John Rumm wrote:
On 05/03/2011 16:28, The Natural Philosopher wrote:
John Rumm wrote:
On 04/03/2011 08:55, sm_jamieson wrote:

There are some caps (more expensive) marked "for audio" etc.
Is this just marketing, or is there anything different ?
Surely in the power supply they cannot make much difference ?
I am sceptical myself.

When you think that an amplifier is in its simplest analysis a voltage
(or current) modulated variable output power supply, you appreciate
the PSU is a fairly critical part that is actually in circuit with the
load.

Fairly NON critical part actually. Since the feedback that controls the
output is taken from across the load, not the power supply.

While true, the key point remains; you are better off not allowing
excess PSU generated noise through to the output in the first place than
you are attempting to compensate for it later.




Smplistic exaplanatin: A voltage regulator is an audio amplifier
configured to have zero AC input and a precise DC output :-)

I.e it makes little difference where you put your supply ripple
rejection. Its the same amount of componentry.

It s not hard to ensure that the supply rail rejection of an audio
amplifier even before feedback is applied is well up in the 80-100dB
range: Adding feedback takes it to the completely inaudible.

To think otherwise is to be beguiled by HiFi Bull**** merchants.

Remember, when te music stops, so does the current drain on a class B
amp. There is bugger all ripple to break through.

Under power, the music is so damned loud you wont hear any hum
breakthrough that's 100dB down on the signal anyway.

The only time you hear it is if its faulty, or you are clipping to it.

Or in terms of odd harmonic distortion if there is resistance in it and
your supply rejection is crap. But as I said, supply rejection is one of
the things that is relatively easy to get very very good . Constant
current balanced pair phase splitters.. even cascaded transistors remove
miller effects from the high voltage swing transistors..Christ even the
'all on a chip' power amps have all that.

The proof of the pudding is in the final distortion products you achieve
at highish powers. .