Fusion Signal Integrity Ext: via simulation capability?

[ragnar_tollin]

Hi,

I was looking at the SI extension, but in the videos from what I can tell, it cannot simlate impedance and xtalk for vias? At least all the ads are only for diff pair xtalk. I am specifically looking for a tool that can do this for a multilayer board.

I tried customer support but they gave me nonsense answers, except that I need to pay for it to "test" all the features, but I want to know that it has the features I need before purchasing, so I will give this a last shot here. I want to be able to simlulate diffpair including vias, layer trasitions, anti-pads etc. Can it do this?

[jorge_garcia]

Hello @ragnar_tollin,

If you are on a Fusion commercial trial, you can try the extensions for free for 14 days if memory serves. When you pick a signal for analysis, the entirety of that signal including vias are taken into account for the impedance analysis. Nearby signals on other layers are also taken into account, so for example if you have a gap in a polygon this will be shown as a change in impedance in the trace.

I'll share this with our devs for double confirmation, if I've gotten any of the above wrong I'll come back to correct it.

Let me know if there's anything else I can do for you.

Best Regards,

[jorge_garcia · Accepted solution]

Hi @ragnar_tollin,

So the devs got back to me, vias are not taken into account but the layer transitions are taken into account when doing the impedance calculation.

Let me know if you have any other questions or concerns.

Best Regards,

[ragnar_tollin]

Thanks for good support, then I will hold off on getting the extension and feel free to share this as a feature request to the dev. team. I suspect however that this is challenging since accoriding to my research 3d simulations is the correct way to find the via impedance. But even 2d simulations would be sufficient for me.

[aholtzma]

I kicked the tires on this extension, and there are probably too many limitations to be useful. On top of via's not being simulated, the pads aren't either. Also the return current is magically transported between reference layers on layer transitions (perhaps due to the no-via limitations). Layer and trace/component transitions are the primary reason you use an EM tool like this and it falls short at this point in time.