Mark wrote:

The state of the art of signal analysis is advanced to the point where
I do not believe it is possible for a difference in __amplifiers__ to be
audible but not measurable. It may not be obvious at first what
measurement is needed, but once that is unraveled, some measurement
should always be able to explain any audible difference (talking about
amplifiers here, mics and loudspeakers are another matter)

Absolutely, but that's the rub!

If a critical listening test should uncover an audible difference
between two amps, the next step should be to make whatever measurements
are needed to identify the cause of the difference. Then the designers
can use that information to make design changes. Something like, amp A
has a more solid sound compared to amp B and we discovered that the
reason is the difference in damping factor (a trivial example).

This is true, BUT we have folks here in this group who are claiming that
low THD levels obviate the need for any other testing. Hell, we have
one person here claiming that TIMD is total bunk.

Please provide an example of an amplifier performance characteristic
that can produce an audible difference that cannot be measured.

I don't know of any that cannot, but I know of a whole lot that never
will, because of the reluctance on the part of some people to investigate
more modern measurements. I know others that never will, because the
audible differences are significant enough that it's not even worth
bothering with.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

"Kevin Aylward" wrote in message
...

Selling a two channel mic amp for $1500 is snake oil. There is zero
chance that its value for money.

Excuse me, but that's tommyrot. I suggest you spend some time with a
Great River preamp and a cheaper preamp with as close a frequency
response as you can find. Do some listening tests, making sure you
match levels carefully. You will almost certainly hear the
difference.

No chance. Hey, what did your horoscope tell you to do today? How about
the tea leaves?

I don't do astrology. I don't do mysticism. I do, however, pay some
attention to what my ears tell me.

Then use both for ten years, and see which one is broken.
More expensive preamps usually do sound better, and they're also
better built, which is part of what professional gear is about.

Excuse me, this is complete nonsense. Again, if the amp has say, 0.01%
thd and imd, frequency response flat to say, 0.2db and all te usual
stuff like reasonable headroom, drives a load correctly and stable etc,
then the all sound identical. End of story. Its been proven in too many
controlled ab tests. Its just not worth debating, and I'm not going to
any more.

Well, there's a relief. And I hope you enjoy the recordings you make with
your $15 preamps -- hey, it's a free country. Me, I'm working with stuff
that has a little more put into it, both for the sake of improved sound and
improved long-term reliability. I don't go nuts with it, but there's a
threshold below which I haven't heard anything I like yet.

There is a sensible limit. Anything better is worthless. You cant get
any better then the weakest link in the chain. This is the speakers.
To get distortion lower then 1% costs you an arm and a leg. To get
speakers flat to 1db, costs the other arm and leg.

The latter points are quite correct. The former is not. The facts of
information theory are that the total chain is always *worse* than its
weakest link.

In what way does the statement:

"You cant get any better then the weakest link in the chain"

logically contradict your statement?

The statement is correct, but not complete.

Peace,
Paul

wrote:
Perhaps a career in video would suit your talents better. If you can't
hear any of this stuff, maybe you can see some of these things.

I use the Audio Precision analyzer in the design and testing of the
High Speed Mic Preamp as well as mods done to other manufacturers
equipment. Yes, it does show subtle test variations. It also will
show measurements that may be nearly identical, but sound different.

Yeah, like prove it.

It, like any test rig has severe limitations due to the fact it uses
steady state stimulus rather than the violent waveforms of actual
music.

Oh dear...same old 30 year old arguments.

If you can't do gated sine wave test, your equipment aint up to much.


Folks that design gear based on test results will miss the boat every
time.

Complete nonsense.

I think I'll direct you to Doug Self, myself.
http://www.dself.dsl.pipex.com/ampin...o/subjectv.htm

Test gear will tell you if something is wrong, only your ears
will tell you when something is right.

Not as far as basic amplifier design associated with distortion, BW,
noise etc is concerned. Its a done deal.

Yes, once the bugs are worked
out on a design, there comes a point when you turn off the analyzer
and turn on the ears.


More bull****.

Ears are only needed to tell if things like tone controls are set right,
compression ratios are ok, and so forth, as these are subjective. The
design of an audibly straight piece of wire with gain, is a solved
engineering issue, and for the life of me, I see no reason whatsoever to
listen to a power amp or mic amp at any point in its design. Its just a
matter of designing to well understood numbers.

I shall point out, I am a guitarist, have been so for 35 years. I use a
Fender tube twin reverb. It has a distinct sound, even when clean. It
also measures significantly different from a transistor amp. There is
therefore no reason to suggest that they should sound the same. So,
sure, some amps sound different, but for good reason.

Kevin Aylward

http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Lorin David Schultz wrote:
"Kevin Aylward" wrote:

Oh lord, please deliver us from the ignorant.

Hey, I have a few oxygen free cables that I can sell you, only $500
for a pair, but make sure you connect them the correct way from the
speaker to the amp. They are colour coded to make this easier for
you.



Kev, baby, relax -- I'm on your side on the snake-oil issue. It's
just an ENORMOUS leap from "many esoteric audio products are a
rip-off" to "all properly-designed amps sound the same." I couldn't
believe you really meant that, so I figured you must be probing for
reaction.

I am trying to keep the prose down. "properly" is not really an accurate
word. My statment is on the lines of:

Given certain basic specifications are met, all plain power amps sound
the same.

Typically, full power BW say, 50Khz, unconditional stable in all loads
with a phase margin close to say, 90 degrees, zout say 0.05 mohms, and
so forth. The point, being that there are no audible differences that
cant be accounted for by measurements.

Sure, those Crowns that spec 0.5% thd at 1Khz might well be detectable,
but I am not the one to claim that it actually is:-) I give sufficient
conditions, not necessary ones.

Kevin Aylward

http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Scott Dorsey wrote:
Kevin Aylward wrote:

Hey, I have a few oxygen free cables that I can sell you, only $500
for a pair, but make sure you connect them the correct way from the
speaker to the amp. They are colour coded to make this easier for
you.

So you have decided that, since much of the high end world is
completely bunk,

Yep. Its a marketing scam.

that therefore there are no audible differences
between amplifiers with low THD numbers?

Not at all. One amp may have a BW of 1khz, and another with 100Khz.


There is quite a jump between these two statements.
--scott

Given certain basic specifications are met, all plain power amps sound
the same.

Typically, full power BW say, 50Khz, unconditional stable in all loads
with a phase margin close to say, 90 degrees, zout say 0.05 mohms, and
so forth.

Kevin Aylward

http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Mike Rivers wrote:
In article
writes:

The point is, it that it costs only a few bucks in electronic
components to produce 0.002% thd, imd, and all the rest of that
wonderfully useless 1000v/us slew rate performance. Like we want to
listen to FM 101 through our mic amps.

