Reply
 
Thread Tools Display Modes
  #41   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



Arny Krueger wrote:

"Eeyore" wrote:
Arny Krueger wrote:


For one thing, modern output transistors have Ft up in
the megahertz range. Back in the 70s people were still
struggling with output devices with Ft in the KHz range.


Has no influence on the change in transconductance of
output devices with load current.


But it does have something to do with the shape of such transients as show
up when you there is some crossover distortion.


Not directly. The open-loop bandwidth of the amplifier (which can be higher
with modern high fT devices) does that. The larger bandwidth means more NFB
can be applied at the higher audio frequencies which in turn reduces the level
of the high harmonic order THD products.

You COULD build an indifferent amplifier design with high fT devices and not
get that benefit you see.


But thanks for bringing up the issue of change in transconductance of the
output devices, particularly at low current levels. That's another issue
that modern devices sometimes do better at.


I doubt that it'll be much different. It's the basic device physics you see.
You can reduce the impact by cleverer design though.


Stop talking irrelevant nonsense.


Lighten up, Graham. So seem to forget about all the truely weird nonsense
that we see on Usenet. How long has it been since you were on AAPLS?


I expect better from you.


How many amplifiers have YOU designed btw ?


More than I'd care to admit here. The experience made me a believer in
commercial amps. :-(


You're aware that I've designed a fair number of successful commercial designs
I believe.

Graham


  #42   Report Post  
Posted to rec.audio.tech
Peter Larsen[_2_] Peter Larsen[_2_] is offline
external usenet poster
 
Posts: 724
Default Stereo receievers: THD of .04 vs. .08

Eeyore wrote:

And the bias current in the output stage is a far larger fraction of
the maximum load current. It may operate entirely in Class A with
many or even most loads. Not many people are using 600 ohm working.


Let that apply for the poweramp too ... not at an unrealistic
concept for home use ... and there is NO issue.


Would you seriously suggest an amp or so of quiescent current would be
acceptable ? It certainly WILL fix the problem of course.


Please enhance that point, I can see what you mean, but not how it applies.

Graham



Kind regards

Peter Larsen



  #43   Report Post  
Posted to rec.audio.tech
GregS[_3_] GregS[_3_] is offline
external usenet poster
 
Posts: 664
Default Stereo receievers: THD of .04 vs. .08

In article , Eeyore wrote:


Arny Krueger wrote:

"Peter Larsen" wrote
Eeyore wrote:

Remember, 5532s are class AB and bipolar. ;-)

And the bias current in the output stage is a far larger
fraction of the maximum load current. It may operate
entirely in Class A with many or even most loads. Not
many people are using 600 ohm working.

Let that apply for the poweramp too ... not at an
unrealistic concept for home use ... and there is NO
issue.


IME, most home systems spend much of their time running at very low levels
where neither side of the output stage is ever cut off.


It's not the cutting off per se that causes the distortion. Just modulating
the collector current of an emitter follower output stage by any significant
fraction of the quiescent value will cause distortion due to the change in
----


I put together my "big" amp some years ago. I had to keep tweaking it to get it stable.
What I ended up with was a bridge amplifier with a lot of crossover distortion. Since the scecond channel was
fed the inverted output of the first channel, if you looked at it differetially on a scope, the crossover distortion
vanished. What the hell, I said, and used it like that. It was just fine. I didn't want to jack up the bias any more.
I had a lot of predriver stages.

greg
  #44   Report Post  
Posted to rec.audio.tech
Arny Krueger Arny Krueger is offline
external usenet poster
 
Posts: 17,262
Default Stereo receievers: THD of .04 vs. .08

"Eeyore" wrote in
message
Arny Krueger wrote:

"Eeyore" wrote:
Arny Krueger wrote:


For one thing, modern output transistors have Ft up in
the megahertz range. Back in the 70s people were still
struggling with output devices with Ft in the KHz
range.


Has no influence on the change in transconductance of
output devices with load current.


But it does have something to do with the shape of such
transients as show up when you there is some crossover
distortion.


Not directly. The open-loop bandwidth of the amplifier
(which can be higher with modern high fT devices) does
that. The larger bandwidth means more NFB can be applied
at the higher audio frequencies which in turn reduces the
level of the high harmonic order THD products.

You COULD build an indifferent amplifier design with high
fT devices and not get that benefit you see.


But thanks for bringing up the issue of change in
transconductance of the output devices, particularly at
low current levels. That's another issue that modern
devices sometimes do better at.


I doubt that it'll be much different. It's the basic
device physics you see. You can reduce the impact by
cleverer design though.


