Capacitive and Inductive Coupled EMI model

[newengr]

TL;DR Summary

I would like to create an EMI equivalent circuit model considering both inductive and capacitive coupling in one.

I am new to EMI modeling and just looking into it. I would like to create a circuit schematic to represent the inductive and capacitive coupling paths in one schematic. All that I've been able to find online is references discussing each independently. Is this the best approach if you want to consider both? Any references or suggestions for modeling both together would be helpful. Or if there's a reason not to, that would be helpful too.

 

[Baluncore]

We need a little more information.
What is the source of the EMI ?
What is the target ?

Directional couplers and short antennas can be hard to express on a circuit schematic.

 

[newengr]

The source of the EMI is a MOSFET switching on a DC bus. The target are nearby circuits referenced to a the midpoint of the bus or the -DC rail; both cases work. I extracted the parasitics which include capacitive coupling and inductive coupling terms using Q3D. I have the netlists for capacitive and inductive separate so I can draw the circuit for each. The challenge for me is trying to combine them.

I think combining them would be best if I can model my loads and source well enough. Then my circuit would have the appropriate voltage and current waveforms. Or is it better (or equivalent) to just take my voltage waveform and use it to analyze capacitive coupling the use my current waveform to look at inductive coupling? O and I'm neglecting radiated coupling at this point.

 

[Baluncore]

What form does the mutual C and L data take.
Does it employ nodes similar to a spice model?

If you can define the nodes on the EMI generator, and the nodes on the sensitive target, then you might define a scattering matrix that combines the reactive currents being delivered to the target.
Induced currents would appear between two nodes and could be simulated using transformers with coupling coefficients. Capacitive currents due to voltage changes would be between nodes. The method of moments might be considered.

I don't think EM radiation will be important for such a small space. It will all be near field.

 

[berkeman]

The source of the EMI is a MOSFET switching on a DC bus. The target are nearby circuits referenced to a the midpoint of the bus or the -DC rail

Probably shared impedance coupling will dominate the crosstalk, but without seeing your schematic and layout, it's hard to know. Can you share the schematic and layout?

 

[f95toli]

This might be a useful reference

Crosstalk and EMI on microwave circuit boards​

Rider, Todd William
https://krex.k-state.edu/dspace/handle/2097/35562

See section 2.3