Oh, I guess Kevin is coming from that other audio universe where a
"preamp" is something that goes between an FM tuner, cassette deck, CD
player, or phonograph, and has Volume, Balance, and Tone controls as
well as an input selector.

Are you the only one in the NG so far not to know my universe?

http://www.anasoft.co.uk/band/bio.htm

His comment about $500 cables was the
tip-off. That is indeed an industry with a segment that runs on snake
oil.

Jim Williams' work with professional recording and production gear is
a different world

Not at all. Its same ****, different suckers.

where people actually do perceive actual sonic value
for money, they don't just think it sound better because it costs
more.

Of course they do. Its a well known psychological effect.

Kevin Aylward

http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Mike Rivers wrote:
In article
writes:

Yes, but like anyone who posts, its always for a purpose. He expects
a return. No one does anything that they know will be net detriment
to themselves. Its basic evolution.

You read this. Send me money.

I don't get much return on the investment in time that I make here,

Its a statistical thing. What's the loss verses the gain. You arnt
losing anything much by posting here, if you were, you wouldn't be
posting. Its that simple.

I
just hope to get famous some day. At least as famous as you anyway. Or
maybe I already am.

The problem with specs is that often the spec are wrong or
misleading. The specs tell the story if they are correct and fully
understood.

Jim williams (among others) resolves this problem by actually
listening to equipment after he modifies it. Chances are his customers
will find that it sounds better to them, too. If not, I guess he
wouldn't be able to stay in that business.

Nonsense. That's not how the real world works. Billions pray and give
money to churches, and they all get f'all back.


Selling a two channel mic amp for $1500 is snake oil. There is zero
chance that its value for money.

Money is money, but value is relative. It may not be worth $1500 to
you, but many people spend more than that for a mic preamp and are
really happy that it gives them better results than the preamp in the
console that they were using previously. Ask him this - how many have
been returned for a refund?

Yeah, like somone has just been suckered for a few big ones, and they
want to tell everyone about what a big mug they are?

Look, dudes this all well understood 101 psychology. Peoples perceptions
are based on their belief of the facts, not the facts.

If the only benefit is that they belief that they get "better" results,
when in fact, they don't, its a big a scam as horoscopes.

Kevin Aylward

http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Scott Dorsey wrote:
Kevin Aylward wrote:
Scott Dorsey wrote:
Kevin Aylward wrote:
Scott Dorsey wrote:
input transistors. That's, dare I say, just plain daft, and don't
work.

Its like nails scratching nails down a blackboard to us analogue
designers. There are so many reasons not to try this. Indeed, in my
current designs, there arnt any, its all current sources.

(Note that by input I was referring
to the negative-going input on the differential pair that is used
for feedback, but of course if you offset one side of the
differential pair, you offset both of them)

You also offset the output, which don't do very nice things to the
speaker.

No. There is a capacitor in the way.
That's the point. You want
that offset on the capacitor.

One doesn't use output capacitors for power amps nowadays. Amps are
DC coupled. Don't know the last time I saw a single supply amp.
Probable in Dr. Who's TARDIS.

Are you listening to anything I am saying?

Yep, and it makes little sense.

I am not talking about an
output capacitor, I am talking about a capacitor in the feedback loop
between the output and the inverting side of the differential input
pair.

You are obviously not listening to what I am saying. If there is an
offset on the input, there will be an offset on the output, as the amp
is DC coupled throughout. To prevent this output offset doing nasty
things to the speaker, one would then need an output cap.

I hope you mean the cap that is in the feedback network. It is not in in
the loop, i.e. connected from output to input. The cap goes from the
negative input to an ac ground via a resister.


Let's go through this again:

1. If there is no offset, you don't need a capacitor

2. If the offset is larger than the signal, you use a capacitor.

3. If the offset is smaller than the signal, make it larger.

This is not anything particularly difficult here.

Apparently it is for you, as what you say here doesn't make any sense.
What are you trying to say here? Draw your circuit.

Where is your offset? Where is your capacitor?

Email me a SuperSpice schematic file.

Read the Sanyo
capacitor handbook for a description of the whole thing. Read Doug
Self's series on amp design.

Sure, Self knows his stuff, however I am already an expert in amp
design, I have been so for over 20 years, and have no need for any
instruction in such matters.

What part of "I designed the Studiomaster MOSFET 1000 20+ years ago" did
you miss?


However, the design is easily understood from inspection by anyone
that understands circuit design. There is no rocket science here.
All the blocks are standard stuff. The only bit that might need
explaining is the somewhat novel second feedback loop on the output
transistors on one of the circuits. Oh..and that some of the diodes
are zenors for simulation purposes.

Yes, but not being able to read the numbers makes it problematic for
use as an example circuit to point out DC setpoints.

I dont really understand what you are trying to say here. DC setpoints
are trivial to determine by inspection with basic design knowledge. For
example, the voltage at the collectors of the input stage npn
transistors are forced to 4 Vbes (+~1V volt resister drop) by the common
mode feedback. Some cascode voltages are not important, within reason
they can be anything. The second stage current mirror has around a 1V
across its emitter resister etc.

Kevin Aylward

http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Kevin Aylward wrote:

You are obviously not listening to what I am saying. If there is an
offset on the input, there will be an offset on the output, as the amp
is DC coupled throughout. To prevent this output offset doing nasty
things to the speaker, one would then need an output cap.

Uhh... Kevin... if an amplifier is DC coupled throughout, then there are
no coupling caps and no need to worry about the issue.

I hope you mean the cap that is in the feedback network. It is not in in
the loop, i.e. connected from output to input. The cap goes from the
negative input to an ac ground via a resister.

That's not a coupling cap, that is part of the low-pass network used for
a DC servo. That's a different issue altogether (since you can have
massive nonlinearity there at high frequencies and get away with it).

It's still a good idea to use film caps for such things, because if you
make the input impedance high enough (say, with a J-FET front end), you
don't need a very large cap. The major problem with electrolytics is
that their tolerances are very wide and they age poorly, so the turnover
frequency on your servo will vary a lot from amp to amp.

Let's go through this again:

1. If there is no offset, you don't need a capacitor

2. If the offset is larger than the signal, you use a capacitor.

3. If the offset is smaller than the signal, make it larger.

This is not anything particularly difficult here.

Apparently it is for you, as what you say here doesn't make any sense.
What are you trying to say here? Draw your circuit.

I am taking about in ANY circuit. These are general-purpose rules that
exist in any application.

Sure, Self knows his stuff, however I am already an expert in amp
design, I have been so for over 20 years, and have no need for any
instruction in such matters.

What part of "I designed the Studiomaster MOSFET 1000 20+ years ago" did
you miss?

Ummmm.... I hate to tell you this, but that's not going to get you a whole
lot of respect around here.

