Do high current rail splitters exist?

[e100]

Low distortion 10-100 w class AB audio amp modules often require a split rail power supply.
Thinking aloud, would it be possible to use a single supply class D audio amp to generate a high current virtual ground to satisfy the power requirements of a class AB split rail amp? Would the RF noise leak into the class AB amp making it as bad as a class D amp?

[dmills]

Seems like a hard work alternative to a 10kuF cap in series with the speaker (The usual way to skin a single supply amp if you do not want to bridge the load, you usually want to bridge the load)!

A CCM buck converter running 50% duty cycle would also probably get it done, and has the virtue of simplicity (CCM so if is effectively two quadrant).

Regards, Dan.

[e100]

You got me thinking, wouldn't it make sense if all class AB amps operated in bridge mode by default?
Lower voltage power supply = cheaper components, or am I missing something?

[dmills]

Lower effective load impedance, so the SOA issues are about the same, but now if you want to support a 4 ohm load you need to be able to drive into 2 ohms, and all the non ideal behaviours become a much bigger issue.

Also twice the component count....

Regards, Dan.

[C]

Think of problem areas and what is effected.

A differential audio input is great as it removes a lot of noise.
But you have to remember that with only two connections you have equal of a PWM input if it is not a balanced signal input. A DC offset can build over time. So you need a 0 signal input to counter the building DC offset. Most times it has the bad name of ground.

With a single output you are putting un-equal loads on each power rail and counting on a return path to balance the load on other rail.

With differential output you are putting an equal load on each rail but need 2X circuit. By knowing that rail load is equal the circuit can better adjust to rail changes. You can save of power supplies.
The PWM like DC offset of input can be easily removed using the zero signal input.

The thing often forgotten is that with differential output you are not trying to make a positive output side match a negative output side with each built from different transistor types ( NPN. PNP). But are trying to make two NPN based sides match & two PNP sides match. A positive/negative side miss match becomes a common mode signal to output which is ignored.