Switching Regulator Help: Just for the hardcores and Gurus

I have to design an off-line switching regulator with the following specs:

Environment: Class C
Input transient protection: for class C environment
Vac input: 90-245V
Vout: +55Vdc +/-1V from no load to full load
Noise: 0.5Vpp at full load
Current: 35A maximum (100%)
Switching Frequency: 25KHz (?)
Efficiency Expected: better than 70%
Conducted-mode noise: better than +80dBu

In general, my choices are:

1- To comply with the conducted-mode noise: A 7 sections lowpass filter at both AC lines. The coils wound on pot-cores and ceramic caps to ground

2- For input transient protection: Gas-filled spark gaps on both AC power lines and AC quick circuit breakers

3- Topology: Full-bridge oscillator tied to a planar non-resonant transformer.


I am looking forward:
1- To have an analysis on the specs above to know if something important was missed.

2- To know if my choice on the oscillator topology and transformer is a good one. Suggestion on alternative topologies

3- To have suggestions as to PNs for switching elements, class of capacitors for the output filter, ferrites, etc...

4- To have some hints regarding constructive aspects.

5- Literature and design/simulation software.



I deeply appreciate any help.

NandoPG

Re: Switching Regulator Help: Just for the hardcores and Gu

For such a high-power power supply I would have expected a PFC requirement. In fact, the input range and the 70% efficiency suggest a PFC is expected.
The switching frequency seems awfully low, perhaps it's a typo.
Also, I am not too sure about the 80dBuV conducted emmissions limit. I think it should be 60dBuV.

I do not see a load transient response requirement.

The topology is fine for this power level, though you may also consider a phase-shifted ZVT full-bridge.

A good book would be Abraham I. Pressmann's "Switching power supply design".

Re: Switching Regulator Help: Just for the hardcores and Gu

Would you have the PFC standard described in a pdf document that you could upload here?

The switching frequency was choosen so low to make possible an additional noise reduction set by linear regulators present in each major sub-section of the equipment this power supply will feed. Would that make sense?

With 25KHz the conducted-mode noise is a bit worse, being actually 80dBu

Load transient response requirement is really missing, thanks for that..

What would I gain using a ZVT type of full-bridge?

Thank you very much for your post,

NandoPG

Re: Switching Regulator Help: Just for the hardcores and Gu

I don't have a document about power line harmonics, but please go to the Texas Instruments website and look for appnotes on PFC. A relatively popular chip is the UC3854, also from TI. There are newer ones too, such as UC3855. Just look for their datasheets and appnotes.

The ZVT will reduce dissipation in the bridge transistors, since they switch at zero voltage. A chip that can do that is the UC3875, from TI. There are newer ones, too. Again, search for ZVT or appnotes on UC3875. There are issues associated with the ZVT at low load currents, but a compromise can be made.

I have to say that I question the 25kHz switching frequency, especially at this power level, since the magnetics will be huge. As for the noise, it can be cleaned up using some additional small LC filters before the linear regulators, if that is really needed. I think you should seriously consider a higher switching frequency.

https://focus.ti.com/lit/an/slua144/slua144.pdf