I dont really understand what you are trying to say here. DC setpoints
are trivial to determine by inspection with basic design knowledge. For
example, the voltage at the collectors of the input stage npn
transistors are forced to 4 Vbes (+~1V volt resister drop) by the common
mode feedback. Some cascode voltages are not important, within reason
they can be anything. The second stage current mirror has around a 1V
across its emitter resister etc.

What I am trying to say is that if we are going to discuss amplifier design
in the abstact, we need both to be able to refer to a single concrete
design. Any design at all, it just has to be one that we can both read
the part designations on.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

"Kevin Aylward" writes:

Are you the only one in the NG so far not to know my universe?

http://www.anasoft.co.uk/band/bio.htm

Hi Kevin -

I see you're a Steely Dan fan. I've had some interest in the band on and
off over the years, mostly because of the very cutting edge recording
technology they embraced, and the resulting sound of some of their
recordings.

Take a look at the "Aja" album as one example -- a tech breakthrough of a
sort at the time (except for the tape saturation of the drums on some
cuts -- but that's another story).

Look at the production credits. Others on this list probably have first
hand info, but I can tell you in general that *every one* of the studios
used for that album were ground-breaking in that they used heavily
modified if not completely custom gear, with most of the tweaking
evaluations *done by ear* -- just as Mr. Williams, Michael Grace, Dan
Kennedy, John Hardy and others do in the final stages of their
design/build processes.

I suggest reading some of the interviews with Scheiner, Schnee, and
Schmitt; woowee! These guys would chew up and spit out
"numbers'n'labcoat" types if the sonic performance wasn't what they
thought it ought to be, no matter what the test gear said.

There really is a bigger world out there in pro-audio on many fronts, and
it includes custom, spendy gear embraced by folks using their ears. Unlike
consumer high-end, in pro high-end brand names, marketing, status, or
price mean little; performance is everything.

Rarely does the $100 mic pre do the job. And the cost of the parts used is
only a small piece of the total picture. You pay guys like the designers
mentioned above for their time and field experience, as opposed to someone
fresh out of school (or someone who perhaps is a career academic) who can
run numbers through a simulation. The latter is a excellent *starting*
point, but that's all it is.

Frank Stearns
Mobile Audio
--

Mark wrote:

The state of the art of signal analysis is advanced to the point where
I do not believe it is possible for a difference in amplifiers to be
audible but not measurable.

Yet it is amazing how many otherwise competent people still
believe in magic.


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

Arny Krueger wrote:

Don't you need two convolutions - one of the target room and one of
the room at hand?

Actually you need four total in real time but not for that
reason. The room at hand is not an issue because the target
is being simulated on 'phones.

I missed the mark when I said any stereo convolver would
work. It actually requires four FIR filters calculated from
the measurements taken from the room and also from the
'phones. Three sweeps must be recorded in all for me to
calculate the four FIRs used in real time. In use the L and
R channels are convolved with two of the calculated FIRs and
summed for the left ear and similarly for the right ear with
the other two FIR's. I should have said it can be done by
any DAW or plug that can do stereo _cross_ convolution.
That leaves only plugs because I know of no DAW that can do
that intrinsically.

Voxengo's Pristine Space VST plug is capable with eight
inputs, eight convolutions and eight outputs which can be
routed and summed arbitrarily. I think his lite version of
it is stuck at simple stereo but I'm going to talk to him
about adding the cross capability to it.

What's nice about his convolver is it offers true, zero
latency partitioned convolution which is pretty much
required for a mixing environment.

The SIR convolver may do it but I haven't used it lately to
know for sure. I'm pretty sure that SIR doesn't have zero
latency, however.

The net of all this is that you should hear your monitors
and room over the phones at the point where you took the
measurements. Works remarkably well for me with my HR824's
in a nasty, unoptimized room over nasty R.S. Titanium
'phones. Of course the position is static, but hey.

I've made an AudioMulch patch using Pristine Space and
Virtual Cable to intercept and process anything that would
be sent to my sound card. I point all my sound apps
including "Multimedia" to Virtual Cable, have Mulch take its
input from that, process it and send it to the sound card.

This doesn't seem to be as specific to a particular pinnae
shape as people in surround seem to think it should be.
I've asked a few others to appraise what I calculated using
mine and they've found it pretty convincing.


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

Arny Krueger wrote:

I would consider that to be an unproven hypothesis. AFAIK nobody has
done any revealtory technical tests that show what say, a Great River
does that is extra wonderful. Nobody has done the controlled listening
tests that would be required to conclusively show that its the
preamp's performance, and not the hype that people *hear*.

A few years back Lynn Fuston set up a big shootout with
about all the pres around at the time. I was told by a
participant in the blind listening tests, which allowed
switching while listening to a performance, that once a few
outlier dogs were eliminated the remainder were pretty
nearly indistinguishable as Kevin contends. Was anybody
here involved in the shootout who could confirm or dispute
that?

I found it telling that no summary of the listening tests
was published as a result, just a CD of test data to judge
for yourself. I don't think anyone has set up an ABX from
that CD but, because all results were recorded serially from
the performers, no two are quite the same and you'd be
judging the performance as well as the pres.


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

On 26 Apr 2005 10:33:14 -0700, "Mark" wrote:

Please provide an example of an amplifier performance characteristic
that can produce an audible difference that cannot be measured.

Without touching any of the third-rail issues like "depth", "air",
or the like, I do sometimes wonder if it's really possible to
measure, or even properly observe, all of the stuff that
plain amplifiers do at very low signal levels.

We tend to often, like Kevin, make sweeping assumptions about
monotonicity, but careful looks at individual components'
transfer curves often show big small-signal errors. (And
giant shrimp.)

Modern amplifiers operated at typically one watt or less at 0VU have
important operating levels to maybe 60dB below that. If the
noise background is only (!) 40dB lower than that, realistic
measurements are difficult.

What kind of numbers are necessary at -60dBVU to not be audible?
I've never heard anybody discuss it; don't mean it ain't happened.

Thanks,

Chris Hornbeck
"This has been an account for those who don't keep them"
J-LG, _Tout Va Bien_ 1972

Bob Cain wrote:

A few years back Lynn Fuston set up a big shootout with
about all the pres around at the time. I was told by a
participant in the blind listening tests, which allowed
switching while listening to a performance, that once a few
outlier dogs were eliminated the remainder were pretty
nearly indistinguishable as Kevin contends. Was anybody
here involved in the shootout who could confirm or dispute
that?

You can order the CD and find out for yourself. I didn't go to that one,
but I did go to the Boston Pre Party where there was a very wide range of
sound on the preamps (in spite of no cheap preamps being selected). On
the other hand, there was a wide range of topology and technology respresented
as well.

I found it telling that no summary of the listening tests
was published as a result, just a CD of test data to judge
for yourself. I don't think anyone has set up an ABX from
that CD but, because all results were recorded serially from
the performers, no two are quite the same and you'd be
judging the performance as well as the pres.

Give it a listen. The CD is interesting.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

"Kevin Aylward" wrote in message
. ..

