP备:10050787号 京公网安备:11010502037710 )

GMT+8, 2024-5-4 03:52 , Processed in 0.013883 second(s), 6 queries , Gzip On, Redis On.

eetop公众号 创芯大讲堂 创芯人才网
返回顶部