On 9 Oct 2004 15:26:05 GMT, B&D wrote:

On 10/8/04 11:20 AM, in article , "Chung"
wrote:

The listener listens to the amp, *NOT* the
amp with the feedback loop open. You can have amps with vastly different
open loop responses, but if the closed-loop response differences are
beneath audible thresholds, then the amps will sound the same. (For
example, there are dozens of op amps that can be used in the signal
path, all with different open-loop responses. But many of them will
sound identical if used properly.)

That isn't true at all.

Yes it is, and proveably so. Do you have any evidence beyond mere
assertion, for your opinion?

The open loop response of an amplifier is critical to the performance of the
entire amp - and its response does influence the sound of transients. As
someone who designs amplifiers in general, I have seen it and I am currently
working on a upgrade to an existing design where the performance is to
change the feedback to allow faster rise times and load changes in a
reliable manner.

As someone else who has been designing amplifiers - for about thirty
years now - I agree that it is *preferable* to start with a reasonably
linear open loop response. However, it is *not* essential, for sonic
transparency. BTW, how do you propose to change the *feedback* to
improve transient performance?

One "listens" to am amplifier in its entirely - both the open and closed
loop response.

A well-designed amplifier will not allow the feeback loop to lose
control. Hence, you will *only* hear the closed loop response.

A previous poster mentioned a source slew rate - and that is
important to keeping the amp under control - and I did mention "sufficiently
slow loop" will be audible - and I suspect most amps do this.

It is sufficient to place a filter at the input, which will not allow
slew rate limiting at full output. Crude, but effective. Your
'suspicion' is irrelevant, unless you can *demonstrate* the existence
of this effect in a real-world amplifier.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

On 9 Oct 2004 15:26:51 GMT, B&D wrote:

On 10/8/04 11:20 AM, in article , "Chung"
wrote:

Here's where you are wrong. The listener listens to the amp, *NOT* the
amp with the feedback loop open. You can have amps with vastly different
open loop responses, but if the closed-loop response differences are
beneath audible thresholds, then the amps will sound the same. (For
example, there are dozens of op amps that can be used in the signal
path, all with different open-loop responses. But many of them will
sound identical if used properly.)

Put a capacitor as the feedback on the op amp - and put in a rising edge of
a square wave and note the transient and long term response. Do it with a
rising edge of various input signals.

Why? What has this to do with the closed loop response of a linear
audio amplifier?

Do it with a variety of capacitors and rising times and you will see that
the initial response and final response are different. THIS is a very
simplified example of what I am talking about (though not 100% accurate -
just to illustrate my point).

Sure it's not 100% accurate, because you are describing an integrator,
not an amplifier.

Have you ever used op amps?

Lots - and no one in history has ever attempted to use an integrator
as a linear amplifier.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

On 9 Oct 2004 15:31:52 GMT, B&D wrote:

On 10/8/04 1:08 AM, in article ,
" wrote:

The listener will hear both the open loop and closed loop response. In the
steady state condition you will hear the closed loop response. In
transients the sound will be influenced by the open loop response.

Sure, if you're referring to obsolete dreck designs like a 741 op amp and
the like.

If one can hear this going on then that amplifier cannot be said to be
'nominally competent'. What news!

Why worry about it these days?

I am not particularly worried about it myself, but when you design an
amplifier using global feedback, you do need to take this into account - not
to mention that a lot of feedback can make an amplifier less stable as well.

*Any* amount of negative feedback makes the amplifier less stable -
but it's not usually a problem.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

On 9 Oct 2004 15:31:04 GMT, B&D wrote:

On 10/8/04 1:11 AM, in article , "Stewart
Pinkerton" wrote:

On 7 Oct 2004 03:13:59 GMT, B&D wrote:

The listener will hear both the open loop and closed loop response. In the
steady state condition you will hear the closed loop response. In
transients the sound will be influenced by the open loop response.

That is not necessarily true, and is certainly *not* true of a
well-designed amplifier which absorbs transients at its input.

