The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> RF Design >> Design OSC
https://designers-guide.org/forum/YaBB.pl?num=1408021507

Message started by AZADBAKHT on Aug 14^th, 2014, 6:05am

Title: Design OSC
Post by AZADBAKHT on Aug 14^th, 2014, 6:05am

HI guys

i wanna design Frequency Synthesizer.when i design VCO,for OSC Resonance i constrained chose over 750 ohm Resistor.
in you're opinion,my design is valid?

Title: Re: Design OSC
Post by loose-electron on Aug 14^th, 2014, 5:06pm

Your question is not understandable.

Title: Re: Design OSC
Post by AZADBAKHT on Aug 14^th, 2014, 11:37pm

i'm so sorry loose-electron...
i try to edit my question

Title: Re: Design OSC
Post by aaron_do on Aug 15^th, 2014, 2:01am

Hi,

Its more accurate to model the inductor's parasitic resistance in series with the ideal inductor. Is that what you are asking?

Aaron

Title: Re: Design OSC
Post by AZADBAKHT on Aug 15^th, 2014, 6:36am

Hi aaron_do

yes,i know this resistance it's parasitic model for tank.
but,in you're opinion i can build an tank with 750 ohm resistance(parasitic model)?
at RLC parallel tank:

R=√(L/C)]*Q >>>> (L=1nH , C=10fF)

the formula illustrate for have R=750 must Q is high!!

you think i can build the this tank??

Title: Re: Design OSC
Post by aaron_do on Aug 15^th, 2014, 8:04am

Well...you didn't specify your operating frequency or what technology you are using to design your inductors. Where did you get that value for C? It looks like your operating frequency is extremely high...

If your PDK has inductors you can easily check. Otherwise you can use an EM simulator.

Aaron

Title: Re: Design OSC
Post by AZADBAKHT on Aug 15^th, 2014, 9:33am

ok ,i working at CMOS-90nm Technology
VCO tuning range is 5GHZ-10GHZ then C capacity is vicinity 40fF-10fF
and i simulate the design and calculate Q
Q is 4-6.66

with this Q i can't design an Tank with 750 ohm loss
nonetheless,in you're opinion i can use a real Resistance in Tank for post layout simulation?

Title: Re: Design OSC
Post by loose-electron on Aug 15^th, 2014, 10:06am

At the frequency stated you need a much better model of the inductors and the non ideal parasitics. To get meaningful results you will also need a set of parasitics for every node in the system.

Title: Re: Design OSC
Post by AZADBAKHT on Aug 15^th, 2014, 12:16pm

tnx a lot sir loose-electron.
how i can set parasitic for every node in system?
my software is cadence.

Title: Re: Design OSC
Post by AZADBAKHT on Aug 15^th, 2014, 1:01pm

tnx a lot sir loose-electron.
how i can set parasitic for every node in system?
my software is cadence.

Title: Re: Design OSC
Post by aaron_do on Aug 15^th, 2014, 6:48pm

Since you know what post-layout simulation is, you know how to get the parasitics at all the nodes. However, you need some way to extract inductance too. Try an EM simulator like ADS momentum. As loose-electron implied, if you model the inductor as just an ideal inductor with a parallel resistance at that frequency, your results will be wrong.

regards,
Aaron

Title: Re: Design OSC
Post by AZADBAKHT on Aug 15^th, 2014, 10:59pm

ok
understand,can i use the cadence-virtuoso software for design inductor real?
or i must use the ADS?
i don't know how to set the parasitic model at all node in the system in cadence-virtuoso.
you know how that?
tnx a lot Aaron.

Title: Re: Design OSC
Post by loose-electron on Aug 16^th, 2014, 5:23pm

Suggest that you talk with your design team locally about getting better models. Everybody sets up their model sets differently so you need to contact whoever is responsible for the design tools and semiconductor models that you are using

Title: Re: Design OSC
Post by AZADBAKHT on Aug 16^th, 2014, 10:40pm

ok
tnx a lot for you're help sir loose-electron

Title: Re: Design OSC
Post by aaron_do on Aug 17^th, 2014, 6:31am

Hi,

Quote:

i don't know how to set the parasitic model at all node in the system in cadence-virtuoso.

after creating your layout, you need to create an extracted model using Calibre or Assura. But you need to design the inductors separately using an EM simulator.

Actually these are the something like the basics of using the software, and it would be better if you could find somebody in your own lab who can go over this with you (I'm assuming you're doing this at a University).

regards,
Aaron

Title: Re: Design OSC
Post by AZADBAKHT on Aug 17^th, 2014, 8:51am

tnx a lot sir Aaron

really,you help me a lot.
i haven't team and advance lab in my university . so i have very question
i'm sorry for have this problem. :-[

Title: Re: Design OSC
Post by aaron_do on Aug 17^th, 2014, 6:10pm

Hi,

So you're the first in your university to work on such a project? That sounds tough, especially for RF work since you will probably need to design the inductors in an EM simulator and then import the s-parameter models.

Assuming you're using Assura, you need to run DRC and LVS. Once your circuit has passed LVS, you need to run QRC. After QRC, an "av_extracted" view will then appear in your library. You need to run simulations on this av_extracted view. You can do so in one of two ways. First you could go to your environment settings in ADE and type in av_extracted at the front of the switch view list. Alternatively you could create a "config" view which will allow you to change the view to simulate block-by-block.

Since this is RF, you shouldn't run QRC on the inductors (unless they are part of the PDK). You need to use an EM simulator to find the S-parameters of the inductor. You can then include the s-parameters in the simulation using the nport component from analogLib...

There are a lot of details I haven't included here because it would be too long. Try running QRC and simulate the extracted view first. Feel free to ask further questions later.

regards,
Aaron

Title: Re: Design OSC
Post by AZADBAKHT on Aug 17^th, 2014, 11:39pm

Hi
Yes i'm first person to work at this project.
tnx a lot for you're help.i try to design...

Title: Re: Design OSC
Post by totowo on Nov 16^th, 2014, 5:38pm

Hi, AZADBAKHT,

" R=√(L/C)]*Q "
This equation is only approximately correct @SRF(Self-Resonance Frequency), where omega=1/sqrt(LC). The SRF is definitely not your oscilation frequency.

You should use R=omega*L*Q to estimate the impedance, where the omega is the oscilation frequency you want.

The different R values roots in the Series-Parallel Transform.

AZADBAKHT wrote on Aug 15^th, 2014, 6:36am:

Title: Re: Design OSC
Post by AZADBAKHT on Nov 16^th, 2014, 9:22pm

Hi totowo,
tnx a lot for your explain.