Stop talking irrelevant nonsense.


Lighten up, Graham. So seem to forget about all the
truely weird nonsense that we see on Usenet. How long
has it been since you were on AAPLS?


I expect better from you.


You're getting it. Most of AAPLS stop thinking about the time you say
transconductance., if not well before.

How many amplifiers have YOU designed btw ?


More than I'd care to admit here. The experience made me
a believer in commercial amps. :-(


It's like what Scott said in a recent post - relying on people who do
something better than you.

My amp designs in the mid 80s (???) got to the point where there I was kinda
stopped at 0.03% THD at full power at 20 KHz, and less than that at all
frequencies and power levels below that with excellent stability into just
about any load. I wanted better, if I was going to bother building any of
them up into a finished component. The DH 200 did better than that, and if
my amp couldn't whip a DH 200, why bother?

You're aware that I've designed a fair number of
successful commercial designs I believe.


Yes, but you seem to be very worried about things that just don't happen
with the modern SS amps I've tested, even fairly rugged cheap locomotives
like the QSC USA series.


  #45   Report Post  
Posted to rec.audio.tech
Chris Hornbeck Chris Hornbeck is offline
external usenet poster
 
Posts: 1,744
Default Stereo receievers: THD of .04 vs. .08

On Fri, 11 Jan 2008 18:11:27 +0000, Eeyore
wrote:

Would you seriously suggest an amp or so of quiescent current would be
acceptable ? It certainly WILL fix the problem of course.


The real danger of this line of inquiry, and the parallel
one of intrinsic linearity, is that the inquirer will end
up in the untouchable cast, those sad degenerates who
experiment with single-ended triodes and such-like
proscribed topologies.

Following the social stigma, a new life as a crack dealer
is only a short step. Woe and rue the day that you ask
too many questions.

Much thansk, as always,

Chris Hornbeck

"There's little that's impossible, but it becomes more complicated if
you move between different systems." - Mike Rivers, in another context


  #46   Report Post  
Posted to rec.audio.tech
Trevor Wilson[_2_] Trevor Wilson[_2_] is offline
external usenet poster
 
Posts: 724
Default Stereo receievers: THD of .04 vs. .08


"Arny Krueger" wrote in message
...
"Eeyore" wrote in
message
JANA wrote:

With the distortion you mentioned, there will be no
difference that you can hear. Even at 0.1% THD
distortion, you cannot hear that.


The distortion spec is measured at close to full power.


The real distortion at lower powers where most of the
listening is done may be as much as TEN times higher with
typical Class AB bipolar transistor designs.


No way with modern amps. Any big rises at low levels that may be shown in
published tests is due to noise. The truth outs if someone actually
publishes a spectral analysis.

For one thing, modern output transistors have Ft up in the megahertz
range. Back in the 70s people were still struggling with output devices
with Ft in the KHz range.


**Nope. The Japanese had a goodly number of complementary devices, with
decent Voltage, current and SOA ratings which extended well into the tens of
MHz. By 1974, Hi-Rel had true Triple Diffused, TO3 devices with a 20MHz fT
and 20 Amps capacity. Exceptionally linear and very potent devices. I recall
them vividly, since Motorola had announced their intention to commit to
Epitaxial Base devices (ca. 2MHz), since it was impossible to produce Triple
Diffused devices (according to Motorola). Other manufacturers also had some
impressive devices back then. The Marantz 2325 receiver used 2SA747/2SC1116
devices. These posessed an fT of 15MHz.

Having said all that, good gain linearity over a wide current range has only
been a relatively recent development.

Trevor Wilson




  #47   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



Peter Larsen wrote:

Eeyore wrote:

And the bias current in the output stage is a far larger fraction of
the maximum load current. It may operate entirely in Class A with
many or even most loads. Not many people are using 600 ohm working.


Let that apply for the poweramp too ... not at an unrealistic
concept for home use ... and there is NO issue.


Would you seriously suggest an amp or so of quiescent current would be
acceptable ? It certainly WILL fix the problem of course.


Please enhance that point, I can see what you mean, but not how it applies.


How it applies is to swamp the non-linear effect that results from changes in
output device transconductance when biased to lower values of quiescent
current.

If you're not familiar with that effect please say so.

Graham

  #48   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



Arny Krueger wrote:

"Eeyore" wrote in

I expect better from you.


You're getting it. Most of AAPLS stop thinking about the time you say
transconductance., if not well before.


This ISN'T a.a.p.l-s !

This group has the word tech in its title. It's reasonable to expect a level of
technical competence when posting here.

