It is applicable with any input signal source. However, applicable just means that the property still holds, not that there is necessarily a risk of metastability. That depends on whether you can guarantee that you meet the set-up and hold times. A Mk 1 finger has no such guarantee BUT it is also very unlikely that your finger press is going to just happen within a window 10 picoseconds wide which occurs once per clock edge with a 10MHz clock. Even if it does, the result will just look like additional switch bounce at worst. However, if you have a very high-frequency repetitive signal which presents an opportunity at each transition to violate the set-up time and do so in almost exactly the same timing relationship each time, and it is not tolerant of perceived excess edges, it makes the probability of failure simply low, not astronomically low. Hence the failures once per month in 10s of units running continuously. Sean On Fri, Jul 1, 2011 at 3:25 PM, Michael Watterson wrote= : > On 01/07/2011 20:03, Sean Breheny wrote: >> Given that metastability is a fundamental property of anything which >> makes discrete yea/nay decisions (although the probability may be tiny >> of a failure), I don't know how you can possibly say that it just >> "does not apply" to these various buses. >> >> Sean > If it's applicable, then it's applicable with ONE CPU and a Mk1 finger > pressing a button. > > If it's applicable, then it's applicable with arbitrary numbers of CPU > in any clock relationship. > > I can think of more believable bad reasons to use one clock. > -- > http://www.piclist.com PIC/SX FAQ & list archive > View/change your membership options at > http://mailman.mit.edu/mailman/listinfo/piclist > --=20 http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .