HFSS accuracy for spiral inductor design

Hi guys,

I want to know how accurate is HFSS to help me guess the value and quality factor of a planar spiral inductor on silicon (integrated circuit).

I want to know if, based on the layers description (insulators, metals and silicon) HFSS is able to simulate losses and give to me an accurate set of S parameters.

Thanks a lot

Come on guys! I can't believe you don't have any aidea

Why not ?

Since it is based on FEM method, the algo itself will include the loss mechanism. However, I do not have pratical experience on that topic.

At my work we have compaired HFSS simulations with measured results
and have found good correlation. You need to spend some time setting
up the simulation correctly.

We also simulated the same inductors with momentum but the results where not that accurate.

Regards

rf

Humungus,

for integrated circuit,a planar spiral inductor on silicon is too small for HFSS. tendness towards surface in high freq may cause much different.
So U can get a approache inductance value, but cannot get a sure Q.

those just disturb me, so I must process a serial planar spiral inductors to make sure.

So, to evaluate the Q factor of an integrated inductor, the strategy is try and error, isn't it?

Hi
Once I simulated such an inductor in HFSS and sonnet in order to compare the results. The inductor had 8-9 turns.
The sonnet gave me better S paramter results compare to HFSS. I can say that HFSS was not so accurate with my settings.

there are 3 way to simulate spiral inductors:
1. 3D simulator like hfSS and Sonnet. they are the most accurate solution but require a lot of simulation time (hours!) and alot of memory and high speed processor!

2. 2.5D simulator like momenum sim. of ADS. they are much faster than 3D simulator (in minutes by using RF mode not Microwave!) and very lower memory. The accuracy of them is a little lower than 3D's.
plesase be careful when using this simulators; they approximate the thikness of metal strips to zero thikness sheets. you must use via layers to simulate the thikness!

3.using ASITIC which is the fastest spiral simulator so far.It divides the spiral to smaller transmition line and solves the circuit by some simplifications. Sim. time is in seconds. although it's accuracy is lower than other method, it can help for finding initial guess of dimensions for required inductance and optimum Q factor @ the frquency of your interest.

thank you