momentum simulation

I got a problem when I use momentum to simulate inductor/transformer. My inductor spiral is at M6 and is connected to underpass at M5 through via5 in D5. So, my substrate setup is like this

------Strip m6 (thickness expansion down)
D5 [VIA] via5
------Strip m5

When I check the structure in 3D view, I found via5 starts from top surface of m5 to the top surface of m6. However, via5 is assumed inserted between top surface of m5 and the bottom surface of m6. I guess it is because "thickness expansion down" option of m6, which expand the dielectric layer with the same thickness of m6 and with the same permitivity of D5.

Does anyone have experience to solve this problem? Thanks.

What's happening here ? -)

simulation credit

You should substract the thickness of m6 by thickness of via5.

For example, if now your m6 thickness is a, and via5 thickness is x then you should change your m6 thickness to (a-x).

I've aksed Agilent supporting engineer. He confirmed my set-up is correct and this "error" is result by Momentum. In fact, Momentum is 2-D simulator. So, when we use thickness expansion of metal trace, such "error" will happen. Anyway, the Momentum simulation results are close to those by Sonnet. Here is 3-D structure of the transformer I simulated. As shown in the picture, via8 (grey) is overlaid with M9 (green). But, actually, via8 should only reach to the bottom of M9.

did you try to simulate in EMDS? I often use EMDS because it's faster. However, I think Momemtum is more accurate than EMDS.

I haven't tried EMDS because we don't have license. Do they use the same engine?

rfkevin,

Do you think it would make any difference that the via extends into the entire thickness of the top metal? I would have thought that it wouldn't make any difference because you have "metal intersecting metal" and the two become a single metal.