SMPS transformer questions

[AW-tech]

I am trying to design a switching supply for a bass amplifier (120W RMS @ 4 ohms). When calculating the power supply requirements, do I use RMS or peak current?
I've been using the following site: Halbbrückengegentaktwandler
I am wondering if anyone has any experience with this site. It seemed to recommend some pretty large wire cross sections (a = 3.17mm squ. for 9.5A), and I wonder if I could get by with less.

I read somewhere here that it would be good to use a center tapped secondary to reduce diode losses. Would the current requirements of each of these secondary halves be less than if a single secondary were used? When winding a center tapped secondary, do you wind one half, then the other half on top of that? Or do you wind them together (bifilar style)?

What kind of tape do you use and where does one get it at a reasonable price in the US?

[Sch3mat1c]

For 10A I would go with around 14AWG, or 1.6mm. Figure 2000A/in^2 (or around 3.1A/mm^2). More for smaller transformers and cheaper construction (i.e., use a fan!), less for bigger transformers and cooler operation. The program may be basing it on a much lower current density, or possibly taking into account proximity effect (although this is generally a consideration only for multilayer windings, which requires assuming a winding structure, something the program doesn't go into).

Multistrand construction or litz wire is something else to consider.

A FWCT rectifier is handy for low voltages, typically 12V and less. Diode loss is less important for higher voltages, where you can use a FWB rectifier instead, saving on turns ratio.

Because each half of the winding conducts only half the time, the RMS rating goes up slightly, so you need slightly larger wire than expected (i.e., slightly more than half the cross sectional area of a FWB rated winding).

You can wind them however you want; bifilar reduces voltage capacity somewhat, dramatically increases parasitic capacitance (potentially increasing switching loss and ringing), while dramatically reducing leakage inductance between the ends (which might make the rectifier work better, easier to snub I suppose). What's more important is leakage to the primary, which is often wound in two sections (on a solenoidal bobbin, for E or C type cores), or completely covering a toroid.

Tim