Continue to Site

Welcome to EDAboard.com

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.

Why Hold time does not depend on clock frequency ?

Status
Not open for further replies.
Because it depends on the time between two events-the rising edge of a clock and the change in state of data. Hold time is just that, TIME, not FREQUENCY. The concept of hold time is: the data must remain constant for HOLD TIME seconds after the clock edge. It doesn't matter WHEN the next clock edge occurs.
 
well , while calculating we consider setup time margin , so if you are calculating frequency then it will be clock to Q + combo delay + setup margin at capture flop .... similar to that we have hold margin .. then why we do not count it in calculating the frequency ?

Rahul J
 

Hi Rahul,
Hold time does not depend on frequency because, this check is done at the same edges. where as the setup check is done after one clock cycle
hold check is done at the same edge to make sure that the data using at that cycle is the one latest, not the old data

Sasi
 

Hold time must be considered if the MINIMUM (clock-to-Q + combo delay) is less than the minimum hold time. This would rarely be the case. So, hold time is often not considered in setting the maximum frequency.
 

incase hold time is more than clock to q +combo delay , then can you calculate frequency using hold margin ?

You may calculate the hold margin , but this will result in hold time violation and must be fixed before going for next stage in backend.

Rahul J
 

If there is a hold violation, it needs to be fixed. Afterwards it needs to be checked whether there is any setup violation or not and if found any setup violation it needs to be fixed.

It is sufficient to calculate frequency using setup time.

Regards
 
  • Like
Reactions: pdude

    pdude

    Points: 2
    Helpful Answer Positive Rating
The ground rule is : 1>Hold violation leads to chip failure. There is no way you can save the chip (without ECOs ofcourse) without closing hold because no change of clock frequency can remove the neg slack, because hold does not depend on time period.
2> When doing a hold fix always check if it has caused setup vio in some other path and vice versa.
 
  • Like
Reactions: pdude

    pdude

    Points: 2
    Helpful Answer Positive Rating
There seems to be confusion here. People seem to be getting this backwards

Setup time, like hold time, DOES NOT depend on frequency. Setup time is nothing more than the minimum time requirement that data must be stable before the clock edge. That's it. It has NOTHING to do with the previous clock edge.

However, from a system point of view both setup and hold time, along with propagation delays, will figure into maximum frequency limits.

The original question was why doesn't hold time depend on frequency; it was not a question about overall system timing.

Said another way, frequency depends on setup and hold time; setup and hold time do not depend on frequency.
 
  • Like
Reactions: pdude

    pdude

    Points: 2
    Helpful Answer Positive Rating
HI Barry

agree there is confusion in my question , i want to ask that while calculating max frequency , we dont consider hold time requirement.
 

HI Barry

agree there is confusion in my question , i want to ask that while calculating max frequency , we dont consider hold time requirement.

Sorry, but that's exactly NOT what you asked. You asked why hold time doesn't depend on frequency. As I explained in post #9, frequency depends on hold time, not the other way around.

I breathe when I sleep; that doesn't mean I sleep when I breathe.
 
Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top