How can I simulate an SMD inductor using CST?

[ac478]

Suppose I want to view the EM fields generated by a basic LC oscillator using SMD components mounted on a PCB. What would be the most reasonable approach for this task? Particularly for inductors, it doesn't seem to be feasible to import 3D CAD models into CST and define material data there, because not every inductor has a detailed CAD model. In certain CAD models, where the inductor is housed in a compact case the CAD file does not even include the coil itself. So, what alternative methods could be used here? Should I simply search for different inductors until I find one with a well-detailed CAD model?

 

[FvM]

In certain CAD models, where the inductor is housed in a compact case the CAD file does not even include the coil itself.

Sounds reasonable considering the CAD model purpose. It's mechanical PCA modeling, not magnetic design.

I see however no problem to identify magnetic path and coil dimensions of many popular SMD inductors.

 

[ac478]

Sounds reasonable considering the CAD model purpose. It's mechanical PCA modeling, not magnetic design.

I see however no problem to identify magnetic path and coil dimensions of many popular SMD inductors.

So this means I should model a simple inductor in CST, myself, with coil dimensions and core material type taken from the datasheet and this would be enough right? It doesn't have to be a 1:1 replica of the real component, in terms of packaging (the SMD inductors I work with will not have a magnetic shielding) and other details that are unnecessary for the EM simulation.?

 

[BradtheRad]

You may find a CAD-based simulation out there to guide you. One big difference is whether a coil has a core of air or of metal. A metal core attracts and concentrates flux so it's hard to be sure what sort of invisible field extends out of the inductor.

My own home-brew simulator might offer insights as to inductor behavior. Flux field becomes strongest at the same moment when current becomes strongest. Current rises in accord with inductive time constant (by formula L/R). However counter-emf hits maximum immediately. As current rises, counter-emf declines.

My Youtube video 'Tank Circuit in operation (animated)'. It depicts inductor flux field building and collapsing in accord with current flow back and forth in an LC tank.

 

[volker@muehlhaus]

Don't do it.

For 99% of use cases, it is a waste of time to model the SMD inductor in such detail. Instead, use the CST inductor model which simulates the inductor as a black box. This inserts inductance into the signal path, and you can analyze your PCB as desired.

More detail is necessary only for special cases, e.g. if you have two closely spaced inductors which couple through the components's magnetic field. But that's a case that you want to avoid during PCB outing anyway, by placing them at 90° offset.

 

[ac478]

Don't do it.

For 99% of use cases, it is a waste of time to model the SMD inductor in such detail. Instead, use the CST inductor model which simulates the inductor as a black box. This inserts inductance into the signal path, and you can analyze your PCB as desired.

More detail is necessary only for special cases, e.g. if you have two closely spaced inductors which couple through the components's magnetic field. But that's a case that you want to avoid during PCB outing anyway, by placing them at 90° offset.

Hello, thank you very much for your answer.
I have a small question: If I define the location, where the SMD components will be attached on the PCB, as a discrete port and add the lumped components in the schematic view (I think this is what you meant or adding the lumped component directly in 3D view), will the 3D simulator also take into account fields created by these lumped components?

I know simulating the fields generated by an LC oscillator sounds trivial and uneccesary but I have to do it, manufacture the PCB, measure the near-field with probes and compare the simulation and measurement results. I need the measurement data and the simulation results from CST to match together, not perfectly but approximately.

 

[volker@muehlhaus]

as a discrete port and add the lumped components in the schematic view

Yes, or maybe CST allows to insert an ideal lumped inductor directly into the model without using a port. I'm using another EM tool (Empire XPU) where both options are available.

I have to do it, manufacture the PCB, measure the near-field with probes and compare the simulation and measurement results. I need the measurement data and the simulation results from CST to match together, not perfectly but approximately.

I see ... modelling the lumped L will give you the correct results except for the local fields directly around the inductor. If you think of the wirewound coil, the fields from all the individual turns add up inside the device, and cancel outside the device. What remains is the field from the current through the device, which you would also see when replacing the inductor with a through line + lumped inductance.

In my EM tool (Empire XPU) there are SMD components which are built that way: they have a current path at some height above the PCB, with the L value inserted as a lumped component. This should be fine for your use case in CST as well, unless you really need to study the local fields directly at the device.