Yeah, like somone has just been suckered for a few big ones, and they
want to tell everyone about what a big mug they are?

Look, dudes this all well understood 101 psychology. Peoples perceptions
are based on their belief of the facts, not the facts.


So this apples to everything then - no guitar pickup sounds better than
another, no amp sounds better than another. This is obviously untrue. Some
electronic designs manifestly sound 'superior' to others, and there is
variance in manufacturing quality. Some mic preamps will sound better than
others and this is not necessarily a function of price. I've found a few
devices here and there that I'm pretty happy with and my little setup is
'budget' through and through.

As for the value for money aspect, you're guilty of thinking like most
consumers in that value climbs on a fixed linear scale, when in fact value
is variable and the 'value scale' is more geometrical . Most people assume
if you pay twice as much you get twice as much 'value'. But you will often
pay ten times as much for twice as much 'value', because of the nature of
the mass-market economy. In addition, value is different from individual to
individual. Most advertising dollars are in effect spent trying to create a
uniform impression of value.

You're right, though, in that consumers with the dosh often believe they are
getting a percentage return in value on their item that costs 1000 percent
of what the 'common folk' use, and advertisers are more than happy to
accomodate this perception. I haven't really noticed much of that around
here, though, not like in the audiophile groups.

jb

Bob Cain wrote:

I found it telling that no summary of the listening tests
was published as a result, just a CD of test data to judge
for yourself. I don't think anyone has set up an ABX from
that CD but, because all results were recorded serially from
the performers, no two are quite the same and you'd be
judging the performance as well as the pres.

I did order the CD and found exactly what you just said - the
performances all differed enough so that one could easily distinguish
them based only on the performances.

Some sick, demented pommy nut case wrote:


You are obviously not listening to what I am saying. If there is an offset
on the input, there will be an offset on the output, as the amp is DC
coupled throughout.


** There is always a DC voltage relative to ground at the inputs of a
bipolar differential pair due to input *BIAS* current. If that bias current
is say 20 uA for each device and the resistance to ground for the signal
input side is 47 kohms, then the DC offset will be 0.94 volts. This may be
positive or negative depending on the use of PNP or NPN devices.

Normal practice is to match the input devices and make the global feedback
resistor the same value as the input to ground resistance (ie also 22
kohms ) to eliminate DC offset at the speaker output - since when the
speaker output is at ground potential, the same bias current flows in both
resistors hence they will both have the same DC voltage drop.

The use of caps ( typically electros) to feed the input signal into to amp
and couple the gain setting resistor to ground for AC voltages is an
essential part of this scheme - and incidentally applies a **DC bias
voltage** to them.

If one or both these caps is replaced with a link, then the output of the
amp will take on a DC offset - in the case of the feedback cap possibly
equal to the full supply rail.

Eliminating the cap in the NFB loop is not desirable - since the amp then
has full gain down to DC. When such an amp is driven into clipping with an
asymmetrical signal ( speech, music, singing voice etc) the output acquires
an *average* DC offset of many volts for the duration of that clipping.
This does not happen with other amps and will cause large displacements of
the voice coil of a woofer.

See: http://sound.westhost.com/clipping.htm



............ Phil

On Tue, 26 Apr 2005 18:25:41 GMT, "Kevin Aylward"
wrote:

Mike Rivers wrote:
In article
writes:

The point is, it that it costs only a few bucks in electronic
components to produce 0.002% thd, imd, and all the rest of that
wonderfully useless 1000v/us slew rate performance. Like we want to
listen to FM 101 through our mic amps.

Oh, I guess Kevin is coming from that other audio universe where a
"preamp" is something that goes between an FM tuner, cassette deck, CD
player, or phonograph, and has Volume, Balance, and Tone controls as
well as an input selector.

Are you the only one in the NG so far not to know my universe?

http://www.anasoft.co.uk/band/bio.htm

His comment about $500 cables was the
tip-off. That is indeed an industry with a segment that runs on snake
oil.

Jim Williams' work with professional recording and production gear is
a different world

Not at all. Its same ****, different suckers.

Uh oh...

On Tue, 26 Apr 2005 18:05:16 GMT, "Paul Stamler"
wrote:

"Kevin Aylward" wrote in message
k...

Selling a two channel mic amp for $1500 is snake oil. There is zero
chance that its value for money.

Excuse me, but that's tommyrot. I suggest you spend some time with a
Great River preamp and a cheaper preamp with as close a frequency
response as you can find. Do some listening tests, making sure you
match levels carefully. You will almost certainly hear the
difference.

No chance. Hey, what did your horoscope tell you to do today? How about
the tea leaves?

I don't do astrology. I don't do mysticism. I do, however, pay some
attention to what my ears tell me.

Maybe this guy can't tell the difference. Not everyone has ears like
musicians and engineers... you can be trained but some people can just
plain hear better than others... like some people have better vision
or can run faster.

Al

On Tue, 26 Apr 2005 12:25:12 GMT, "Lorin David Schultz"
wrote:

"Arny Krueger" wrote:

Hey, if you are good enough at managing translation, you can even
master with headphones.



If you say so... I sure wouldn't want to, and it's not something I'd
recommend to most people.

You must be either extremely well-versed in compensation techniques or
highly skilled in the art of self-delusion! g

I vote for: B

Al

On Tue, 26 Apr 2005 07:08:55 -0400, "Arny Krueger"
wrote:

Nobody has done the controlled listening
tests that would be required to conclusively show that its the
preamp's performance, and not the hype that people *hear*.

I don't need a controlled listening test to know that I like the sound
of my GR pre a lot more than some other stuff. I liked it better than
the Forsell pre that I had for example.

Al

On 26 Apr 2005 09:44:19 -0400, (Mike Rivers)
wrote:


In article writes:

Yes, but like anyone who posts, its always for a purpose. He expects a
return. No one does anything that they know will be net detriment to
themselves. Its basic evolution.

You read this. Send me money.

I don't get much return on the investment in time that I make here, I
just hope to get famous some day. At least as famous as you anyway. Or
maybe I already am.

The problem with specs is that often the spec are wrong or misleading.
The specs tell the story if they are correct and fully understood.

Jim williams (among others) resolves this problem by actually
listening to equipment after he modifies it. Chances are his customers
will find that it sounds better to them, too. If not, I guess he
wouldn't be able to stay in that business.

Selling a two channel mic amp for $1500 is snake oil. There is zero
chance that its value for money.

Money is money, but value is relative. It may not be worth $1500 to
you, but many people spend more than that for a mic preamp and are
really happy that it gives them better results than the preamp in the
console that they were using previously. Ask him this - how many have
been returned for a refund?

It's not only the issue of sound, but one of build quality and
reliability. $1500 mic pres are not made in China on the cheap.

Al

"Kevin Aylward" wrote in message
...

