Flyback Lm value in CRM/BCM design

I would like to use an offline PFC flyback controller that operates in critical/boundary conduction mode.
http://cds.linear.com/docs/en/datasheet/3748fa.pdf

I am stuck in trying to figure out what my magnetizing inductance Lm for my flyback transformer should be! The datasheet only recommends part numbers, but I would like to know how they figured these out since my requirements are different.


In the datasheet, my switching frequency can vary between 4kHz and 150kHz to accommodate this mode. Do I make my Lm calculation at the maximum or the minimum frequency? I think the maximum current draw will force Ton to be maximized, and Ton can be largest at low frequency - so 4kHz. If I use my turns ratio and Vin/Vout ratio to calculate my duty cycle, I can then use the V = Ldi/dt formula to determine the L at minimum input voltage/maximum load current. Am I on the right track with this, or completely off?

I am now finding that I should not use the minimum switching frequency of 4khz, but instead choose a switching frequency minimum that will size the inductor for me. But I still think it is correct that my switching frequency minimum must be used for max load and max current draw (Vin min).

Are you sure you linked that right document? That's not a PFC controller....

Oh my, you are right. Sorry for flopping my question so badly!

This is the part I'm using actually:
https://www.linear.com/product/LT3798

I came to a value of 450uH, considering 70 VAC as minimum input and max load on output as 18V and 1.5Amps. That should give me a minimum switching frequency at 40kHz (during the peak of the line voltage sine wave at max power) .

Your figures are in the ballpark, more or less.

My simulation uses a 450 uH coil, and a DC supply of 70 V. (I was not sure how to use a 70 VAC supply as your plan states.)

Duty cycle is 50 percent (typical for a flyback).

Performance is adjusted to provide your output specification of 18 V at 1.5 A.



I tried various operating frequencies. A suitable frequency appears to be 50 kHz.
If I make it slower then the output voltage rises.
Faster and the output voltage drops gradually.

The waveforms appear to be more efficient at 50 kHz.