# Changing coupling coefficient during transient simulation

Hi.
I would like to simulate two coupled inductors with coupling coefficient changing during the simulation.
what is the right way to define this?

See

couple.qsch (7.1 KB)

See

couple2.qsch (11.3 KB)

Hi.
thanks for the answer, but in my configuration (80kHz) this method has some convergence issues.
Can I just use the â€śk L1 L2 V(k)â€ť definition?

You think if it could be done that way, I wouldnâ€™t do it? Of course it canâ€™t be done.
And why didnâ€™t you try such a simple method yourself?

Hi.
I understand that my suggestion is not a valid solution, but I wanted to ask if there is additional solution to this issue? One that doesnâ€™t impose convergence problems.
Maybe Mike could update the solver to let the user to change the coupling coefficient during the simulation as was done in MicroCap.

I donâ€™t see any problems with convergence. I demonstrated the operation of magnetically coupled inductors at a frequency of 100 kHz. I changed the coupling coefficient linearly from 1 to 0 in 1 ms.
You just have little experience in simulations.
My method is quite simple and is based on physical laws. I doubt that MC does exactly what you described.
For example, I can change the magnetic coupling coefficient according to a sinusoidal law, but can you do the same in MC?
I think you are misguided by the capabilities of MC.
You must have seen parametric analysis, where the parameter is the coupling coefficient. Qspice allows you to do parametric analysis.
Give a link or a screenshot of how MC does it.

1 Like

I beg your pardon. I underestimated MicroCap.

1 Like

Do you think this leakage + magnetizing model can work in your case?
This is an equivalent circuit in my personal study note. Here I use arbitrary inductance device to setup k dependent inductance formula. In this example, the fixed coupling is at k=0.5, and compare to this model, when V(k)=0.5, both secondary output are matched.

If this circuit can work, I can make a sub-circuit symbol for you to use it.

k coupling model in my note

1 Like

I still have some issues with convergenceâ€¦