Sure, Self knows his stuff, however I am already an expert in amp
design, I have been so for over 20 years, and have no need for any
instruction in such matters.

Then you are a fool. Only a fool thinks he has nothing more to learn.

Peace,
Paul

"Bob Cain" wrote in message
...

Mark wrote:

The state of the art of signal analysis is advanced to the point where
I do not believe it is possible for a difference in amplifiers to be
audible but not measurable.

Yet it is amazing how many otherwise competent people still
believe in magic.

Rot. All you need is a measurement almost nobody has thought of making yet.

Let me give an example. What is the output impedance of a power amplifier?
Say, 0.05 ohms at a particular frequency, to pull a number out of Mr.
Aylward's hat. So you measure it at a level of 1 watt equivalent. Measure it
at a different power level. Will you get the same output impedance? Or is
the output impedance level-dependent? Ain't never seen *that* spec in a
manufacturer's literature, or in most test reports either, but it's one
worthy of investigation. I hereby deliver the topic free to any EE in need
of a thesis topic.

Thing is, Mr. Aylward claimed that equivalent performance on a small number
of tests (THD, frequency response, phase linearity and output Z -- were
there more?) meant that audible performance must therefore be equivalent. If
the amplifiers' performance is in fact perfectly described by those
particular tests, of course he's right. But I contend that it isn't, and
therefore it's possible for different amplifiers to be audibly different
despite similar performance on Mr A's tests, and the long-term experience of
many people, professional recording engineers who earn their daily bread
using the gear, bears out that contention. They're ALL self-deluded idiots?
I don't think so.

Peace,
Paul

"Arny Krueger" wrote in message
...
Bob Cain wrote:

I found it telling that no summary of the listening tests
was published as a result, just a CD of test data to judge
for yourself. I don't think anyone has set up an ABX from
that CD but, because all results were recorded serially from
the performers, no two are quite the same and you'd be
judging the performance as well as the pres.

I did order the CD and found exactly what you just said - the
performances all differed enough so that one could easily distinguish
them based only on the performances.

I would be very glad if someone with the resources repeated the experiment
using a reproducible sound source. Say, a Yamaha Reproducing Piano.

Peace,
Paul

Bob Cain wrote:


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein



It is a little dificult to get the exact quote, but none of them use
"described"

Its more like,

"A scientific theory should be as simple as possible, but no simpler"

or

"Everything should be as simple as possible, but no simpler"

Kevin Aylward

http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Bob Cain wrote:
Mark wrote:

The state of the art of signal analysis is advanced to the point
where I do not believe it is possible for a difference in amplifiers
to be audible but not measurable.

Yet it is amazing how many otherwise competent people still
believe in magic.

Indeed. A belief in magic is the norm, to wit the billions that believe
in a god.


Kevin Aylward

http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Phil Allison wrote:



You are obviously not listening to what I am saying. If there is an
offset on the input, there will be an offset on the output, as the
amp is DC coupled throughout.


** There is always a DC voltage relative to ground at the inputs of a
bipolar differential pair due to input *BIAS* current. If that bias
current is say 20 uA for each device and the resistance to ground for
the signal input side is 47 kohms, then the DC offset will be 0.94
volts. This may be positive or negative depending on the use of PNP
or NPN devices.

As usually, Phil misses the boat completely.

We are not discussing DC bias offsets, well I'm not anyway. We are,
apparently, discussing the use of putting additional bias on the cap of
the feedback network, say several volts in order to keep the cap always
polarised in the same direction.


Normal practice is to match the input devices

{snip drivel}

Look, Phil, I am an expert in amplifier design, there is *nothing* you
can tell me, so don't try and teach your grandpa to suck eggs.

Since your so clever, how about giving the readers a detailed analysis
of http://www.anasoft.co.uk/EE/circuits...ortionAmp2.jpg and
how the output stage contributes to the 0.001% thd at 20khz of this amp.
(oh..some diode are zeners)

Kevin Aylward

http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Scott Dorsey wrote:
Kevin Aylward wrote:

You are obviously not listening to what I am saying. If there is an
offset on the input, there will be an offset on the output, as the
amp is DC coupled throughout. To prevent this output offset doing
nasty things to the speaker, one would then need an output cap.

Uhh... Kevin... if an amplifier is DC coupled throughout, then there
are no coupling caps and no need to worry about the issue.

Pardon!!!!

If there is DC on the + input there will be DC on the output of a DC
coupled amp. Dah...

This is getting out of control mate. You don't know anything do you?


I hope you mean the cap that is in the feedback network. It is not
in in the loop, i.e. connected from output to input. The cap goes
from the negative input to an ac ground via a resister.

That's not a coupling cap, that is part of the low-pass network used
for a DC servo.

Yep. This is indeed getting daft. You dont know anything about amp
design do you.

What bloody servo?

There is a potential divider from the output to ground, consisting of
two resisters and a cap. The tap point is connected to the neg amp
input. The ground side has the cap so that that gain for DC is unity.


That's a different issue altogether (since you can
have massive nonlinearity there at high frequencies and get away with
it).

Oh dear...what are you on about dude...You just stringing words
together, aimlessly.



It's still a good idea to use film caps for such things, because if
you make the input impedance high enough (say, with a J-FET front
end), you don't need a very large cap. The major problem with
electrolytics is that their tolerances are very wide and they age
poorly, so the turnover frequency on your servo will vary a lot from
amp to amp.

You are simply spouting of words with no idea what you are talking
about. Its clear that you are out of your depth here. Look sonny, your
discusing this with a *real* analogue designer.

Even if the amp design had a bloody servo, its low frequency pole would
be a non issue. Whether its 1H or 5 hz wont matter squat.


Let's go through this again:

1. If there is no offset, you don't need a capacitor

2. If the offset is larger than the signal, you use a capacitor.

3. If the offset is smaller than the signal, make it larger.

This is not anything particularly difficult here.

Apparently it is for you, as what you say here doesn't make any
sense. What are you trying to say here? Draw your circuit.

I am taking about in ANY circuit.

I want to know about *this* circuit. The one you claim is usefull to
bias up the cap in the feedback network of 1000's of bog standrd amp
designs in order to prevent reverse bise on the feeback network cap.

These are general-purpose rules
that exist in any application.

And you don't know or understand any of them.

Oh dear...why has it taken me this long to figure out that you are
totally clueless on electronics.


Sure, Self knows his stuff, however I am already an expert in amp
design, I have been so for over 20 years, and have no need for any
instruction in such matters.

What part of "I designed the Studiomaster MOSFET 1000 20+ years ago"
did you miss?

Ummmm.... I hate to tell you this, but that's not going to get you a
whole lot of respect around here.

Oh... its 200Khz power bandwidth, 0.002% thd, I don't know of many power
amps that even today compete with that spec. However, I am not
responsible for the mains switch welding on.


