Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.
clock buffer has the same rise and fall times.i.e input transition=optput transition.
it is the must and should condition for clock buffer,but normal buffer may not be.
Clock buffers are badly needed in a processor, because the clock spreads all over the processor chip and has to enable every part of the chip.So buffering is needed to boost up the signal and make it to supply the clock signal to every part..
Signal buffers are rarely used for boosting up, because they pass through flops, where they get regenerated. they are used for introducing delays and acquiring inputs from pheripherals..
There are actually two kinds of clock buffers often found in libraries.
One is just regular buffers, that are kind of optimised for better mark/space ratio and less variation from PVT.
The other is actually a dummy kind of clock buffer found in some libraries. This is not a real buffer, but a buffer that will expand to a complete clock-tree during PnR
Generally there are separate entity in the library are there like CLKBUF and CLKINV
but
In 65nm liberary there is not any entity like CLKBUF or CLKINV so could anybody tell me which buffer of inverter should we use for clock tree
Why clock buffer should have same rise and fall time , how will it affect timing ?
According my understanding, same rise and fall time will help in keeping the clock plus symmetric properties. As we know clock is most import part and we don't like change properties. (rise and falling edged rates.)
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.