Graham

  #49   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



Arny Krueger wrote:

"Eeyore" wrote in

You're aware that I've designed a fair number of
successful commercial designs I believe.


Yes, but you seem to be very worried about things that just don't happen
with the modern SS amps I've tested, even fairly rugged cheap locomotives
like the QSC USA series.


The QSC USA series is a COMPLETE HEAP OF CRAP and illustrates many classic
design flaws and is a good example of just how NOT to build a high performing
amplifier. The USA series is also most certainly not *modern* by any reasonable
understanding of the word.

If you can't hear the flaws of a USA series you must be completely deaf.

Graham




  #50   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



Trevor Wilson wrote:

"Arny Krueger" wrote


For one thing, modern output transistors have Ft up in the megahertz
range. Back in the 70s people were still struggling with output devices
with Ft in the KHz range.


**Nope. The Japanese had a goodly number of complementary devices, with
decent Voltage, current and SOA ratings which extended well into the tens of
MHz. By 1974, Hi-Rel had true Triple Diffused, TO3 devices with a 20MHz fT
and 20 Amps capacity.


Hi-Rel ?


Exceptionally linear and very potent devices. I recall
them vividly, since Motorola had announced their intention to commit to
Epitaxial Base devices (ca. 2MHz), since it was impossible to produce Triple
Diffused devices (according to Motorola). Other manufacturers also had some
impressive devices back then. The Marantz 2325 receiver used 2SA747/2SC1116
devices. These posessed an fT of 15MHz.


I recall being impressed by Sanken devices the first time I was designing a
serious amplifier around 1980. I settled for Hitachi lateral mosfets instead.
Obtainability was a serious issue back then.


Having said all that, good gain linearity over a wide current range has only
been a relatively recent development.


With modern low-Z driver stages, current gain linearity is no longer an issue
the way it was.

Graham



  #51   Report Post  
Posted to rec.audio.tech
Trevor Wilson[_2_] Trevor Wilson[_2_] is offline
external usenet poster
 
Posts: 724
Default Stereo receievers: THD of .04 vs. .08


"Eeyore" wrote in message
...


Trevor Wilson wrote:

"Arny Krueger" wrote


For one thing, modern output transistors have Ft up in the megahertz
range. Back in the 70s people were still struggling with output devices
with Ft in the KHz range.


**Nope. The Japanese had a goodly number of complementary devices, with
decent Voltage, current and SOA ratings which extended well into the tens
of
MHz. By 1974, Hi-Rel had true Triple Diffused, TO3 devices with a 20MHz
fT
and 20 Amps capacity.


Hi-Rel ?


**Yep, Hi-Rel. They were a small manufacturer back then. A mate discovered
them and used their devices for his amplifiers. Very impressive devices. I
have a spec sheet somewhere. When I locate it, I will scan and post, if you
like. Purely historical, of course. The devices have been unavailable for
quite awhile. FWIW, the numbers a ED203 and EB203.



Exceptionally linear and very potent devices. I recall
them vividly, since Motorola had announced their intention to commit to
Epitaxial Base devices (ca. 2MHz), since it was impossible to produce
Triple
Diffused devices (according to Motorola). Other manufacturers also had
some
impressive devices back then. The Marantz 2325 receiver used
2SA747/2SC1116
devices. These posessed an fT of 15MHz.


I recall being impressed by Sanken devices the first time I was designing
a
serious amplifier around 1980. I settled for Hitachi lateral mosfets
instead.
Obtainability was a serious issue back then.


**By 1980, Japanese transistors were much more easily available. The 2SD555
and it's complement were my favourites at the time.



Having said all that, good gain linearity over a wide current range has
only
been a relatively recent development.


With modern low-Z driver stages, current gain linearity is no longer an
issue
the way it was.


**True enough, but, as you may recall, I have a preference for zero Global
NFB amplifiers. Every bit of linearity is helpful.

Trevor Wilson


  #52   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



Trevor Wilson wrote:

"Eeyore" wrote
Trevor Wilson wrote:
"Arny Krueger" wrote

For one thing, modern output transistors have Ft up in the megahertz
range. Back in the 70s people were still struggling with output devices
with Ft in the KHz range.

**Nope. The Japanese had a goodly number of complementary devices, with
decent Voltage, current and SOA ratings which extended well into the tens
of MHz. By 1974, Hi-Rel had true Triple Diffused, TO3 devices with a 20MHz
fT and 20 Amps capacity.


Hi-Rel ?


**Yep, Hi-Rel. They were a small manufacturer back then.


I have a very vague idea I may have come across the name but in those
pre-internet (crikey - pre-fax even ! we still had a telex machine ) days one
was very constricted to using devices that local distributors carried. Typically
that meant RCA, Motorola and Thompson for highish power devices.