I dont really understand what you are trying to say here. DC
setpoints are trivial to determine by inspection with basic design
knowledge. For example, the voltage at the collectors of the input
stage npn transistors are forced to 4 Vbes (+~1V volt resister drop)
by the common mode feedback. Some cascode voltages are not
important, within reason they can be anything. The second stage
current mirror has around a 1V across its emitter resister etc.

What I am trying to say is that if we are going to discuss amplifier
design in the abstact, we need both to be able to refer to a single
concrete design. Any design at all, it just has to be one that we
can both read the part designations on.
--scott

Discussing amplifier design with you is a complete non starter as you
know so little about it. All you know is some bits and bob you picked up
from the back of an envelope.

Nothing you have said makes any sense. You just use technical words with
no understanding as to their meaning.

Kevin Aylward

http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Some sick, pathetic, pommy maniac & ranting nut case wrote:
Phil Allison wrote:


You are obviously not listening to what I am saying. If there is an
offset on the input, there will be an offset on the output, as the
amp is DC coupled throughout.


** There is always a DC voltage relative to ground at the inputs of a
bipolar differential pair due to input *BIAS* current. If that bias
current is say 20 uA for each device and the resistance to ground for
the signal input side is 47 kohms, then the DC offset will be 0.94
volts. This may be positive or negative depending on the use of PNP
or NPN devices.

As usually, Phil misses the boat completely.


** That "boat" must the Titanic - and Manic Kev just went down with the
ship.

There can indeed be large DC offsets at the input that are not at the
output.



We are not discussing DC bias offsets, well I'm not anyway.


** Scott specifically referred to them in his earlier post.


Normal practice is to match the input devices....


** Kev snipped all the points I made that show how wrong he is.


Look, Phil, I am an expert in amplifier design,


** Only in his sick and demented mind.


there is *nothing* you can tell me,


** I am certain that IS true.

There is nothing anyone can tell a maniac.





.............. Phil

Scott Dorsey wrote:
Mark wrote:

The state of the art of signal analysis is advanced to the point
where I do not believe it is possible for a difference in
__amplifiers__ to be audible but not measurable. It may not be
obvious at first what measurement is needed, but once that is
unraveled, some measurement should always be able to explain any
audible difference (talking about amplifiers here, mics and
loudspeakers are another matter)

Absolutely, but that's the rub!

If a critical listening test should uncover an audible difference
between two amps, the next step should be to make whatever
measurements are needed to identify the cause of the difference.
Then the designers can use that information to make design changes.
Something like, amp A has a more solid sound compared to amp B and
we discovered that the reason is the difference in damping factor (a
trivial example).

This is true, BUT we have folks here in this group who are claiming
that low THD levels obviate the need for any other testing.

I hope you are not referring to me, as this is not an accrate reflection
of my approach.

As an analogue designer, one considers many factors, such as power BW,
noise, dynamic basing offsets etc.

For the sake of brevity, one often makes simplified statements on the
assumption that the reader understands the implications of those
statements. This can result in misunderstandings as to the true intent.
Generally, one debates as if the other is on a similar basis. For
example, a liar is more likely to assume that whoever he encounters is
also a liar. That is, you can spot a liar as they usually are always
accusing others of lying. Liaring is always on their mind, so they just
naturally assume that that is also true for other minds.

Why is this relevent here?

When a seasoned designer makes an *implication* that say, 0.01% thd at
20khz, might be all that is required, it is made on the assumption of
what, essentially, automatically follows from achieving such spec in a
typical design.

For example, to achieve such low distortion at high frequency, usually
means that large amounts of feedback is used. This means that the unity
gain frequency must be very high. This automatically means that the BW
of the amp is going to be large, all things being equal, therefore,
testing for BW, is pretty much redundant, as it is also going to be
large. Of course one can deliberately contrive an amp for the purposes
of onupmanship that has low distortion, and 1khz BW.

What about noise testing?

Well, when I measured my MOSFET 1000 design on the bench I obtained some
interesting results.

1 Distortion from oscilater 0.0018%
2 Distortion from amp output 0.0015%

So, why the lower distortion from the amp?

Hint: a practical measure of distortion, automatically makes moise
measurements somewhat redundant.

So, of course, one spec does not tell all, but achieving certain specs,
often means that many over specs all follow in line. People just don't
go around deigning power amps with 0.001% thd, that have +/-10dbs
frequency variations.

Hell, we
have one person here claiming that TIMD is total bunk.

Essentially, it is.

For any competently designed amp, it is zero. Indeed, the claims as to
what it even is are dubious. There are claims that this "TID" is due to
the open loop BW not being greater then the required audio BW. The
argument being that the feedback cant get back quick enough, so that
there are transient errors. This is of course, completely bogus, and is
based on a complete misunderstanding of how feedback amplifier work.

In the past there were some amps that had an ok small signal BWs, but at
full power, slew rate limits caused the amp to triangle out. This type
of performance can result in large IMD, or characteristically to the
name of TID, if a high frequency square wave was superimposed on a sine
wave, the sine wave would undergo more harmonic distortion than without
the square wave. TID is a bit bogus, in that if 20Khz distortion is low,
so will "TID". Amps that slew limit at 15 khz, have dreadfull 20khz thd
and 19khz/20Khz imd.

If one designs an amp to have a full power BW of say, 60khz, and band
limits the input to say, 40Khz, then TID is essentially, zero. With
normal signals, no filtering is usually necessary.

I have specifically analysed the effect of superimposed slew rates on
sine distortion. The conclusion of the analysis is that one does really
want around twice the audio BW as a full power BW. During slew, the
input diff pair becomes unbalanced, which results in larger distortion.


Please provide an example of an amplifier performance characteristic
that can produce an audible difference that cannot be measured.

I don't know of any that cannot, but I know of a whole lot that never
will, because of the reluctance on the part of some people to
investigate more modern measurements.

You mean those of us that don't look for pink unicorns either.

Kevin Aylward

http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

reddred wrote:
"Kevin Aylward" wrote in message
. ..

Yeah, like somone has just been suckered for a few big ones, and they
want to tell everyone about what a big mug they are?

Look, dudes this all well understood 101 psychology. Peoples
perceptions are based on their belief of the facts, not the facts.


So this apples to everything then - no guitar pickup sounds better
than another, no amp sounds better than another. This is obviously
untrue.

Of course it isn't. Anything that involves conscious awareness is pretty
much arbitrary. How do you wish to diffine "better"?

For example, I absolutely hate the sound of the bridge pick position on
a guitar. I never use it. Others, love the sound. Should I list the
bands that I detest and those I hate?

What we "like", i.e. what our conscious awareness perceives "is
something that we want to experience more of", is software programmed by
environmental experience, i.e. memes.

