inductor em

plz

i want to know what is the best EM simulator to simulate the spiral inductor on sillicon substrate accuratly

thanks

Momentum of Agilent in MW Mode.
It gives quite accurate results.

I am trying to simulate spiral inductors on silicon too. I will use TSMC0.18um process. but I cannot find the updated process parameters, such as the thickness of metal layer 6 and metal layers, permittivity and permeability of substrate, oxide layers. Could you guys tell me what they are or where I can find them? Thanks.

I found in one of TSMC documents mentioning that the thickness of a AlCu layer used to create inductor is 2 um, is that true?

i used AMS 0.35 technology , it has 4 metal layers . the 4th layer was about 1micron

and does anyone know , which is the best in sipral inducor simulation

sonnet or HFSS

thanks

Hi, Khouly:

You can give IE3D a try. I know that one major semi-conductor company tested all the major EM simulators in the market on modeling spiral inductors and they told me that IE3D yields the best accuracy. Regards.

Not familiar with Si inductors, be careful if you have multilayer substrate (more than 4-5 layers), IE3D might be not the right tool. I did some comparison of multilayer inductor analysis in IE3D, Sonnet and measured values, Sonnet predicted self-resonsnce within 50MHz, IE3D was 2 GHz away! Has something to do with layer number, Sonnet works fine for many layers (checked up to 16!)

FlyHigh

IE3D is good for any layers. However, you need to make sure that you use fine enough cells. The regular rule of thumb of 20 cells per wavelength does not work. For spiral inductors, you need to make sure the inner most section to be divided into at least 2 cells. Then, you will get good results. The reason is that, even a spiral is small, the current is changing very fast. You do need more cells than regular structure. The trick may be mentioned in the IE3D user's manual.

for fast spiral inductor simulation and optimization use ASITIC.
for accurate simulation and verification use sonnet or hfss or even A-D-S.
Note: A-D-S is 2.5D EM simulator so it is not very accurate and u must add long vias for simulating metal thickness or interwindings capacitance which determines self resonance and changes Q of the inductor specially in symetric spiral inductors used in LC tank oscillators and mixers.

BEST!

Well, yes, the uncarefull meshing might be the case why I did't get the right result, although I analyzed multi layer thick metal 3-D helical inductors rather than spiral but current crowding might happen there as well. I start to test quality of solution (convergence) by gradually increasing per wavelength mesh density. The solutions were changed to some extent but they were getting much slower as well, so I had to quit at some point. Acctualy I didn't test accuracy because of speed issue. It was more than 2 years ago, I didn't use IE3D since, but some of my colleagues did and they are quite happy with it. Very successful microstrip and CPW circuits such as 40GHz filters, planar antenna and planar baluns were made base on IE3D simulation. Also current version is 10 and I was working with 6 so there were many improvements since.

OEA has a dedicated program for the simulation of inductors on-chip, the program can be found at www.eecad.com. The bad thing is that the licensing fee is very hefty, and the program which is possible to download at eecad is encrypted (needs a decrypting key of some kind). Does anyone have any experience with this OEA program? Or, better yet, does anyone have the decrypting key for the program? :P

/Pim

I used Maxwell 8 to simulate some Spiral/Rectangular inductors on chip. But be not sattisfied with the postprossecing interface.
So I want to try Maxwell 9 or IE3d or even good FEMLAB.

I used ASITIC for Rectangular spirals and it is very close to measurements.
For Octagons it was 25% off!! That does not refer to the latest release which might have had some improvements

Momentum is a good tool but again 2.5 D simulator....

hope that helps

colombo2

Sonnet is faster and more accurate in spiral inductor simulation if you know some of the tips and trick.