PFC Inductor Design

[fethiyeli]

Hi everyone. I need to design a PFC inductor using toroidal core. When I looked at the Magnetics website, I couldn't figure out which current I should use. I mean sometimes they use DC current for core selection whereas sometimes they use peak current. I just want to make sure LI^2 calculation. Which current do I need to use ?

Magnetics - Inductor Design with Magnetics Powder Cores

Magnetics offers a number of inductor and transformer design tools and literature to assist engineers in optimizing their Magnetics components. Contact our Applications Engineers with your design questions.

www.mag-inc.com

Magnetics - Inductor Design with Magnetics Ferrite Cores

Magnetics offers a number of inductor and transformer design tools and literature to assist engineers in optimizing their Magnetics components. Contact our Applications Engineers with your design questions.

www.mag-inc.com

Sometimes in 20kHz air conditioner PFC circuits, I see being used of reactor as shown below in mH values. Do I really need mH values like 3.5mH or something like that ? What are the benefits of using that value of inductance ?

Line Reactor | Air Conditioner Reactor | Yunlu

This line reactor is widely used for power-factor correction of passive or partial PFC schemes and is suitable for schemes with a frequency lower than 20KHz.

ylenergytech.com

 

[cupoftea]

You mean a Boost PFC?
The powder cores are non linear, so there inductance varies with the current through them. Pretend though, at first, that you are using a linear inductor, of a certain fixed value L. Make this L, that L that gives you approx 10% ripple at the peak of the mains. Then you do a powder core inductor that has that value of inductance L, when it is passing the mains peak current level. (i mean the filtered mains input current to the product, not the peak of the current spike).
So anyway, say your PFC is 1kW....and VAC is 240VAC, then your mains input current is 1kW/240 = 4.17A RMS. So peak mains input current is 5.9A.....so you design your powder core inductor to give you the inductance that gives you 10% ripple "at 5.9A"....ie, at the peak of the mains.

I abbreviated that a bit so i hope you followed.

You actually do a boost converter design for the static case of mains peak being the input voltage.

0.5LI^2.....just use the I rms mains input current for the I.


To calculate the peak B, (you will need to eventually) you need to build it up bit by bit. So you increment H a tiny bit, then calculate dB/dH (= u0.ur)...then use that to calculate where the next B value will be...then calculate the dB between those values...then add up all the dB's, to give you the B peak value....its kind of funny with non linear inductors, you need to do a "bit by bit" calculation to get the peak B.
Theres some threads in this forum about it, with the excel calculation sheet.

If you increment H by very small steps, then this method is pretty accurate to calculate B pk.

...remember, increment H, calculate dB, increment H, calculate dB, increment H, calculate dB, etc etc....then add up all the dB's.
You have to do it like this as the permeability varies with the current.

I am sure you are aware of "bit by bit" methods...eg....integration where you consider a circle to be made up ot millions of little isosceles triangles.

Remember that dB/dH = u0.ur
and reluctance = N^2/L
and reluctance also = path length / (uo.ur.A)


In here you find a folder on "non linear inductors, if you want....

https://drive.google.com/file/d/1WvElhDNfHI0EBei4rouSNEdWWTO1oCtJ/view?usp=sharing

Also in the "PFC" folder, there is a folder called....
"PFC _non linear inductors"

..this coudl be good for you.

Attached also is an LTspice sim of a PFC so you can practivce your inductor value on it....LTspice is free