For the trap state, I think you don't need it. Only in the address driven
one cycle, the W and addresses should be concerned. In other cycles, it does
not matter at all what the W and the addresses are.

For the setup/hold times, you need to consider them. That means, for
example, only at the end of the address driven cycle, the addresses are
valid. They are not valid at the beginning of that cycle.

On Wed, Nov 12, 2008 at 3:00 PM, Ben Schnur <b-schnur@northwestern.edu>wrote:

> Hello all,
>
> Does anyone know whether we need to include a trap state in our Mealy FSM?
>
> That is, can we assume there are no invalid input combinations?
>
> I cannot see how it is possible to reduce the number of inputs to 1
> without assuming this.
>
> Also, in our timing diagram, do we need to worry about setup/hold
> times for input values or
> are we ignoring them since we are doing a digital simulation that
> ignores delays?
>
> Any feedback is appreciated.
>
> Thanks,
>
> Ben
>
>
Received on Wed Nov 12 15:51:04 2008