A mate discovered
them and used their devices for his amplifiers. Very impressive devices. I
have a spec sheet somewhere. When I locate it, I will scan and post, if you
like. Purely historical, of course. The devices have been unavailable for
quite awhile. FWIW, the numbers a ED203 and EB203.


I'd be interested to see those if you can find them thanks.


Exceptionally linear and very potent devices. I recall
them vividly, since Motorola had announced their intention to commit to
Epitaxial Base devices (ca. 2MHz), since it was impossible to produce
Triple Diffused devices (according to Motorola). Other manufacturers also

had
some impressive devices back then. The Marantz 2325 receiver used
2SA747/2SC1116 devices. These posessed an fT of 15MHz.


I recall being impressed by Sanken devices the first time I was designing
a serious amplifier around 1980. I settled for Hitachi lateral mosfets
instead. Obtainability was a serious issue back then.


**By 1980, Japanese transistors were much more easily available.


Not so easy IME.


The 2SD555 and it's complement were my favourites at the time.


I must check the data on that.


Having said all that, good gain linearity over a wide current range has
only been a relatively recent development.


With modern low-Z driver stages, current gain linearity is no longer an
issue the way it was.


**True enough, but, as you may recall, I have a preference for zero Global
NFB amplifiers. Every bit of linearity is helpful.


MY opinion is that zero global NFB is a classic example of navel gazing. There
is NO sensible basis for avoiding loop NFB totally. Indeed it you want very low
output Z it's the only way you can get it.

Personally, I use local and overall NFB in roughly similar amounts. It works
well.

Graham

  #53   Report Post  
Posted to rec.audio.tech
Mr.T Mr.T is offline
external usenet poster
 
Posts: 2,108
Default Stereo receievers: THD of .04 vs. .08


"Eeyore" wrote in message
...
Is it a "conventional Class AB bipolar design" ?


It is class AB, and it is bipolar.


It's not conventional.


Brilliant wording on your part, since you get to decide what is
"conventional" it seems.
Anything other than a basic text book design is non-conventional right?

MrT.


  #54   Report Post  
Posted to rec.audio.tech
Arny Krueger Arny Krueger is offline
external usenet poster
 
Posts: 17,262
Default Stereo receievers: THD of .04 vs. .08

"Mr.T" MrT@home wrote in message

"Eeyore" wrote
in message ...
Is it a "conventional Class AB bipolar design" ?


It is class AB, and it is bipolar.


It's not conventional.


Brilliant wording on your part, since you get to decide
what is "conventional" it seems.
Anything other than a basic text book design is
non-conventional right?



Right now, the standard textbook design would be Doug Self's Blameless
amplifier design.

http://www.dself.dsl.pipex.com/ampins/dipa/dipa.htm

The distortion that Graham is talking about, Self calls "gm doubling"
distortion.

"It is not generally appreciated that moving into Class-AB, by increasing
the quiescent current, does NOT simply trade efficiency for linearity. If
the output power is above the level at which Class-A operation can be
sustained, THD increases as the bias advances into AB operation. This is due
to so-called "gm-doubling" (ie the voltage-gain increase caused by both
devices conducting simultaneously in the centre of the output-voltage range,
in the Class-A region) putting edges into the distortion residual that
generate high-order harmonics much as under-biasing does. This vital fact is
little known, presumably because gm-doubling distortion is at a relatively
low level and is obscured in most amplifiers by other distortions."


  #55   Report Post  
Posted to rec.audio.tech
Arny Krueger Arny Krueger is offline
external usenet poster
 
Posts: 17,262
Default Stereo receievers: THD of .04 vs. .08

"Eeyore" wrote in
message
Arny Krueger wrote:

"Eeyore" wrote
in

You're aware that I've designed a fair number of
successful commercial designs I believe.


Yes, but you seem to be very worried about things that
just don't happen with the modern SS amps I've tested,
even fairly rugged cheap locomotives like the QSC USA
series.


The QSC USA series is a COMPLETE HEAP OF CRAP and
illustrates many classic design flaws and is a good
example of just how NOT to build a high performing
amplifier. The USA series is also most certainly not
*modern* by any reasonable understanding of the word.

If you can't hear the flaws of a USA series you must be
completely deaf.


Mere assertions with no supporting facts provided, neither theoretical or
experimental.

