Tskew and timings between flip-flops

Thread Starter

Avyhe2956

Joined Sep 14, 2016
2
I do not understand how to approach questions of this kind.

INTRODUCTION: I know the following formulas:

The conditions that must exist between FF1 and FF2: (including Tskew)

(A) T_Clk >= t_pd(FF1)+t_pd(LOGIC)+t_setup(FF2)-T_skew
(B) t_cd(FF1)+t_cd(LOGIC) >= t_hold(FF2)+t_skew

But, I still do not understand, for example, how to deal with the exercises below?

For (1): what is the minimal Tclk? I prefer a full explanation than a final answer. In addition, How can I see the condition for Thold saved?

For (2): what os the minimal Thold?

I sit on it for a long time and can not understand.

NOTE: I know that NOT inverts the specimen - but I can not seem to get it reflected in the formula and to understand it.


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
NOTE: I never saw a formal way for this topic and I really want to understand it.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Solution: (And i waiting for inspection / repairs / comments / answers)

FOR (1):
0.5*T_clk >= t_pd(FF)+t_pd(A/B)+t_setup(FF)
--> 0.5*T_clk >= 20+10+5
-->T_clk>=70ns

HOLD:
I do not completely understand how it should be. Like a or like b?
(a). t_cd(FF)+t_cd(A/B) >= t_hold(FF) + 0.5*Tclk ?
(b). t_cd(FF)+t_cd(A/B) >= t_hold(FF) ?

and why?

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

FOR (2):

HOLD between FF1-->FF2:
I do not completely understand how it should be. Like a or like b?
(a). t_cd(FF)+t_cd(or)+t_cd(AND) >= t_hold(FF) + 0.25*Tclk + t_cd(NOT) ?
(b). t_cd(FF)+t_cd(or)+t_cd(AND) >= t_hold(FF) + t_cd(NOT) ?

and why?
 
Top