CML Divider division ratios

[Eng.islammostafakamal]

I'd ask about the CML Divider as i design for 2by3 Cell at the first two stages in the divider chain as
it should operate with high frequencies

First stage should divides from 3.5G --> 2.4G
second stage should divide from 1.75G --> 800M

My problem is that the first stage can't cover the whole range of frequencies. it operates well till 3G only

the same problem with stage two too.

how could i extend the transient frequency range ?

Note: i have tried to increase delay of the last latch to support low frequencies but i have found that 3G is the edge of proper operation.

thanks in advance

 

[saro_k_82]

Are you coming to say that you can operate the divider beyond 3GHz and not below.
The CML flops dont have a problem with low frequencies. Why do you think that they fail at lower frequencies or what happens when they fail?

 

[Eng.islammostafakamal]

At first

thanks for your Help

I will give you an example
i wanna make a divider which operates at 3.5G

The first CML stage operate properly from 3.5 ---> 3.6GHZ

So, the worest case for the second stage is to operate
from (3.5/3=1.167GHz) --> (3.6/2-->1.8GHZ)

My problem is that i couldn't adjust the second stage to cover the whole range (1.167 --> 1.8 GHZ).

It covers only 1.8GHZ -->1.4 GHZ

So,How could i extend the covering range ?

Thanks for advance and sorry for replay latency as i was having the Final exams

 

[dick_freebird]

I am guessing that you have some "ballistic timing" that makes
the /3 work, but it is not robust enough. Presumably you have
logic to flip the active clock edge back and forth, maybe
it is too fast and you are getting by at present by violating
some setup/hold, and land on the wrong side of things
when cycle time gets long enough.

What happens when you run at extremely low clock rate?
Which is the "normal", and which the "abnormal" case for
logic functionality? Where does timing begin to bend the logic?
These questions will probably lead you to the gap.

You should be able to observe in SPICE (et al) how and why
the thing breaks up.