Not worthy of any serious technican, let alone an self-proclaimed engineer.
:-{




  #56   Report Post  
Posted to rec.audio.tech
Peter Larsen[_2_] Peter Larsen[_2_] is offline
external usenet poster
 
Posts: 724
Default Stereo receievers: THD of .04 vs. .08

Eeyore wrote:

Let that apply for the poweramp too ... not at an unrealistic
concept for home use ... and there is NO issue.


Would you seriously suggest an amp or so of quiescent current would
be acceptable ? It certainly WILL fix the problem of course.


Please enhance that point, I can see what you mean, but not how it
applies.


How it applies is to swamp the non-linear effect that results from
changes in output device transconductance when biased to lower values
of quiescent current.


If you're not familiar with that effect please say so.


My level of understanding is that the devices are always ON in the quiescent
current equivalent power range, ie "so" said.

Graham



Kind regards

Peter Larsen





  #57   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



"Mr.T" wrote:

"Eeyore" wrote

Is it a "conventional Class AB bipolar design" ?


It is class AB, and it is bipolar.


It's not conventional.


Brilliant wording on your part, since you get to decide what is
"conventional" it seems.


Anyone with reasonable experience of amplifier circuitry is likely to get
the gist of what I meant.


Anything other than a basic text book design is non-conventional right?


To elaborate, I'm primarily referring to the classic emitter follower
output stage with all the output devices biased into conduction at idle. I
already said that design tweaks can ameliorate cross-over THD, so yes, I'm
excluding examples of those.

As for the Crown Macro Tech, I suggest you look at the schematics and see
for yourself why I call call it unconventional.
http://www.crownaudio.com/pdf/legacy...schematics.zip

What else is it you want to know ?

Graham

  #58   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



Arny Krueger wrote:

"Mr.T" MrT@home wrote
"Eeyore" wrote

Is it a "conventional Class AB bipolar design" ?

It is class AB, and it is bipolar.

It's not conventional.


Brilliant wording on your part, since you get to decide
what is "conventional" it seems.
Anything other than a basic text book design is
non-conventional right?


Right now, the standard textbook design would be Doug Self's Blameless
amplifier design.

http://www.dself.dsl.pipex.com/ampins/dipa/dipa.htm

The distortion that Graham is talking about, Self calls "gm doubling"
distortion.

"It is not generally appreciated that moving into Class-AB, by increasing
the quiescent current, does NOT simply trade efficiency for linearity. If
the output power is above the level at which Class-A operation can be
sustained, THD increases as the bias advances into AB operation. This is due
to so-called "gm-doubling" (ie the voltage-gain increase caused by both
devices conducting simultaneously in the centre of the output-voltage range,
in the Class-A region) putting edges into the distortion residual that
generate high-order harmonics much as under-biasing does. This vital fact is
little known, presumably because gm-doubling distortion is at a relatively
low level and is obscured in most amplifiers by other distortions."


Whilst that is kinda true it's far from anything like the whole story.

As the amplifier output current transitions the crossover region, the
transconductance of both devices varies widely. This is inherent to transistor
operation. The transconductance is not a fixed number, it varies as the current
changes. That variation is most extreme in the crossover region where device
currents are low.

Graham


  #59   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



Arny Krueger wrote:

"Eeyore" wrote
Arny Krueger wrote:
"Eeyore" wrote

You're aware that I've designed a fair number of
successful commercial designs I believe.

Yes, but you seem to be very worried about things that
just don't happen with the modern SS amps I've tested,
even fairly rugged cheap locomotives like the QSC USA
series.


The QSC USA series is a COMPLETE HEAP OF CRAP and
illustrates many classic design flaws and is a good
example of just how NOT to build a high performing
amplifier. The USA series is also most certainly not
*modern* by any reasonable understanding of the word.

If you can't hear the flaws of a USA series you must be
completely deaf.


Mere assertions with no supporting facts provided, neither theoretical or
experimental.


From listening.


Not worthy of any serious technican, let alone an self-proclaimed engineer.
:-{


I'm aware you dismiss any listening tests that don't agree with your opinion.
Possibly you dismiss ALL listening tests that can't be replicated by deaf
people using the ABX method too ? What's the THD of a USA series at 100mW and
1W for example and what does the distortion analyser output (Self's Figs 14 a,b
,c ) look like (or provide a spectum analysis).


Graham


  #60   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



Peter Larsen wrote:

Eeyore wrote:

Let that apply for the poweramp too ... not at an unrealistic
concept for home use ... and there is NO issue.


Would you seriously suggest an amp or so of quiescent current would
be acceptable ? It certainly WILL fix the problem of course.


Please enhance that point, I can see what you mean, but not how it
applies.


