The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> RF Design >> CML divider self oscilation frequency
https://designers-guide.org/forum/YaBB.pl?num=1421750016

Message started by raja.cedt on Jan 20^th, 2015, 2:33am

Title: CML divider self oscilation frequency
Post by raja.cedt on Jan 20^th, 2015, 2:33am

Dear all,
Can any one explain how a cml divider will have self oscillation frequency, since it is a two stage circuit each stage should contribute 90 deg at oscillation frequency. But each stage can provide 90 deg phase shift at very high frequency where almost gain drops to very low(I am assuming tail node is not contributing any phase). I think I missing some thing about the cross-coupled pair, I am modelling them as -ve resistance only.

Can any one explain where i am missing/point me to useful doc

Thanks in advance,
Raj.

Title: Re: CML divider self oscilation frequency
Post by loose-electron on Feb 6^th, 2015, 10:58am

Suggestion here- please post a set of block diagrams and schematics of what you have. The question as it is right now requires too much guessing on what is going on.

jerry

Title: Re: CML divider self oscilation frequency
Post by raja.cedt on Feb 6^th, 2015, 3:38pm

Sorry..
Question is how can two stage cml circuit oscillates here?

Title: Re: CML divider self oscilation frequency
Post by loose-electron on Feb 9^th, 2015, 10:41pm

the device is a cascade of

amplifier - latch - amplifier - latch - back to input

As the device operates in a linear manner, what is the gain and phase relationship of each stage ? At te functional level that will tell you about the oscillatory behavior of the device.

Title: Re: CML divider self oscilation frequency
Post by Jeffrey987 on Mar 14^th, 2015, 2:33pm

the latch has a resistance of -2/gm (like in a LC oscillator). Write down the impedance of the load (Cl//Rl//(-2/gm) and you will see when barkhausen is satisfied (Aloop = gm Load gm Load)

Title: Re: CML divider self oscilation frequency
Post by raja.cedt on Mar 16^th, 2015, 2:50am

Hi jeffrey987,
Thanks for your reply, in-case of RLC it is clear but here only R//(cl//-2/gm), which boils down to Cl//-ve res hence this contribute only +ve phase shift and again this will be 90 only at high frequency by that frequency gain of diff pair will very low.

Correct me if I am wrong.

Raj.