Continue to Site

why do 3 winding mutual inductors get unstable

Status
Not open for further replies.

ahero_hu

Newbie level 5
Hello,

My simulation schematic is a voltage source stimulating an inductor L0, coupled to another 2 inductors (L1 and L2, through k1, k2). each of them is connected in series with a 50 Ohm resistors. The voltage source is 64.3V-amplitude 2MHz sine wave.
The simulation result of V_L0, V_L1, V_L2 are shown in the second graph, representing transient voltage drop across L0, L1, L2, respectively.

Does any one know what is the cause of this oscillation?
View attachment mind sim.bmp

The mutual inductances are used to model the EMI coupling in EMC BCI test.

Seems like a simulation artefact. most likely a problem specific to your simulator. I would try a more realistic coupling slightly below 1. You will never experience k=1 with a real component anyway.

ahero_hu

ahero_hu

Points: 2
K=1 for K1 and K2 need to be adjusted for practical case. This implies L1 and L2 also with K which will aid under transition and cause oscillations in the simulations.

Thank you, FvM, it is indeed solved by reducing the coupling factor.

---------- Post added at 17:15 ---------- Previous post was at 17:11 ----------

Hi, kspalla,
I am actually curious about how this oscillation is generated?
How do you analyze this?

I am actually curious about how this oscillation is generated?
I doubt, if there's a simple answer. I expect it can be explained by the finite numerical accuracy of the simulator. The strange thing is however, that the circuit seems to have sufficient losses to act nonideally. Technically, the three inductors should simply act as one.

magic of simulator

Thank you, guys. It either has something to do with simulator or "inductance matrix stability", as I heard from somebody else.

That's why connect a 1GOhm to each node and try to set the initial node voltages as 0 or 1 V etc..

Yes, but according to the schematic, all three inductors have one side connected to gnd node.

I don't work with cadence, so I can't say if published cadence schematics must have that bad appearance and readability they usually have at edaboard...

ahero_hu

Points: 2