How it applies is to swamp the non-linear effect that results from
changes in output device transconductance when biased to lower values
of quiescent current.


If you're not familiar with that effect please say so.


My level of understanding is that the devices are always ON in the quiescent
current equivalent power range, ie "so" said.


There's FAR more to it than that. The 'switching' ON and OFF isn't really the
problem per se.

Graham



  #61   Report Post  
Posted to rec.audio.tech
Peter Larsen[_2_] Peter Larsen[_2_] is offline
external usenet poster
 
Posts: 724
Default Stereo receievers: THD of .04 vs. .08

Eeyore wrote:

How it applies is to swamp the non-linear effect that results from
changes in output device transconductance when biased to lower
values of quiescent current.


If you're not familiar with that effect please say so.


My level of understanding is that the devices are always ON in the
quiescent current equivalent power range, ie "so" said.


There's FAR more to it than that. The 'switching' ON and OFF isn't
really the problem per se.


This is getting clearer from the current crop of new posts, very
interesting, thanks!

Graham



Kind regards

Peter Larsen



  #62   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



Peter Larsen wrote:

Eeyore wrote:

How it applies is to swamp the non-linear effect that results from
changes in output device transconductance when biased to lower
values of quiescent current.


If you're not familiar with that effect please say so.


My level of understanding is that the devices are always ON in the
quiescent current equivalent power range, ie "so" said.


There's FAR more to it than that. The 'switching' ON and OFF isn't
really the problem per se.


This is getting clearer from the current crop of new posts, very
interesting, thanks!


To elaborate, the cross-over region introduces a highly non-linear 'kink'
into the transfer characteristic of an amplifier at exactly the audibly
most critical point. No amount of obfuscation can deny that basic fact.
The smartest techniques to ameliorate the situation include 'moving' the
kink away from the zero signal region by various methods and 'stretching'
the kink so its effect at any given level is reduced.

Graham

  #63   Report Post  
Posted to rec.audio.tech
Arny Krueger Arny Krueger is offline
external usenet poster
 
Posts: 17,262
Default Stereo receievers: THD of .04 vs. .08

"Peter Larsen" wrote in message

Eeyore wrote:

How it applies is to swamp the non-linear effect that
results from changes in output device transconductance
when biased to lower values of quiescent current.


If you're not familiar with that effect please say so.


My level of understanding is that the devices are
always ON in the quiescent current equivalent power
range, ie "so" said.


There's FAR more to it than that. The 'switching' ON and
OFF isn't really the problem per se.


This is getting clearer from the current crop of new
posts, very interesting, thanks!


There are two effects:

(1) The transconductance of the active devices is a function of current,
often with the largest changes happening at low current levels.

(2) As the output stage cycles, there are periods when just one device or
just the other is turned on, and a smaller region around zero where both
devices are on. Even if the devices are each perfectly linear, this change
would still occur.


  #64   Report Post  
Posted to rec.audio.tech
Arny Krueger Arny Krueger is offline
external usenet poster
 
Posts: 17,262
Default Stereo receievers: THD of .04 vs. .08

"Eeyore" wrote in
message
Arny Krueger wrote:

"Eeyore" wrote
Arny Krueger wrote:
"Eeyore" wrote

You're aware that I've designed a fair number of
successful commercial designs I believe.

Yes, but you seem to be very worried about things that
just don't happen with the modern SS amps I've tested,
even fairly rugged cheap locomotives like the QSC USA
series.

The QSC USA series is a COMPLETE HEAP OF CRAP and
illustrates many classic design flaws and is a good
example of just how NOT to build a high performing
amplifier. The USA series is also most certainly not
*modern* by any reasonable understanding of the word.

If you can't hear the flaws of a USA series you must be
completely deaf.


Mere assertions with no supporting facts provided,
neither theoretical or experimental.


From listening.



Sighted evaluations.

I'm aware you dismiss any listening tests that don't
agree with your opinion.


If you mean that I tend to dismiss all listening tests, irregardless of
outcome, that aren't done in a professional manner, then I'm guilty as
charged.

Possibly you dismiss ALL
listening tests that can't be replicated by deaf people
using the ABX method too ?


What's this deaf people stuff Graham? You're talking trash! :-(

What's the THD of a USA series
at 100mW and 1W for example and what does the distortion
analyser output (Self's Figs 14 a,b ,c ) look like (or
provide a spectum analysis).


Tell us Graham. You're the professional engineer.

Why would you believe my measurements, Graham. In your view, I do ABX tests
with deaf listeners. If I'm that idiotic, why believe my measurements?

