Coupled magnetics for interleaved boost converter with Kool Mu Core

[elecTomas]

Hi all,


I am currently designing a three-phase interleaved boost converter for a high-power low-cost application. For that reason, I would like to couple the three inductors into a single core, so costs reduction can be optimized.

I have calculated the inductor parameters obtaining the following results:
- Lequivalent = 0.6mH
- Lself-inductance = 2.56mH
- Lm = 1.56 mH
- Lleakage = 1mH

I want to couple the three inductors into a Kool Mu core I already have. My design approach for the inductor is the information provided by the core manufacturer for the air gap distributed cores. So starting point by calculating a first apporach of number of turn by using L=Al*N^2

Truely, I am not a mangetics expert at all, so from the results I have for me coupled inductor I do not know how to move on in the desig procedure.

Can anyone give some hints?

Thanks

 

[BradtheRad]

It's a great idea, but do you expect there could really be three different flux fields emanating from one core, acting independently?

 

[schmitt trigger]

The relationships between the main and leakage inductances are very critical and related to duty cycle, for proper ripple current steering. Otherwise you'll end up with terrible output ripple.

 

[FvM]

It's not completely clear how you arrived at the leakage parameters in post #1. Quite obviously the coupling factor depends besides winding distribution also on the powder core permeability.

The Al value is valid for an unformly distributed winding. Besides main and leakage inductance, you'll also care for total ampere turns and possible core saturation.

 

[Orson Cart]

Brad the rad has a point, one core for 3 phase PFC seems unlikely, can you post a schematic? have you modelled it?

 

[FvM]

As far as I understand, it's a multiphase buck converter, not a PFC, so the three inductors have ideally equal DC current, but different AC component.

I have seen multiphase designs with (partially) coupled inductors. It's not generally impossible if the coupling is low, e.g. < 0.6, but the designs didn't actually convince me. In any case the coupling causes recirculating currents loading the power switches.