How to simualte CMOS spiral inductor ?

[cswang]

Hi !
Anyone know which EM tool can precisely predict CMOS spiral inductor Q and model ? Where can find the example or document ?

 

[cwcwecan]

First, You should have the process parameters
(e.g. Si:Er=11.9, Loss Tangent, and hight
Every Poly's , Loss Tangent, and hight
Every Metal's conductivity)

Then, Import the inductor's gds file to the EM simulator and Key in the process parameter above.

You can use ADS's momentum, Sonet, HFSS, Microwave Office, IE3D as your EM Simulator. According to my experiences, if the process parameters are right, you can get simulation results exactly.

 

[cswang]

First, You should have the process parameters
(e.g. Si:Er=11.9, Loss Tangent, and hight
Every Poly's , Loss Tangent, and hight
Every Metal's conductivity)

Then, Import the inductor's gds file to the EM simulator and Key in the process parameter above.

You can use @DS's momentum, Sonet, HFSS, Microwave Office, IE3D as your EM Simulator. According to my experiences, if the process parameters are right, you can get simulation results exactly.

I know that , but in my experement , to estimate inductor is not't hard , but Q is really difficult! Each tool will get different Q value and also different from our measurement data !
any idea ?
thanks!

 

[Manoochehr]

Use ASITIC

ASITIC is a CAD tool that aids RF/microwave engineers to analyze, model, and optimize passive metal structures residing on a lossy conductive substrate. This includes inductors, transformers, capacitors, transmission lines, interconnect, and substrate coupling analysis.

To download ASITIC look at the site of Berkeley University(eecs.berkeley.edu). I have used it some times ago and it was realy good software.

 

[harkonnen]

EM-Simulating for accurate Q on MMIC indutors can be an extensive task, especially with high Q values.
In case your eddy-currents (lossy substrate) domiante the loss it might be a little easier. But when your main loss is the bulk conductivity and skin-loss (SOI etc.), it becomes difficult. Then it depends how thick your microstrip line is compared to the skin-depth and how closely coupled adjacent lines are.
Basically you will have to understand how the simulator handels losses and how it models skin effects.