An illustration. Suppose when one is young, one eats something, and then
shortly afterwards, one gets ill. What often happens is that one
develops a dislike for that particular food. The body system makes an
assumption that it was that food that made one ill, and takes steps to
prevent it occurring again. This is an evolutionary mechanism to allow
us to deal with a wide range of differing environments. It would be way
to complicated to have hard coded genetics instructing us as to what
food was poisonous and what was not. There are way too many
possibilities. Evolution solves this problem by making our perceptions
programmable. All of our emotion (like, hate, envey) are programable as
to when they are insigated. This is addressed specifically at
http://www.anasoft.co.uk/replicators/emotions.html, and
http://www.anasoft.co.uk/replicators/index.html

Do you like the Clingon battle song by the way?

Kevin Aylward

http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Frank Stearns wrote:
"Kevin Aylward" writes:

Are you the only one in the NG so far not to know my universe?

http://www.anasoft.co.uk/band/bio.htm

Hi Kevin -

I see you're a Steely Dan fan. I've had some interest in the band on
and off over the years, mostly because of the very cutting edge
recording technology they embraced, and the resulting sound of some
of their recordings.

Take a look at the "Aja" album as one example -- a tech breakthrough
of a sort at the time (except for the tape saturation of the drums on
some cuts -- but that's another story).

Look at the production credits. Others on this list probably have
first hand info, but I can tell you in general that *every one* of
the studios used for that album were ground-breaking in that they
used heavily modified if not completely custom gear, with most of the
tweaking evaluations *done by ear* -- just as Mr. Williams, Michael
Grace, Dan Kennedy, John Hardy and others do in the final stages of
their design/build processes.


First, I have no idea exactly what was tweaked, and what was changed, so
none of this means much.

However, I don't see that we are referring to the same issues. Of
course, audio requires significant subjective setup of equipment. For
example, tone controls, compression, reverb etc. No one is denying this.

It may also be that, e.g. tone control frequencies were manually hacked.
None of this is pertinent to what I am addressing. That is, e.g. the
audible difference between a well designed cheap mic amp with say, 0.02%
thd, 10V/us slew rate, and that of a 0.001%, 3000v/us slew rate, is
ZERO.

Of course, people may well prefer tube mic preamps for their sound. This
is a justifiable claim. 1% thd is well documented to be detectable.
There is no scientific principle that dictates that the more accurate
the better it is consciously perceived.


I suggest reading some of the interviews with Scheiner, Schnee, and
Schmitt; woowee! These guys would chew up and spit out
"numbers'n'labcoat" types if the sonic performance wasn't what they
thought it ought to be, no matter what the test gear said.

I don't claim that all equipment sounds the same, or that straight piece
of wire with gain specs sounds better. I claim that any set of
comparable systems that exceeds certain, easily achievable specs, sound
the same. i.e. a graphics with the same band frequencies, and said
specs. I shouldn't have to state that I am obviously, not claiming that
say, amps with great specs with different filters sound the same. I am
assuming a comparing of apples with apples as taken as read.


There really is a bigger world out there in pro-audio on many fronts,
and it includes custom, spendy gear embraced by folks using their
ears. Unlike consumer high-end, in pro high-end brand names,
marketing, status, or price mean little; performance is everything.

Rarely does the $100 mic pre do the job.

I have to disagree on this one. Modern design allows for a mic amp to be
audible a straight piece of wire with gain. Excepting, of course, where
one wants the colouration of a tube amp. I even suppose, they may be
some that like an old 741 mic pre amp.

And the cost of the parts
used is only a small piece of the total picture. You pay guys like
the designers mentioned above for their time and field experience, as
opposed to someone fresh out of school (or someone who perhaps is a
career academic) who can run numbers through a simulation. The latter
is a excellent *starting* point, but that's all it is.


Well, yes.


Kevin Aylward

http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

playon wrote:
On Tue, 26 Apr 2005 18:05:16 GMT, "Paul Stamler"
wrote:

"Kevin Aylward" wrote in message
...

Selling a two channel mic amp for $1500 is snake oil. There is
zero chance that its value for money.

Excuse me, but that's tommyrot. I suggest you spend some time with
a Great River preamp and a cheaper preamp with as close a frequency
response as you can find. Do some listening tests, making sure you
match levels carefully. You will almost certainly hear the
difference.

No chance. Hey, what did your horoscope tell you to do today? How
about the tea leaves?

I don't do astrology. I don't do mysticism. I do, however, pay some
attention to what my ears tell me.

Maybe this guy can't tell the difference. Not everyone has ears like
musicians and engineers... you can be trained but some people can just
plain hear better than others... like some people have better vision
or can run faster.


Billions are trained from birth that an invisible, all powerful, can do
anything, be everywhere at once, boggy man in the sky made us, and even
have personal experiences to "prove" it. So, what's your point?

Kevin Aylward

http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

On Wed, 27 Apr 2005 09:39:58 GMT, "Kevin Aylward"
wrote:

playon wrote:
On Tue, 26 Apr 2005 18:05:16 GMT, "Paul Stamler"
wrote:

"Kevin Aylward" wrote in message
...

Selling a two channel mic amp for $1500 is snake oil. There is
zero chance that its value for money.

Excuse me, but that's tommyrot. I suggest you spend some time with
a Great River preamp and a cheaper preamp with as close a frequency
response as you can find. Do some listening tests, making sure you
match levels carefully. You will almost certainly hear the
difference.

No chance. Hey, what did your horoscope tell you to do today? How
about the tea leaves?

I don't do astrology. I don't do mysticism. I do, however, pay some
attention to what my ears tell me.

Maybe this guy can't tell the difference. Not everyone has ears like
musicians and engineers... you can be trained but some people can just
plain hear better than others... like some people have better vision
or can run faster.


Billions are trained from birth that an invisible, all powerful, can do
anything, be everywhere at once, boggy man in the sky made us, and even
have personal experiences to "prove" it. So, what's your point?

As a guitarist, do you have a day job?

Al

Kev, thanks a bunch for dropping by and explaining everything.

On Wed, 27 Apr 2005 09:36:11 GMT, "Kevin Aylward"
wrote:

reddred wrote:
"Kevin Aylward" wrote in message
. ..

Yeah, like somone has just been suckered for a few big ones, and they
want to tell everyone about what a big mug they are?

Look, dudes this all well understood 101 psychology. Peoples
perceptions are based on their belief of the facts, not the facts.


So this apples to everything then - no guitar pickup sounds better
than another, no amp sounds better than another. This is obviously
untrue.

Of course it isn't. Anything that involves conscious awareness is pretty
much arbitrary. How do you wish to diffine "better"?

For example, I absolutely hate the sound of the bridge pick position on
a guitar. I never use it. Others, love the sound. Should I list the
bands that I detest and those I hate?

What we "like", i.e. what our conscious awareness perceives "is
something that we want to experience more of", is software programmed by
environmental experience, i.e. memes.

