The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Modeling >> Passive Devices >> Basic Inductor modeling question
https://designers-guide.org/forum/YaBB.pl?num=1282192629

Message started by aaron_do on Aug 18^th, 2010, 9:37pm

Title: Basic Inductor modeling question
Post by aaron_do on Aug 18^th, 2010, 9:37pm

Hi all,

I have designed an inductor in Virtuoso, and I exported a gds file and imported it into ADS momentum for EM simulation. OK now that I have my S-parameter data, how do I actually put it back into cadence so that I can run schematic simulations, and also will it be able to pass LVS?

I was thinking I could create an S2P file and use a 2-port black box, but I don't think it would be able to run simulations such as PSS or maybe even DC analysis in spectre. Also, i'm not sure if it can pass LVS. It would probably be better to create a proper RLC model of the inductor, so if somebody could point me in the right direction I would be very grateful.

thanks,
Aaron

Title: Re: Basic Inductor modeling question
Post by tkhan on Aug 19^th, 2010, 4:20am

http://edocs.soco.agilent.com/display/ads2009/Broadband+SPICE+Model+Generator
http://edocs.soco.agilent.com/display/ads2009/About+Narrowband+SPICE+Model+Generator

My colleague used these, but he said when compared with measurements neither of the models get all the extracted parameters correct.

Title: Re: Basic Inductor modeling question
Post by aaron_do on Aug 19^th, 2010, 6:32am

thanks. I managed to find that too, so now I need to figure out how use the spice file. I figure I can use cdl in cadence, but i'm not sure if there are any compatibility issues...

Aaron

Title: Re: Basic Inductor modeling question
Post by oermens on Aug 19^th, 2010, 8:21am

See http://www.designers-guide.org/Forum/YaBB.pl?num=1153231989

Title: Re: Basic Inductor modeling question
Post by pancho_hideboo on Aug 19^th, 2010, 8:22am

aaron_do wrote on Aug 18^th, 2010, 9:37pm:

I have designed an inductor in Virtuoso,
and I exported a gds file and imported it into ADS momentum for EM simulation.
OK now that I have my S-parameter data,
how do I actually put it back into cadence so that I can run schematic simulations,
and also will it be able to pass LVS?

Use Agilent Momentum on Agilent RFDE or Agilent GoldenGate with Cadence Assura.

If you run only simulation, you don't need Cadence Assura.

See chapter-10 of http://cp.literature.agilent.com/litweb/pdf/rfde2009/pdf/rfdemom.pdf

Title: Re: Basic Inductor modeling question
Post by aaron_do on Aug 19^th, 2010, 11:22pm

Hi pancho_hideboo,

thanks. Yeah I just found out that I have access to both RFDE and GoldenGate. So which one is better? I was told that Agilent GoldenGate is basically supposed to replace RFDE...Anyway I'll see if I can figure it out.

thanks,
Aaron

Title: Re: Basic Inductor modeling question
Post by pancho_hideboo on Aug 20^th, 2010, 8:30am

aaron_do wrote on Aug 19^th, 2010, 11:22pm:

Yeah I just found out that I have access to both RFDE and GoldenGate.
So which one is better?

It depends on purpose.

I'm using all of native ADS, RFDE and GoldenGate according to purpose.

See http://www.designers-guide.org/Forum/YaBB.pl?num=1268404959/7#7

Title: Re: Basic Inductor modeling question
Post by sheldon on Aug 22^nd, 2010, 5:38am

Aaron,

You can use the n-port with Spectre for pss and dc analysis. You
might want to use the most recent release MMSIM, since the
robustness of the n-port has been improving over time. The RF Kit
had an example of post-layout simulation with an inductor imported
from a third party tool. If you have problems, you should talk to your
local support team.

Best Regards,

Sheldon

Best Regards,

Sheldon

Title: Re: Basic Inductor modeling question
Post by aaron_do on Aug 23^rd, 2010, 4:02am

thanks. I wasn't aware of that. However, I still won't be able to pass LVS if I simply use an n-port. I'm having a bit of trouble starting up with GoldenGate, so for now I want to try and get this model working without RFDE or GoldenGate.

I managed to create a spice netlist from my s2p file, and I am able to import that into an RLC schematic, but now I need to know how to link the schematic to the layout so that I can run LVS when I use my inductor. Any help is appreciated.

thanks,
Aaron

Title: Re: Basic Inductor modeling question
Post by rfcooltools.com on Sep 7^th, 2010, 11:14am

aaron_do,

Use the n-port to simulate the inductor.
Place the n-port in a separate schematic view call it schematic_ind for example. then when you simulate set your switch viewlist in your simulation environment options to have schematic_ind before schematic. Or if you are familiar with the cadence hierarchy manager use that view in the config. you still will have a schematic view but it will only contain the pin names.

To handle the inductor in LVS is a separate step. First off an inductor without any special recognition layers in LVS is basically a short. So you will need to figure out how to resolve this, you may want to get some cad support on this, but here is what I typically do. Basically you need to identify in your pdk a layer that either acts as an LVS ignore or if your pdk has a metal resistor place it in your schematic and layout thus breaking up the shorted pins. if your inductor has three terminals then you will need at least two metal resistors. The metal resistors can be small remember it's only needed to remove the short.

Finally, if you plan to do this often you may want to work with your cad group to define a fool proof methodology moving forward. The reason is that of the two methods of LVS'ing I described above both can be jeopardized by carelessness during the layout. For example if a piece of extraneous metal is accidentally inserted into the region where the inductor resides and shorts out a turn or two and is not caught through visual inspection then it will be LVS clean, but desired performance will be incorrect.

http://rfcooltools.com

Title: Re: Basic Inductor modeling question
Post by aaron_do on Sep 7^th, 2010, 7:15pm

Hi,

my previous PDK had metal resistors, but this one doesn't seem to. Sounds like a good idea though. Anyway as you say its not fool proof, so I suppose I may not be able to do LVS on that part of the circuit. Thanks anyway,

Aaron

Title: Re: Basic Inductor modeling question
Post by modelman on Feb 1^st, 2011, 5:12am

FWITW

there are tools which create lumped element networks which can be used instead of the s-parameter blocks

one can also write Matlab optimization code to create a lumped element network that matches the s-parameters