Calculate and design toroid transformer

[Jackson75]

Hello! A newbie at this forum.
Could anyone give me some basic guidelines or step by step info the best way to design a toroid transformer for use in car audio SMPS, push-pull design.
I've earlier designed some flyback powersupplies, half and fullbridge designs so I'm not new to the electronics.
But now it's time to do some car audio!

Also, I like to know methods of avoiding flux walking. What kind of ferrite material which suits best for the transformer.

Thanks for now!

[Perry Babin]

How much output power do you need?

I don't think flux walking is a significant problem for power supplies used in car audio amplifiers.

I don't know what's available in Sweden. I'd use either F or P (preferred) material if it's available. 77 material may also be suitable but I've never used it.

http://www.mag-inc.com/pdf/Ferrites/...aterials_P.pdf

[Jackson75]

As usual I'm heading for power in the kW range.
Let's start with a 1kW supply, generating about +/- 80V peak rectified secondary voltages.
I know voltage regulation and loop compensation in forward converters, but I try not use any feedback at all for the car audio design, at the moment. We'll see how much it drops, and I'll eventually be forced to add voltage regulation to the design, later.
Earlier, my transformers were made of ferroxcube 3F3, 3C85 materials. In bridge designs I head for a Bmax of 250mT.
Maybe it's no big difference to design a good push-pull transformer for use at 50kHz and 14V?

[Perry Babin]

You may find it more useful to post this in the following forum. There are a lot of members there that are more familiar with the components available in the EU.

http://www.diyaudio.com/forums/forum...?s=&forumid=67

[Eva]

3F3 and 3C85 are perfectly ok. 250mT may be a bit on the high side from the core loss point of view. Remember that an unregulated push-pull is always operating at maximum B during high line conditions, while a regulated one operates at more or less constant B.

Push-pulls tend to be self balancing as long as fast MOSFET turn-off, drain RC snubbers and some dead time are used. It works in such a way that in case one leg has higher magnetizing current, its drain voltage rises faster thus reducing the volts-seconds product in that direction and compensating for the imbalance. This only happens after drain voltage is free to rise (gate already discharged).

Resonant operation is desirable by providing optimum dead time to allow for zero Vds turn on, otherwise high EMI results. Parasitic capacitance or too heavy RC snubbers should be avoided as they slow down the crossover process thus increasing dead time requirements.

Leakage inductance between both primaries is critical, ideally they should be wound bifilar or in pairs of wires, (AB AB AB), as this results in the lowest EMI and MOSFET turn-off losses. Low leakage inductance between primaries and secondaries also helps.