Transients in this case would be something like a plucked string of a
guitar, etc.

In which case they will be accurately amplified by any competent audio
amplifier, since the bandwidth of the highest frequency component of
such 'transients' does not exceed 5kHz, let alone 20kHz.

But your point is true - to use feedback properly, you MUST take slew rate
into account and make sure the input cannot exceed the speed of the feedback
and delay - probably by a factor of 10 or more to make the sound less or
inaudible.

A factor of three is more than adequate.

Absorbing the leading edge of the transients through filtration or other
means, you have to be careful as well since you are removing and altering
the signal in a benign manner.

No, you are simply limiting the bandwidth, in the case of an audio
amplifier, to something like 30-40kHz. Any competently designed modern
SS amplifier will have no problem exceeding this bandwidth without
encountering slew rate limiting, and a well-designed one will have an
*open loop* bandwidth greater than this.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

"Put a capacitor as the feedback on the op amp - and put in a rising edge
of
a square wave and note the transient and long term response. Do it with a
rising edge of various input signals.

Do it with a variety of capacitors and rising times and you will see that
the initial response and final response are different. THIS is a very
simplified example of what I am talking about (though not 100% accurate -
just to illustrate my point)."

This is testable as to the threshold that is audible, do the tests. If a
threshold can be exceed then it can be audible. As is also other factors
in amps,ex. distortion, freq response differences, etc.; all of which
illustrate the same thing. The core point in all of this is to
demonstrate that in amps today these can be heard as difference, including
your's and other proposed design differences that make the $800 and the
$8000 amp clearly different using listening alone testing.

On 10/10/04 12:03 AM, in article , "Stewart
Pinkerton" wrote:

But your point is true - to use feedback properly, you MUST take slew rate
into account and make sure the input cannot exceed the speed of the feedback
and delay - probably by a factor of 10 or more to make the sound less or
inaudible.

A factor of three is more than adequate.

Wow - you wouldn't get a job at our company! :-) We tend to make it about
10x to guarantee control under all circumstances, I am not sure I would make
that under the umbrella ('competent') we have been using if it were only 3x
the time. Given the nature of analog circuitry and the SOTA in power
transistors, feedback should easily be able to be accomplished much faster
than that.

Absorbing the leading edge of the transients through filtration or other
means, you have to be careful as well since you are removing and altering
the signal in a benign manner.

No, you are simply limiting the bandwidth, in the case of an audio
amplifier, to something like 30-40kHz. Any competently designed modern
SS amplifier will have no problem exceeding this bandwidth without
encountering slew rate limiting, and a well-designed one will have an
*open loop* bandwidth greater than this.

I would agree that a 'competantly' designed amp iof we choose to define it
as such. I am not sure what one of those beasts would be - we have used
this term quite a bit - and I am curious as to the specifications and design
goals of such a beast?

Now, I do know that most amps ought to be able to do as you describe, but
what if 50-80% of amplifiers would have limiting to 22.5kHz - would they be
competent?

Might be a good thing for us to define, even somewhat broadly, the
characteristics of one of these 'Platonic ideal' competent amplifiers.

I will start:

1. Absolutely stable into all loads, 0 Ohms to open circuit, from +90 to
-90 phase angles.

2. Rated power is at full power, full bandwidth (band limited white noise)
into 8, 6 and 4 ohms

On 10/10/04 12:02 AM, in article , "Stewart
Pinkerton" wrote:

On 9 Oct 2004 15:26:51 GMT, B&D wrote:

On 10/8/04 11:20 AM, in article , "Chung"
wrote:

Here's where you are wrong. The listener listens to the amp, *NOT* the
amp with the feedback loop open. You can have amps with vastly different
open loop responses, but if the closed-loop response differences are
beneath audible thresholds, then the amps will sound the same. (For
example, there are dozens of op amps that can be used in the signal
path, all with different open-loop responses. But many of them will
sound identical if used properly.)

Put a capacitor as the feedback on the op amp - and put in a rising edge of
a square wave and note the transient and long term response. Do it with a
rising edge of various input signals.

