Two-switch forward: clamping diodes current ratings

Hello,

I need to design a 300W 48V->12V two-switch forward converter and I need to know how to properly calculate the two clamping diodes current ratings. They're only conducting the transformer magnetizing current (during the core reset phase) hence the question might be: is there any rule of thumb to aproximate the magnetizing current relative to the primary transformer curent (peak/average)?

Thank you very much for your time.

Iavg=Vin*ton²*fsw/(2*Lmag)
I_RMS²=ton*fsw*[3*(Ipeak*Lmag)²-3Îpeak*ton*Lmag*Vin+(ton*Vin)²]/(3*Lmag²)

Those are average and RMS current respectively for the primary diodes.

Many thanks for your quick reply! I must admit that I have already got those formulas but (as I couldn't easily measure the magnetizing inductance) I was asking about a practical approach, like magnetizing current being 10% of the primary current or something. Aren't they related?

I could calculate (not measure) the primary inductance though. Is there a rule of thumb to approximate the magnetizing inductance in this case? Btw, the transformer will have a single secondary winding (12V). So, any practical suggestion?

like magnetizing current being 10% of the primary current or something. Aren't they related?

Click to expand...

Yes they are, Primary current=Magnetizing current+Load current*(N2/N1). Once you find out the peak current of the magnetizing inductance, you can estimate easily the magnetizing inductance itself.
The magnetizing current depends on the construction of the xformer, even though there might be some rule of thumb, why would it matter since it varies from a construction to another, in other words, it will not apply to your specific construction with a specific frequency.

Is there a rule of thumb to approximate the magnetizing inductance in this case?

Click to expand...

Here is a better rule of thumb: calculate the magnetizing inductance. That is the best approximation.

Generally the peak current in the diodes is equal to the peak mosfet current at turn off, which lasts for perhaps a 100 nS, so the average current / power in the diodes is very small - as the magnetising current is generally < 5% of the max load current.

For operation above 50kHz, ultra fast and soft recovery diodes are recommended, to reduce losses in the diode and mosfet and for RFI reasons, small snubbers a good idea too.

It pays to have some pcb area for heatsinking for SMT parts (diodes) above that of the bare pad sizing ...

- - - Updated - - -

A realistic calculation of the Lmag should be good enough as long as, for ferrite you do not exceed 180mT Bmax when the fets are on.

Easy peasy said:

Generally the peak current in the diodes is equal to the peak mosfet current at turn off, which lasts for perhaps a 100 nS, so the average current / power in the diodes is very small - as the magnetising current is generally < 5% of the max load current.

Click to expand...

Now that's the kind of answer I was looking for!.. much appreciated!

Easy peasy said:

For operation above 50kHz, ultra fast and soft recovery diodes are recommended, to reduce losses in the diode and mosfet and for RFI reasons, small snubbers a good idea too.

It pays to have some pcb area for heatsinking for SMT parts (diodes) above that of the bare pad sizing ...

Click to expand...

That's what I was actually wanted to know.. if I'll have to go with a TO220 package (and add a small heatsink) or a bare DO-201 package will do the job (I'm using through-hole components only).

I need this converter to be very reliable (24/7) thus I'll mainly go with oversized parts. Btw, should I go with ultrafast & soft recovery diodes or with schottky ones? That recovery stage is quite important, as it could add more stress (voltage spikes) to the Mosfet switches, before the actual clamping process became effective.

true schottkies (silicon) only go up to 45V, MBR10100 is close but not quite ...