I'm curious about the design of RF transformers with purposely designed non-unity coupling coefficient. I can see how this would be done in an IC, but how feasible is it at the discrete component level?
For example, if I made two air core inductors and placed them axially aligned in close proximity, what kind of ranges of coupling coefficient could I hope to achieve? And importantly, how accurate could I get that coupling coefficient to be?
If I wanted to get high, but precise, coupling coefficient, could I use a toroidal transformer? I'm thinking not, since having tight control over the coupling would be fairly difficult.
What piqued my interest on this topic was reading about T-coil peaking, which requires very particular coupled inductors.
Thanks.
Best Answer
If you are looking for some degree of precision in the value of k then use near-unity coupling transformers and use an external shielded inductor to create the uncoupled inductance you require.
From zero to probably around 0.5 coupling coefficient. The trouble is that without an extensive analysis of the fields using a finite element tool, you wouldn't really know if your coupling is within +/- 50% of your target AND, other things in the vicinity could easily influence that coupling.
Use a toroid and an external shielded inductor to create the leakage inductance.