John Fields wrote:
On Sun, 04 Mar 2012 07:27:18 -0600, Lasse Langwadt Christensen
wrote:
John Fields wrote:
On Sat, 03 Mar 2012 20:18:46 -0800, John Larkin
wrote:
On Sat, 03 Mar 2012 16:24:01 -0600, John Fields
wrote:
Yikes, the PEs are completely asynchronous to the clock.
---
Yup, that's how they're designed to be.
sorta...
---
There's all sorts of available pathologies.
---
None that I can see; peruse:
http://www.ti.com/lit/ds/symlink/cd4516b.pdf
for a clue.
--
JF
It might not matter, but when you have PE async to the clock
you will at times violate one of the specs on the first page
preset removal time
-Lasse
---
From figure 3, since the external clock is gated with PE and the
internal clocks forced high while PE is asserted, the state of the
external clock during load time doesn't matter.
You're right in that if PE is de-asserted and the external clock goes
high too soon after PE goes low, the Preset Enable Removal Time spec
will have been violated, but I think (since all the presets have
already been loaded and are stable) all that'll happen is that
there'll be a delay until the next clock high comes along before the
counter starts counting.
I wouldn't count on it, it is the exact same reason some stay away from async
resets. If you happen to deassert it right before a clock edge some flops might
see the clock edge others might not
looking at the internals I think the sames goes for the carry in, hard to tell
what
would happen if the carry is asserted/deasserted and rippling through that next
state logic right when there's a clock edge
-Lasse