Forward converter with peak-current mode controller

[moonz]

Why stable operation of a forward converter with peak-current mode controller, the magnetization current should be less than the reflected current?

I figured out the forward converter, and I'm currently reading app-note(https://www.analog.com/media/en/technical-documentation/app-notes/designing-activeclamp-forward-converters-using-peakcurrentmode-controllers.pdf) where state:

The converter is a peak current-mode controller that controls the peak current as seen by the current-sense resistor, which is the sum of the reflected load current and magnetizing current. It is necessary that the magnitude of reflected load current is always more than the magnetizing current for stable converter operation and well regulated output. This condition is always satisfied, if the magnitude of reflected load current at the input minimum is more than the magnetizing current. Hence, peak-to-peak primary magnetizing current is assumed to be half of the reflected load current at minimum input voltage

Do I understand correctly that it is needed in order for the feedback to 'know' what is really occurring in the load?

If the answer is yes, then how will the forward converter work with a slight load?

[mtwieg]

Why stable operation of a forward converter with peak-current mode controller, the magnetization current should be less than the reflected current?

I figured out the forward converter, and I'm currently reading app-note(https://www.analog.com/media/en/technical-documentation/app-notes/designing-activeclamp-forward-converters-using-peakcurrentmode-controllers.pdf) where state:

The converter is a peak current-mode controller that controls the peak current as seen by the current-sense resistor, which is the sum of the reflected load current and magnetizing current. It is necessary that the magnitude of reflected load current is always more than the magnetizing current for stable converter operation and well regulated output. This condition is always satisfied, if the magnitude of reflected load current at the input minimum is more than the magnetizing current. Hence, peak-to-peak primary magnetizing current is assumed to be half of the reflected load current at minimum input voltage

Do I understand correctly that it is needed in order for the feedback to 'know' what is really occurring in the load?

If the answer is yes, then how will the forward converter work with a slight load?

I've designed peak CMC forward converters before and never observed this, and I can't see any reason why this would be true. In fact, magnetizing current effectively acts as a bit of slope compensation, helping against subharmonic oscillation.

[Faringdon]

Why stable operation of a forward converter with peak-current mode controller, the magnetization current should be less than the reflected current?

It generally is...but yes in light load the magnetizing current peak may be more than the reflected...but who cares, you can still get regulation.

Turn the question round.......why do you want magnetizing current to be greater than reflected load current?
....or, what is it about magnetizing current that you dont like?

But generally, yes,  the load current is what you are trying to control, so why do you want much magnetizing current......magnetizing current is what you are forced to put up with.....due to V = Ldi/dt....but  as Mtwieg says...magnetizing current does give you a  nice bit of slope compensation if you need that.

[moonz]

why do you want magnetizing current to be greater than reflected load current?

I don't want  . Magnetizing current is a detrimental effect for a forward converter, as it doesn't provide power to the load; instead, it increases power loss. Right?

I would like to continue discussing the topic of forward converters. Can someone please review my approach for selecting a core for the converter?

So... Input data:
Input voltage min: 10V
Out Power: 70W
Out Voltage: 54V
Maximum duty cycle(Dmax): 0.6 (I know, i's not typicaly for forward but for MAX17598 it allow)
Frequency switching: 1MHz

1) Calculate out current in load:
Iout=Pout/Vout = 1.3A

2) Calculate turns ratio:
n= Vout/(Vin*Dmax) = 9

3)  Calculate secondary current:
I skipped the inductor selection step, just say that ripple current are 0.33A. The secondary peak current Isec = Iout + Irip = 1.3+0.33 = 1.63A

4) Calculate primary current:
Ipri = Isec * n

5) Calculate magnetization inductance:
Suppose that the magnetization current should be less than 20% of the full load current. Imag = 0.2*Ipri
Then Lmag = 2uH (from V=L*di/dt)

For example choose B66285G0050X187


H = Ipri*Turns/le = 451 (for Al = 1520nH enought 1 turns in primary side) = Ipri/le = 586. Ipri iclude magnetization  current.
B = H * ue * Al = 0.35; It's less then 0.4 (saturation value),and this core is suitable for this design.

Is everything correct in my reasoning or did I make a mistake somewhere?

[jonpaul]

PC200 is the power material

All others like N87 are not for power applications.

Magnetizing current is a natural result of any transformers primary inductance. As it is inductove and not resistive, there is no power loss just VARS.

j

[moonz]

PC200 is the power material

All others like N87 are not for power applications.

What does the PC200 have that the N87 doesn’t?

I wouldn't be so strict because N87 is still used in DC/DC applications. For example TI: https://www.ti.com/lit/ug/tidub83/tidub83.pdf

Also, the manufacturer himself says that the main application is power supplies.

But apparently still this material does not suit me because I want to work at a frequency of 1 MHz

[moonz]

As it is inductove and not resistive, there is no power loss

This is a very strong simplification. This current still makes losses: I2R (on the transistor and winding) plus dynamic losses in the core.

[mtwieg]

PC200 is the power material

All others like N87 are not for power applications.

TDK's PC200 is one of several relatively new ferrite materials (like 3F46 from ferroxcube, 80 from fair-rite, etc). For switching frequencies >500kHz they work well, but these new materials didn't make the old materials obsolete. Also last time I checked these high frequency materials are not available in many form factors.

As it is inductove and not resistive, there is no power loss

This is a very strong simplification. This current still makes losses: I2R (on the transistor and winding) plus dynamic losses in the core.

Agreed, though depending on the design magnetizing current isn't necessarily a bad thing. In some converter topologies, magnetizing inductance helps achieve soft-switching in the semiconductors. But in the case of a hard-switched converter like the forward converter this is not the case. Magnetizing current means there will be core loss, and also slightly increased conduction losses in the FETs.

[MrPWM]

There is no need to make the magnetizing current "half" of the reflected current. After reading that, I don't know why the author even said that. I've designed quite a few forward converters and the magnetizing current is usually a tiny percentage of the reflected current.