Bottom line Graham, you're running true to your - lots of lip, lots of crazy
accusations, and no reliable data. :-(


  #65   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



Arny Krueger wrote:

"Eeyore" wrote
Arny Krueger wrote:
"Eeyore" wrote
Arny Krueger wrote:
"Eeyore" wrote

You're aware that I've designed a fair number of
successful commercial designs I believe.

Yes, but you seem to be very worried about things that
just don't happen with the modern SS amps I've tested,
even fairly rugged cheap locomotives like the QSC USA
series.

The QSC USA series is a COMPLETE HEAP OF CRAP and
illustrates many classic design flaws and is a good
example of just how NOT to build a high performing
amplifier. The USA series is also most certainly not
*modern* by any reasonable understanding of the word.

If you can't hear the flaws of a USA series you must be
completely deaf.

Mere assertions with no supporting facts provided,
neither theoretical or experimental.


From listening.


Sighted evaluations.


Nothing explicitly wrong with those especially when the result isn't subtle
(which it wasn't). We are talking about GROSS differences here.

If I take a Ferrari out for a drive and then a Yugo and come back and say the
Ferrari was faster is that to be denied because it was a sighted test ?


I'm aware you dismiss any listening tests that don't
agree with your opinion.


If you mean that I tend to dismiss all listening tests, irregardless of
outcome, that aren't done in a professional manner, then I'm guilty as
charged.

Possibly you dismiss ALL
listening tests that can't be replicated by deaf people
using the ABX method too ?


What's this deaf people stuff Graham? You're talking trash! :-(

What's the THD of a USA series
at 100mW and 1W for example and what does the distortion
analyser output (Self's Figs 14 a,b ,c ) look like (or
provide a spectum analysis).


Tell us Graham. You're the professional engineer.

Why would you believe my measurements, Graham. In your view, I do ABX tests
with deaf listeners. If I'm that idiotic, why believe my measurements?

Bottom line Graham, you're running true to your - lots of lip, lots of crazy
accusations, and no reliable data. :-(


I'm not one for making great claims for being 'golden eared' and all that but it
is a FACT that certain people's hearing has more acuity than others. I've
recently been involved in tracing low level buzz and hum in a studio control
room along with 2 other techs and I was the only one of us 3 who could reliably
hear what the mix engineer was complaining about AND use that to trace and
eliminate the source.

Any testing using a panel will seriously degrade the likelihood of detecting
REAL audible differences by diluting the numbers whose hearing is up to the
task.

And ABX testing in its very philosophy seems to be more about detecting
differences that everyone can agree they hear rather than differences that only
a few can hear.

Graham




  #66   Report Post  
Posted to rec.audio.tech
Arny Krueger Arny Krueger is offline
external usenet poster
 
Posts: 17,262
Default Stereo receievers: THD of .04 vs. .08

"Eeyore" wrote in
message
Arny Krueger wrote:

"Eeyore" wrote
Arny Krueger wrote:
"Eeyore" wrote
Arny Krueger wrote:
"Eeyore"
wrote

You're aware that I've designed a fair number of
successful commercial designs I believe.

Yes, but you seem to be very worried about things
that just don't happen with the modern SS amps I've
tested, even fairly rugged cheap locomotives like
the QSC USA series.

The QSC USA series is a COMPLETE HEAP OF CRAP and
illustrates many classic design flaws and is a good
example of just how NOT to build a high performing
amplifier. The USA series is also most certainly not
*modern* by any reasonable understanding of the word.

If you can't hear the flaws of a USA series you must
be completely deaf.

Mere assertions with no supporting facts provided,
neither theoretical or experimental.

From listening.


Sighted evaluations.


Nothing explicitly wrong with those especially when the
result isn't subtle (which it wasn't). We are talking
about GROSS differences here.


All the golden ears say that. I've been hearing about gross differences due
to different interconnects and magic capacitors for about three decades.

If I take a Ferrari out for a drive and then a Yugo and
come back and say the Ferrari was faster is that to be
denied because it was a sighted test ?


Irrelevant to anyhing said so far.

I'm aware you dismiss any listening tests that don't
agree with your opinion.


If you mean that I tend to dismiss all listening tests,
irregardless of outcome, that aren't done in a
professional manner, then I'm guilty as charged.


Possibly you dismiss ALL
listening tests that can't be replicated by deaf people
using the ABX method too ?


What's this deaf people stuff Graham? You're talking
trash! :-(


What's the THD of a USA series
at 100mW and 1W for example and what does the distortion
analyser output (Self's Figs 14 a,b ,c ) look like (or
provide a spectum analysis).


Tell us Graham. You're the professional engineer.