An illustration. Suppose when one is young, one eats something, and then
shortly afterwards, one gets ill. What often happens is that one
develops a dislike for that particular food. The body system makes an
assumption that it was that food that made one ill, and takes steps to
prevent it occurring again. This is an evolutionary mechanism to allow
us to deal with a wide range of differing environments. It would be way
to complicated to have hard coded genetics instructing us as to what
food was poisonous and what was not. There are way too many
possibilities. Evolution solves this problem by making our perceptions
programmable. All of our emotion (like, hate, envey) are programable as
to when they are insigated. This is addressed specifically at
http://www.anasoft.co.uk/replicators/emotions.html, and
http://www.anasoft.co.uk/replicators/index.html

Do you like the Clingon battle song by the way?

Kevin Aylward

http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Phil Allison wrote:

You are obviously not listening to what I am saying. If there is an
offset on the input, there will be an offset on the output, as the
amp is DC coupled throughout.


** There is always a DC voltage relative to ground at the inputs
of a bipolar differential pair due to input *BIAS* current. If that
bias current is say 20 uA for each device and the resistance to
ground for the signal input side is 47 kohms, then the DC offset
will be 0.94 volts. This may be positive or negative depending on
the use of PNP or NPN devices.

As usually, Phil misses the boat completely.


** That "boat" must the Titanic - and Manic Kev just went down
with the ship.

There can indeed be large DC offsets at the input that are not at the
output.


I am talking about offsets directly added at the diff pair input, not
before any coupling caps. I am not talking about the inherent amplifier
"input offsets"



We are not discussing DC bias offsets, well I'm not anyway.


** Scott specifically referred to them in his earlier post.

Not in the sense I took them to be. If what is being discussed is
something as mundane as common bias offset voltages due to transistor
bias currents and vbe offsets, then this whole discussion is completely
pointless, and stupid.

My assumption was that Scott is discussing putting a deliberate external
bias on the feedback network capacitor so that if the *signal* across if
the cap reversed in polarity, the net voltage on the cap would still be
the correct direction. This was to account for *AC* signals assumed to
be in 0.5V range.

The idea that one is discussing the small offsets inherent in the amp
circuit itself, under static conditions is totally absurd. There is no
issue with such offsets. For example, yes, my MOSFET 1000 had a typical
offset of 10mv. This is of no consequence to a polar electrolytic
capacitor. It isn't worth disusing, so I wasn't. I am astounded that
could have been Scots intent. Its a non issue.



Normal practice is to match the input devices....


** Kev snipped all the points I made that show how wrong he is.


Look, Phil, I am an expert in amplifier design,


** Only in his sick and demented mind.


Phil, you need the face that facts. I have been a successful analogue
designer for some while now. You on the hand, have nothing to say but
insults.

Best Regards

Kevin Aylward

http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Paul Stamler wrote:
"Bob Cain" wrote in message
...

Mark wrote:

The state of the art of signal analysis is advanced to the point
where I do not believe it is possible for a difference in
amplifiers to be audible but not measurable.

Yet it is amazing how many otherwise competent people still
believe in magic.

Rot. All you need is a measurement almost nobody has thought of
making yet.

It's not a matter of thinking, its a matter of doing. And with the
local disciples of Ned Ludd name-calling everybody who ever measured
anything, why would someone risk posting the results of their
measurements here?

Let me give an example. What is the output impedance of a power
amplifier? Say, 0.05 ohms at a particular frequency, to pull a
number
out of Mr. Aylward's hat. So you measure it at a level of 1 watt
equivalent. Measure it at a different power level. Will you get the
same output impedance?

With good amps, the answer is yes the output impedance is stable. The
output impedance of basic SS amp circuitry is often so low that the
measured output impedance has strong components due to output
inductors and circuit board traces. Those are copper and highly
linear. The output impedance of the active circuitry often takes a
slight downward turn in the 1-10-40 watt range because many output
devices have betas that actually go up as you drive them harder, at
least to a point. It's all pretty moot. The 20-20 KHz output impedance
of most SS amps is dominated by the output inductor if there is one.

Or is the output impedance level-dependent?

Not until the power levels get way up there. The amp I studied output
impedance didn't shift appreciably from 1 to 10 to 100 to 400 watts.

Ain't never seen *that* spec in a manufacturer's literature, or in
most test reports either, but it's one worthy of investigation. I
hereby deliver the topic free to any EE in need of a thesis topic.

Mostly a lost cause.

Thing is, Mr. Aylward claimed that equivalent performance on a small
number of tests (THD, frequency response, phase linearity and output
Z -- were there more?) meant that audible performance must therefore
be equivalent.

It's pretty much the rule that the performance of modern amps exceeds
needs for transparent operation by a fair margin. The most audible
attribute of mnay modern power amps can be heard if you hook them up
in series, separated by attenuators and simulated speaker loads.
Generally, the first thing you hear is slight shift in upper midrange
presence due to the output inductors interacting with the speaker
loads.

If you listen to the "5 times" samples posted at
http://www.pcabx.com/product/amplifiers/index.htm this effect is
audible with most of the amps tested. It was also readily measurable.
I seem to recall that one of the amps lacked an output inductor.
Figuring out which one is left as an exercise for the listeners.

The only one of the tested amps with a posted detailed test report is
shown at http://www.pcavtech.com/pwramp/macrot-5000VZ/index.htm . Hi
Rez plots of frequency response with a simulated speaker load is show
for 1, 10 and 400 watts. Several of the other amps whose sound
samples is posted showed similar effects.

If the amplifiers' performance is in fact perfectly
described by those particular tests, of course he's right. But I
contend that it isn't, and therefore it's possible for different
amplifiers to be audibly different despite similar performance on Mr
A's tests, and the long-term experience of many people, professional
recording engineers who earn their daily bread using the gear, bears
out that contention. They're ALL self-deluded idiots? I don't think
so.

It takes about a day of hard work to transform the measured data into
a report of the kind I referenced above. Sorry about whimping out on
the technical tests for the other amps, but spot checks showed that it
was all pretty much the same in terms of trends.

Paul Stamler wrote:
"Arny Krueger" wrote in message
...
Bob Cain wrote:

I found it telling that no summary of the listening tests
was published as a result, just a CD of test data to judge
for yourself. I don't think anyone has set up an ABX from
that CD but, because all results were recorded serially from
the performers, no two are quite the same and you'd be
judging the performance as well as the pres.

I did order the CD and found exactly what you just said - the
performances all differed enough so that one could easily
distinguish
them based only on the performances.

I would be very glad if someone with the resources repeated the
experiment using a reproducible sound source. Say, a Yamaha
Reproducing Piano.


You and me both. Only, I'm not sure that a piano is a critical enough
source.

For a completely unrelated reason I've been making some recordings of
reproduced sound with measurement mics. The long and short of is that
using more complex and taxing recordings played through good speakers
as a source for mic/preamp tests maybe shouldn't be dismissed so
easily.