Why? What has this to do with the closed loop response of a linear
audio amplifier?

Simple, this is showing the difference between a transient response and a
steady state response can both affect the output of an amplifier.

I wasn't trying to do more than that.

On 10/10/04 12:01 AM, in article , "Stewart
Pinkerton" wrote:

The open loop response of an amplifier is critical to the performance of the
entire amp - and its response does influence the sound of transients. As
someone who designs amplifiers in general, I have seen it and I am currently
working on a upgrade to an existing design where the performance is to
change the feedback to allow faster rise times and load changes in a
reliable manner.

As someone else who has been designing amplifiers - for about thirty
years now - I agree that it is *preferable* to start with a reasonably
linear open loop response. However, it is *not* essential, for sonic
transparency. BTW, how do you propose to change the *feedback* to
improve transient performance?

In the case at work we have nested an analog loop inside of the slower
digital loop to allow for a fast attack response - to supplement the much
slower digital loop that controlled this piece of equipment.

In an audio amp, a feedback is probably best done in an analog manner,
though the sample rates of the digital processors is making digital feedback
more of a possibility.

"I will start:

1. Absolutely stable into all loads, 0 Ohms to open circuit, from +90 to
-90 phase angles.

2. Rated power is at full power, full bandwidth (band limited white
noise)
into 8, 6 and 4 ohms"

All you are doing above is tinkering with the usual "within current
limits" guidelines for comparing amps,ie. if given an untypical load then
all amps have such limitations as above and might distort so as to rise
above the threshold of audibility. In addition, any amp can be modified
to do as above with changes in power supply and number of output devices.
This still does the amps sound differently folk no good, if the $800 amp
can be changed to have the same current limiting as the $8000 amp any
differences as above will disappear, regardless the majic claimed for the
latter amp. Some in the latter category will sound different because each
has a different current limit if pushed far enough, by which we have
demonstrated the obvious. When the marketing/publishing folk claim night
and day differences they are not making appeals to differences to current
limiting but rather some majic the designer put into them.

(Lasse) wrote in message ...

Haven't you read the rules of the Richard Clarks Amp Challenge? In that
case you should. [rules snipped]

I had not, thank you for posting them. They do make sense; if those
are the conditions that are *implicit* in anyone else's claim that
"all power amps are sonically indistinguishable" it's uch easier for
me to see what they're really claimimg.

It does make me wonder what amps (specifically) have not meet Mr.
Clark's conditions!

B&D wrote:
On 10/8/04 11:20 AM, in article , "Chung"
wrote:

Here's where you are wrong. The listener listens to the amp, *NOT* the
amp with the feedback loop open. You can have amps with vastly different
open loop responses, but if the closed-loop response differences are
beneath audible thresholds, then the amps will sound the same. (For
example, there are dozens of op amps that can be used in the signal
path, all with different open-loop responses. But many of them will
sound identical if used properly.)

Put a capacitor as the feedback on the op amp - and put in a rising edge of
a square wave and note the transient and long term response. Do it with a
rising edge of various input signals.

"Put in a rising edge of a square wave"? What exactly are you talking
about? Are you trying to ac couple in just the edge? Putting a capacitor
in the feedback loop would appear that you want to build an integrator.
What is the relevance of that in this discussion? That one could design
a circuit with a non-flat freqeuncy response?



Do it with a variety of capacitors and rising times and you will see that
the initial response and final response are different.

Now, are you talking about a differentiator or an integrator? An
integrator will turn a square wave into a rtiangular wave. As usual, I
find it hard to understand what you are saying...

THIS is a very
simplified example of what I am talking about (though not 100% accurate -
just to illustrate my point).

How does that have anything to do with what I said, which is that
different op amps with vastly different open-loop responses can sound
identical in audio applications? Do you agree with that I said? What was
your point that you tried to illustrate?


Have you ever used op amps?

Since the early '70's. Not to mention designing them in IC's. Of course,
what you wrote so far raised the question of whether you really
understand how to use op amps or not...