Well shoot, guess I'll change my plans. I'll go with a linear regulator for now, deal with the wasted power & slightly lower maximum output power, and have the switching supply as a project to work on where it's not holding up the subwoofer project if it takes a few revisions.

 

My assertion was that switching supplies ARE more of a challenge and that noise MIGHT be an issue. A well designed amplifier with a well designed power supply will have no noise power supply problem, but a few switchers are noisy and can cause problems. A noise level that does not bother 5 volt digital logic may be far to noisy for an audio amplifier.

 

The TDA8920 amplifier IC is old and is being discontinued (last time buy says Digikey).
Its datasheet says that a +30V/-30V power supply is the absolute maximum allowed and recommends +27V and -27V.
Then its output is about 130W into 8 ohms with 0.5% distortion.

 

Reducing the voltage from a non-regulated supply without reducing the efficiency is a real challenge that usually requires adding turns to the primary winding. That can be done on a toroid but it is usually quite tedious and inconvenient. So at that point adding a regulator is the other option.

 

The TDA8920 amplifier IC is old and is being discontinued (last time buy says Digikey).
Its datasheet says that a +30V/-30V power supply is the absolute maximum allowed and recommends +27V and -27V.
Then its output is about 130W into 8 ohms with 0.5% distortion.

I was looking at TDA8920C from NXP which states +/-32.5V max and 170W at 0.5% distortion. I saw that it’s EOL, but I’m not concerned about making more than one.

I also thought about paralleling two of them if I wanted higher power, but realistically I probably don’t need it.

 

I'm designing a powered subwoofer for myself and I've been intending to use a toroidal power transformer I have (60Hz mains frequency). My plan is to have the PSU, amplifier, and speaker all in one box.

In theory a toroidal inductor shouldn't have a magnetic field outside of the toroid, but what about in practice? Suddenly I'm worried about the magnetic field of the power transformer affecting the speaker function, especially since it would be right in the middle of the frequency range of use.

While I have not used this book (I have a much older design book, but can't find it right now for ISBN info), this might be a good start for you- there are many issues other than just the speaker:

https://www.amazon.com/Designing-Bu...keywords=speaker+design&qid=1601311797&sr=8-4

 

I should have mentioned it sooner, but it is probably not possible to create a magnetic field extenal to the speaker that is strong enough to have a detectable effect. Consider that in a normal magnetic speaker the voice coil is only working against the magnetic flux concentrated in that air gap. And consider the level of an external magnetic field strong enough to change that intense and highly focused magnetic field. Any external field intense enough to have any effect would be incredible.
So the whole discussion is about a fake problem. I am dismayed that I neglected to consider that earlier.

 

One thing you do not want to do, as I have seen done by some in the past, is to place a metalic shield cover over the toroid, this effectively creates a shorted turn on the transformer.!
Max.

Surely that's only true if the shield is itself of toroidal form? I'm guessing the speaker problem is usually related to the toroidal core's high permeability which is (AC mains) modulated at 100/120 Hz, a phenomenon exploited in fluxgates.