UPGRADE YOUR BROWSER

We have detected your current browser version is not the latest one. Xilinx.com uses the latest web technologies to bring you the best online experience possible. Please upgrade to a Xilinx.com supported browser:Chrome, Firefox, Internet Explorer 11, Safari. Thank you!

cancel
Showing results for 
Search instead for 
Did you mean: 
Adventurer
Adventurer
2,531 Views
Registered: ‎07-24-2016

Cascading MMCMs

Jump to solution

Greetings,

 

Suppose that one has an on-board oscillator driving a 7-series MMCM, with 200Mhz frequency and 18ps (as per the datasheet) period peak-to-peak jitter. The question is, is it worth it to set the MMCM receiving this clock with the 'minimize output jitter' option, and use its (unchanged frequency-wise) output clock to drive another MMCM, in order to have 'cleaner' clocks in the design? Is this particular jitter two small to even consider that? 

 

Does the 'minimize output jitter' option reduce the jitter of the output clock by a specific percentage/ratio with respect to the input clock?

 

And finally, is cascading MMCMs generally a good idea? Are there any drawbacks?

 

Cheers!

0 Kudos
1 Solution

Accepted Solutions
Highlighted
Historian
Historian
4,564 Views
Registered: ‎01-23-2009

Re: Cascading MMCMs

Jump to solution

No.

 

First, the "minimize output jitter" doesn't specifically mean that the MMCM is used as a jitter cleaner. The "loop filter" in the MMCM (which is part of the analog domain) can be tuned. This provides a trade-off between output jitter and the precision of the output phase (in addition to the ability of the MMCM to track changing input frequencies - like you would see in spread spectrum applications).

 

18ps peak-to-peak jitter is already REALLY GOOD. In fact, I would say too good - are you sure this is not RMS jitter (not peak-to-peak jitter)?

 

Furthermore, the jitter reduction/jitter tolerance is the same in an MMCM synthesizing different frequencies as it is in an MMCM doing the same frequency (1x multiplication) - so having an MMCM in front of the "real" one won't do anything - if the input jitter is too much for the "real" MMCM (and it isn't even close to the case with 18ps), then it would also be too much for the MMCM you put in front of the "real" one.

 

Avrum

Tags (2)
2 Replies
Highlighted
Historian
Historian
4,565 Views
Registered: ‎01-23-2009

Re: Cascading MMCMs

Jump to solution

No.

 

First, the "minimize output jitter" doesn't specifically mean that the MMCM is used as a jitter cleaner. The "loop filter" in the MMCM (which is part of the analog domain) can be tuned. This provides a trade-off between output jitter and the precision of the output phase (in addition to the ability of the MMCM to track changing input frequencies - like you would see in spread spectrum applications).

 

18ps peak-to-peak jitter is already REALLY GOOD. In fact, I would say too good - are you sure this is not RMS jitter (not peak-to-peak jitter)?

 

Furthermore, the jitter reduction/jitter tolerance is the same in an MMCM synthesizing different frequencies as it is in an MMCM doing the same frequency (1x multiplication) - so having an MMCM in front of the "real" one won't do anything - if the input jitter is too much for the "real" MMCM (and it isn't even close to the case with 18ps), then it would also be too much for the MMCM you put in front of the "real" one.

 

Avrum

Tags (2)
Adventurer
Adventurer
2,470 Views
Registered: ‎07-24-2016

Re: Cascading MMCMs

Jump to solution

Hi @avrumw

 

Thanks for the reply.

 

8ps peak-to-peak jitter is already REALLY GOOD. In fact, I would say too good - are you sure this is not RMS jitter (not peak-to-peak jitter)?

 

Take a look at page 8: >datasheet<

0 Kudos