Discussion Forum

Hi

I believe you should study the "d axi set up and try to implement the physics correctly in 3D (hopefully it will come once too from COMSOL) as this is in my view the only way to deal corectly with scin effects and multi-wire cables. There is a few other treads on this and some interesting article references fort the physics

If you manage, it would be of interest for the community, pls put it up on the model exchange

--
Good luck
Ivar

Just some clarifications and further thinking:

I am going to model the individual turns as geometry anyway, because there are only few, and we have to look at near-field effects.

But what about the skin and proximity effects within the single strands of a single turn? The strand diameter is of the order of the skin effect depth.

One important difference is that the strands in the litz wire are connected in parallel, while the turns of the coil are in series. That would make it necessary that literature like Meekers "Continuum Representation of Wound Coils via an Equivalent Foil Approach" would have to be reworked if I'd like to apply it to a single turn out of my multi-turn coil.

Or am I to complicated here, would something like

sigma_effective = sigma_copper / number_of_strands

do the thing? Probably not, because then skin effect still would be way to large.

Hi Matthias

Unfortunately I cannot give you any clear response here, but it's very interesting as this question about how to handle single/multiple strand cables with a correct model is not obvious for me. It must also be somewhat linked to the natural surface resistivity between the strand in the same wire, at which moment is it a bulk object or separate filaments ?

It would take me several days to dig into such a subject (time I do not have) and anyhow one would need to do some detailed measurements to calibrate any model, for me typically a good student study subject (but we have none of those "cheap labors" in our company ;)
(PS: I'm not COMSOL just a user, fan of studying finally physics again)

--
Good luck
Ivar

Hi Ivar,

thanks for your reply.

We are going for real high frequency litz wire, where the single strands are individually isolated, so inter-strand surface resistance is no matter here.

It boils down to a massive parallel connection of tiny strands wound into a coil.

Good hf litz wire follows well-designed patterns to make sure that each strand stays the same length on each position in the cable, averaging out all inner-outer-differences.

I'd say that since the single strand is in terms of diameter similar to the skin effect depth at our frequency (20-100 kHz), I'd totally neglect skin effect.

But I am still not sure about the proximity effect. And if we have no skin, but proximity effects, how to model it? Physically it's the same, but arises from different sources of the field.

Still thinking,

Matthias