LTSpice and Transformer Coefficiency

[ACharnley]

Anyone know how accurate LTSpice is around coefficiency? I'm building a small flyback driver and using 1.00 have a rather excellent response. When I reduce it to 0.998 I get very little indeed. By 0.995 there's nothing.

I'll be lucky to get a transformer that does 0.995?.. so hoping it's a bug. The actual transformer I'm using is this, which doesn't specify the value. https://www.lcsc.com/product-detail/RJ45-Transformer_PROD-Tech-PSTFAQ3416-600T020_C3011570.html

Cheers,

Andrew

 

[ACharnley]

I am an idiot, missed the micro on the henrys!

 

[Papabravo]

I've never heard the term coefficiency (not a word according to the spell checker) before and I think what you are referring to is the coupling coefficient K that represents the deviation from ideality of a real transformer. It turns out that even when you specify K=1, what you actually get is just a small epsilon away from 1. It represents the leakage inductance present in all real transformers. There is an equivalence between the coupled inductor model and one that represents the magnetizing inductance, and the parasitic inductances of the primary and the secondary.

Glad you found the real problem.