[andyr]

I was wondering which was the worser of two possibilities and what are the implications of each situation ... and whether these are realistic scenarios!!  ) :

1.  An output stage of a power amp can deliver a current of X amps (instantaneous or steady state, whichever you prefer to talk in terms of) but the power supply can only deliver < X amps;
OR
2.  The power supply can deliver Y amps but the output stage is only capable of delivering < Y amps.

IE. in scenario 1, the output of the amp is power supply limited; in scenario 2, the amp's output is output stage limited.

Are these two scenarios realistic?  Which is the better situation?  

Regards,

Andy

[_scotty_]

Power supply limiting would be the prefered scenario as the amp would only be clipping the waveform under this condition. Exceeding the maximum current capability of the output devices will likely result in their immediate destruction.
Scotty

[JoshK]

I'll hazard a guess that most amps' driver/input stage can't drive the output stage to destruction regardless of whether the PSU can.  However, i beleive it is more common in most designs for the output stage to clip because of current limitations (PSU). I could be wrong though.

[Dan Banquer]

"Power supply limiting would be the prefered scenario as the amp would only be clipping the waveform under this condition. Exceeding the maximum current capability of the output devices will likely result in their immediate destruction."
Scotty
I'll agree with that, except I would say maximum power capability.
                d.b.

[andyr]

Power supply limiting would be the prefered scenario as the amp would only be clipping the waveform under this condition. Exceeding the maximum current capability of the output devices will likely result in their immediate destruction.
Scotty

Hi Scotty,

I think you are being a bit simplistic here.  Obviously, if you exceed the maximum current capability of the output devices, they will self-destruct.  However, if the DC voltage rails in the amplifier are set so that, with an X ohm speaker, not more than the rated output device current can be delivered, the output devices will not destruct.

Of course, if you use this amp to drive an X/2 ohm speaker, it will have to output more current and so may be driven to destruction!  

Lets do some maths (the info for which has come from a recent post on AA's MUG Forum which referenced an excellent Audioholics article by Dan Banquer entitled "The High Instantaneous Current Spec" - which I hadn't read when I originally posted):

* say the speaker has a uniform impedance of 3 ohms (I know most speakers do not exhibit a uniform impedance curve but Maggies nearly do - and they're almost entirely a resistive load, with little inductance).

* an amp capable of delivering 300w into a 3 ohm speaker will have an output current of 10 ohms (P = I^2 x R).

* the output voltage will be 30v (P = V^2 / R).

If the amp had a PS capable of delivering 20 amps (lets ignore electrical losses, just to make it easy!    ) then the "extra" PS capability will never be used.

IE. the output of the amp is limited by the DC volts the power rails run at ... not by the current capacity of the PS (ie. the rectifiers or mains transformer).  This scenario is my example #2 and I would suggest this is the most common situation for good power amplifiers?  

If you turn up the pre-amp volume control in this scenario then, yes, the power amp will clip as it is trying to output more voltage than the DC rails have available.

I guess my question is ... will my scenario #1 - where the output of the amp is PS current limited, because the power supply is under-specced -  result in the amplifier clipping?  It is not trying to produce more output volts than the DC rails allow it to ... but it is being "starved" of current.

So it certainly won't sound good but maybe it'll sound better coz it's not clipping?

Regards,

Andy

[_scotty_]

My suggestion would be to buy a big enough amplifier to drive the partnering loudspeaker without this question ever arising.
Or to put it another way, you ask a simplistic question you get a simplistic answer.
Scotty

[art]

You seem to be forgetting one thing: the filter caps are responsible for most of the peak output. For short transisents, you can supply the peak demands with an undersized supply.

Pat

[andyr]

You seem to be forgetting one thing: the filter caps are responsible for most of the peak output. For short transisents, you can supply the peak demands with an undersized supply.

Pat

Thanks, Pat ... I had forgotten that simple fact.  So let's follow the "argument" through ...

Ok, yes, transients can be quite adequately "fed" by the filter caps - so the fact that the PS might be undersized doesn't affect anything.  However, the amp will clip if it's trying to output a voltage which is greater than that "allowed" by the DC rails.

With slightly longer transients and an undersized PS, possibly the filter caps don't recharge as fast as they need to, to deliver good sonics?  So this is the case of an amp being PS-limited.

One result of this is that the DC rails dip slightly ... making the amp more likely to clip?     But it's the DC rail voltage which causes clipping ... but if the PS can't drive this back up to the designed voltage quickly enough, under heavy load, this makes the amp sensitive to clipping?

So a PS which is specced properly is a good thing ... and I suppose the way you determine whether the PS is a good one is whether the amp's output is rated X watts into 8 ohms and 2X watts into 4 ohms ... and preferably 4X watts into 2 ohms?

Regards,

Andy