Forums

Essential hazards in CPLD's?

Started by Preben Mikael Bohn October 31, 2003
Hi NG, are essential hazards avoided in CPLD-designs?

If for example I were to design a counter with a subsequent comparator 
in a CPLD, would I be sure that I got no glitches in the output from my 
comparator? Or would I have to filter the output before I used it in 
other part of the CPLD?

Best regards Preben

Preben Mikael Bohn wrote:
> Hi NG, are essential hazards avoided in CPLD-designs?
They can be.
> If for example I were to design a counter with a subsequent comparator > in a CPLD, would I be sure that I got no glitches in the output from my > comparator?
If you use synchronous design and do static timing analysis, your outputs won't glitch. -- Mike Treseler
If you implement a binary counter and compare its output against another
binary value, there is no way on this earth to avoid combinatorial
glitches, no matter what technology or circuit design you use.  The
classical solution to this problem is to Gray-code the counter, so that
only one bit changes at a time.
Of course, you can also re-synchronize the comparator output and thus
suppress all combinatorial glitches...

Peter Alfke
========================================
Preben Mikael Bohn wrote:
> > Hi NG, are essential hazards avoided in CPLD-designs? > > If for example I were to design a counter with a subsequent comparator > in a CPLD, would I be sure that I got no glitches in the output from my > comparator? Or would I have to filter the output before I used it in > other part of the CPLD? > > Best regards Preben
Peter Alfke wrote:
> If you implement a binary counter and compare its output against another > binary value, there is no way on this earth to avoid combinatorial > glitches, no matter what technology or circuit design you use.
I thought so... :-/ > The
> classical solution to this problem is to Gray-code the counter, so that > only one bit changes at a time.
Ah... Of course... Only problem is that I guess this requires a higher macro-cell count and my current design is already at the limit (and there are many counters in it :-)).
> Of course, you can also re-synchronize the comparator output and thus > suppress all combinatorial glitches...
I guess this is the best/"cheapest" way to do it. Thank you for the info. Best regards Preben