On Tue, 15 Jul 2003 08:42:20 +0100, Jim Lesurf
wrote:

In article , Don Pearce
wrote:
On Mon, 14 Jul 2003 09:12:51 +0100, Jim Lesurf
wrote:

Do people still use electrolytics as signal path coupling caps these
days in 'audiophile' equipment? I know that other types show charge
storage effects, but this should, I hope, not be audible with decent
designs handling musical signal patterns.

Plot the AC signal path of an amplifier's output stage and you will find
that the supply decoupling electroloytics are in series with the speaker
- they function as speaker coupling caps. I am not aware of any design
that has done away with this topology yet.

I don't really agree that the above is an adequate description of what is
happening, since it is only a part of the story in this context. Indeed, I
tend to by slightly wary of terms like "signal path" even though I have to
use it at times as my impression is that it is used in a misleading way by
reviewers at times. I'll try to outline why I feel this...

Consider three uses of caps (nominally electrolytic).

1) As a d.c. break (decoupling capacitor) between amplifier stages with no
feedback from one stage to the other.

2) As a d.c. isolation at the 'foot' of a resistive feedback network to
ensure the amplifier gain falls to unit or zero at low frequency.

3) As reservoir caps in a power amp.

In conventional audio equipment the signal pattern is conveyed as a
*voltage* pattern, with whatever current may be required to assert this
voltage being provided. In such circumstances I tend to apply the following
'test' for deciding if something is "on the signal path".

I consider adding a fictitious voltage generator in series with the
relevant component to add a small 'error signal' (be it noise, distortion,
or whatever). I then ask if this then simply adds to the intended signal
with the rest of the system having no way to identify and reject this
addition.

In (1) and (2) any such error signal become part of the signal, and are not
rejected or ignored by the amplifier system. Hence any injected noise or
distortion at such component locations can be regarded as having been
injected into the signal path.

However with (3) the amplifier should have been designed to reject or
ignore any voltage fluctuations on the power rail voltages. My experience
is that well-designed amps do this fairly well. Hence case (3) seems to me
to be different to (1) and (2) *unless* the amplifier design has
effectively no ability to reject fluctuations on the rails due to
imperfections of limitations of the PSU (reservoirs, etc).

I certainly agree that the charge which passes through the loudspeaker was
previously stored in the reservoir capacitors. However I have my doubts
that this means we should regard the reservoir caps as part of the signal
path due to the behaviour of the amplifier in rejecting these fluctuations.
This may not be perfect, but is somewhat different to (1) and (2) were no
rejection may take place.

Makers and reviewers often mention that a given amp uses a side-chain of
some kind for d.c. offset removal, hence moving any caps 'out of the signal
path'. I have my doubts about this since, by the above argument, any error
voltages produced by the cap may still end up being added to the signal.
Hence I am not currently convinced these methods are all 'better' than the
old fashioned d.c. break cap at the foot of a resistive feedback network.
:-)

Slainte,

Jim

I appreciate all this reasoning, Jim, but the way I see it is this:

ALL of the AC signal current that goes through the speakers must pass
through these electrolytics as series elements. That puts them in the
signal path. Whether they are at the "cold" or "hot" end of the
speaker is really neither here nor there.

Actually my use of the word "ALL" above may seem contentious, but any
current that doesn't go through the caps must go instead through the
rest of the power supply - including the diodes - where it will be
chopped and modulated by the mains. Another good reason for big, fat,
low ESR electrolytics.

Just another view.

d

_____________________________

http://www.pearce.uk.com