The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> Literature wanted on Multi-Loop Ring Oscillator
https://designers-guide.org/forum/YaBB.pl?num=1255352311

Message started by Mayank on Oct 12^th, 2009, 5:58am

Title: Literature wanted on Multi-Loop Ring Oscillator
Post by Mayank on Oct 12^th, 2009, 5:58am

Hi,
Can anyone please suggest me some reading material on Multi-Loop Ring Oscillator.
I want to design 1GHz Low-Noise VCO (-60 dBc @ 10kHz offset, -110dBc @ 1MHz). Power and area are major concerns, so dont wanna shift to LC.
Application :- Analog Front End (Jitter Spec is a bit tight).
Reason for such Low-Noise VCO : Reference Freq is Low, so cant keep PLL BW above 900 kHz.

regards,
Mayank.

Title: Re: Literature wanted on Multi-Loop Ring Oscillator
Post by raja.cedt on Oct 12^th, 2009, 9:09pm

hi,
i think this can help you

A 1.25-GHz 0.35- m Monolithic CMOS PLL

Based on a Multiphase Ring Oscillator Lizhong Sun and Tadeusz A. Kwasniewski, Member, IEEE

thanks,
rajasekhar.

Title: Re: Literature wanted on Multi-Loop Ring Oscillator
Post by joel on Oct 21^st, 2009, 6:10pm

Curious about this. I had been taught that ring-osc VCOs are generally larger & higher power ( and jitter) than LC-osc VCOs.

I just finished a PLL design whose VCO is a 3 stage differential ringosc. The oscillator is about 80u X 40u in 0.13/0.35, and simulates fine up to 1.3GHz. I've got some spare area in this shuttle chip and I'm thinking of putting in the same PLL with an LC osc for curiosity's sake, having never desinged one before. The inductor is 100x100, but it's all in upper metal so it's sort of free. The varactor is very small, as far as I understand it. So the base-layer area appears to be much more compact than the ring-osc version. Is this what you've found to be the case in your experience?

By the way, silly newb question. What's a Multi-loop ring oscillator?

cheers, /jd

ps... Googling showed me this, http://www.springerlink.com/content/f856145p86665p52/ . The guy has a 0.18 'multiloop' ringosc running at 7GHz! But it's 440x430, and 50mA! Also he describes it as having a very narrow frequency range. I had also been taught that the big advantage of ringosc over LC VCOs was a wide range, if needed. There's some subtlety here that I don't understand. There also seems to be some patents associated with the multiloop thing. Dangerous? /jd

Title: Re: Literature wanted on Multi-Loop Ring Oscillator
Post by Mayank on Oct 26^th, 2009, 9:41am

Hi Joel,

Quote:

I had been taught that ring-osc VCOs are generally larger & higher power ( and jitter) than LC-osc VCOs.

Ring-Osc VCO's are inferior in performance to LC VCO's as far as noise is concerned. (Obvously as active implementation is far worse than passive in noise performance).
But Ring-Osc VCO's are primarily used to reduce area....Si Real estate costs increases with shrinking technologies...
Power required in Ring-Osc is variable and in designer's hand...If your Noise specs are not so stringent as in LC VCO's,(typically LC VCO's exhibit Phase Noise of -120 dBc @ 1 MHz Offset), You can work with Ring-Osci's at much lesser Power. The thing is you can trade-off Noise(say only -90 or -100 instead of -120dBc as i said earlier) and reduce power consumption of the VCO in ring osci.

But yes, if you try to make a Ring Osci VCO with comparable Noise Performance as that of an LC Equivalent, Ring Osci might take lesser area but it will consume nearly the same power as LC one.

Multi-Loop Ring Osci -- well a fundoo thing... :P but at the same time complex....Normally in a RING OSCI there is one single loop with Odd No. of Inverters. This Loop introduces a phase shift of 180 degrees before gain reaches 0 dB , basically rendering system unstable and causing it to oscillate.
The idea is what if you place multiple Loops in Feedback instead of a single one...This idea came to me sometime back..But someone had already worked on it...[ As always Late :P Well thats usually the case whenever I think I hv thought of something new :D ]

I actually thought of this multi-Loop Idea because i wanted better Noise performance from RING OSCI without burning much power n conserving area. I thought Multiple Loops will control Noise Better...But didnt find any substantial improvement in Noise..There is an alteration in speed..You can attain higher speeds at lesser power but didnt see any significant improvement in Noise performance as i had thought of initially.....Anyways, it was worth exploring... :)

Refer to the paper posted by raja above...It's very exhaustive and a nice read...

regards,
Mayank.

Title: Re: Literature wanted on Multi-Loop Ring Oscillator
Post by Mayank on Oct 26^th, 2009, 9:43am

Joel,
Sorry, forgot about your PLL....

As for your PLL, I would suggest use Ring Osci only if Noise spec is not so tight that it demands power cons. nearly equal to LC.

1 Ques : By differential RO , you mean CML (current mode logic) as in using differential inverters with reduced swings ??

--Mayank.