我不是半导体/ASIC/FPGA 领域的,对跨时钟域,MTBF等了解的很少。
有时候看到个问题,就会想这个问题,解决这个问题。
或许大家可以先看看 新思科技的 《跨时钟域信号同步的IP解决方案》 一文。
http://www.synopsys.com.cn/information/white-paper/ip
另外还有一篇 《ASIC中的异步时序设计》
http://wenku.baidu.com/view/a2bf0c22192e45361066f59c.html
平时多多少少看些参数,多会讲到 Setup time 和 Hold time,既所谓的建立时间和保持时间 Tsu/Thd。
在同步系统中,如果触发器的Tsu ,Thd不满足,就可能会发生亚稳态,发生亚稳态后触发器输出电平不可预测,可能是中间级电平,可能是01,也可能是震荡的,也无法预测何时触发器从亚稳 态中恢复过来,输出正确的逻辑01电平,这个逻辑电平是0是1是随机的,即输出的并不一定就是对的。在异步电路中完全消除亚稳态是不可能,所以我们还是得 想办法减少亚稳态发生几率。
至于解决方法常见的就是锁存器法,结绳法,异步FIFO法........
很多地方多提到MTBF,即平均无故障时间,也给出了计算公式
fclk 是采样时钟频率
fdata 是数据变化频率
tres 是解决亚稳态所允许占用的时间
T0和T1是与具体触发器相关的常数
看新思科技的资料《ASIC中的异步时序设计》上说T0,T1有可能从库供应商处获得,还列出了他们用的测量电路,也从一个侧面说明T0,T1是个特定的值,或者在某个条件下的常数。
tcko1 是第一级同步触发器的tco,即时钟到输出的延时
tsetup2 是第二级同步触发器的 tsetup ,即第二级触发器的建立时间。
MTBF是平均无故障时间,是越大越好,按可以知道降低fclk采样频率可以增大MTBF时间。
此外百度百科上有几个亚稳态解决方法
1 降低系统时钟
2 用反应更快的FF
3 引入同步机制,防止亚稳态传播
4 改善时钟质量,用边沿变化快速的时钟信号
关键是器件使用比较好的工艺和时钟周期的裕量要大。
应该说是降低概率更合适。
再来看看《ASIC中的异步时序设计》一文中人家的计算例子:
对于一个典型的0.25μm工艺的ASIC库中的一个触发器,我们取如下的参数:
tr = 2.3ns, τ = 0.31ns, T0 = 9.6as, f=100MHZ, a = 10MHZ, MTBF = 2.01 days
as是阿秒是10的负18次方秒。
即触发器每两天便可能出现一次亚稳态。即用一个D触发器同步在上面条件下的亚稳态概率是2天出一次,实际使用环境中可能不是那么好测量出来的。
另外我在一个论坛上看到一个坛友说些数据问题,叫他打了2下触发器问题解决了,这个算是碰到亚稳态问题了。
另外还有一篇关于altera 亚稳态的文献
《Managing Metastability with the Quartus II Software》
http://www.altera.com/literature/hb/qts/qts_qii51018.pdf
下面这个网站介绍亚稳态的定义,计算,还有几个PLA芯片对应情况下的10年期的MTBF时间
http://www.interfacebus.com/Design_MetaStable.html
[Cummings, Clifford E. Synthesis and Scripting Techniques for Designing Multi-Asynchronous Clock Designs-讲了亚稳态,同步,最后还有异步FIFO的设计
另外他还有2篇异步FIFO设计的,多可以在网上找到:
1.Simulation and Synthesis Techniques for Asynchronous FIFO Design
2.Simulation and Synthesis Techniques for Asynchronous FIFO Design with Asynchronous Pointer Comparisons]
[TI的 Metastability Performance of Clocked FIFOs 一文讲ACT ABT FIFO的亚稳态,还有时间常数和不同fc,fd下的ACT ABT工艺下 1级同步 2级同步的MTBF时间。
另外还有一篇TI的 Metastable Response in 5-V Logic Circuit 则给出了几种工艺下的74芯片的时间常数,有这个在算MTBF不是很好算吗:
![](data:image/png;base64,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)
]
下面是2个常用的异步输入同步器:
第一个适用于异步输入脉冲宽度比时钟周期大的时候用。
第二个适用于脉冲宽度比时钟周期小的情况,这个电路的技巧就是把脉冲转换成边沿触发电平。sync_out端收到“1”后,异步这级触发器就清零了。这种适用于检测短脉冲,脉冲间隔时间大于4倍采样时钟周期