Why would you believe my measurements, Graham. In your
view, I do ABX tests with deaf listeners. If I'm that
idiotic, why believe my measurements?

Bottom line Graham, you're running true to your - lots
of lip, lots of crazy accusations, and no reliable data.
:-(


I'm not one for making great claims for being 'golden
eared' and all that but it is a FACT that certain
people's hearing has more acuity than others.


No doubt. Young well-trained listeners have been trouncing me in ABX tests
for years. OTOH I can reliably detect all sorts of things that untrained
listeners of any age can't hear. One other advantage of training is that I
can relate what I hear to the right frequency band very well.

I've
recently been involved in tracing low level buzz and hum
in a studio control room along with 2 other techs and I
was the only one of us 3 who could reliably hear what the
mix engineer was complaining about AND use that to trace
and eliminate the source.


I have a headphone amp with lots of gain reserves for tasks like that.

Any testing using a panel will seriously degrade the
likelihood of detecting REAL audible differences by
diluting the numbers whose hearing is up to the task.


Depends on how you make up that panel.

And ABX testing in its very philosophy seems to be more
about detecting differences that everyone can agree they
hear rather than differences that only a few can hear.


Nonsense. The results of an ABX test self-administered by one listener can
be definitive. That's one reason why we invented ABX comparators - so that
we with trained ears could obtain reliable results all by ourselves.

Remember that when I started out with ABX I was in my 30s, and my hearing
and that of many of my friends and associates was far better than it is now.
The way us old dudes get the most sensitive results is by working with young
listeners and training them to hear small differences, and then basing many
of our conclusions on what they hear.


  #67   Report Post  
Posted to rec.audio.tech
Mr.T Mr.T is offline
external usenet poster
 
Posts: 2,108
Default Stereo receievers: THD of .04 vs. .08


"Eeyore" wrote in message
...
To elaborate, the cross-over region introduces a highly non-linear 'kink'
into the transfer characteristic of an amplifier at exactly the audibly
most critical point. No amount of obfuscation can deny that basic fact.
The smartest techniques to ameliorate the situation include 'moving' the
kink away from the zero signal region by various methods and 'stretching'
the kink so its effect at any given level is reduced.


And yet many modern amplifiers when properly measured at 1W (regardless of
max power rating, and price) still exhibit miniscule levels of THD and IMD.
The theory is correct, but not necessarily a real problem.

MrT.


  #68   Report Post  
Posted to rec.audio.tech
Mr.T Mr.T is offline
external usenet poster
 
Posts: 2,108
Default Stereo receievers: THD of .04 vs. .08


"Eeyore" wrote in message
...
Brilliant wording on your part, since you get to decide what is
"conventional" it seems.


Anyone with reasonable experience of amplifier circuitry is likely to get
the gist of what I meant.


Sure I understood your cop-out :-)

Anything other than a basic text book design is non-conventional right?


To elaborate, I'm primarily referring to the classic emitter follower
output stage with all the output devices biased into conduction at idle. I
already said that design tweaks can ameliorate cross-over THD, so yes, I'm
excluding examples of those.


Which is like saying there were problems with the model T Ford. So what?

As for the Crown Macro Tech, I suggest you look at the schematics and see
for yourself why I call call it unconventional.
http://www.crownaudio.com/pdf/legacy...schematics.zip


Sure, by historical standards.

What else is it you want to know ?


Why you bother with such a narrowly defined argument?

MrT.


  #69   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



"Mr.T" wrote:

"Eeyore" wrote

To elaborate, the cross-over region introduces a highly non-linear 'kink'
into the transfer characteristic of an amplifier at exactly the audibly
most critical point. No amount of obfuscation can deny that basic fact.
The smartest techniques to ameliorate the situation include 'moving' the
kink away from the zero signal region by various methods and 'stretching'
the kink so its effect at any given level is reduced.


And yet many modern amplifiers when properly measured at 1W (regardless of
max power rating, and price) still exhibit miniscule levels of THD and IMD.


Miniscule ? Care to give a number ? Can you even give an example. I've never
seen such a number quoted on a spec.


The theory is correct, but not necessarily a real problem.


Most likely corrected by massive negative fedback but I'll bet that the 100mW
THD is still MORE than the 'just prior to clip' THD.

Graham

  #70   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



"Mr.T" wrote:

"Eeyore" wrote

Brilliant wording on your part, since you get to decide what is
"conventional" it seems.


Anyone with reasonable experience of amplifier circuitry is likely to get
the gist of what I meant.


Sure I understood your cop-out :-)


It's not a 'cop-out'.


Anything other than a basic text book design is non-conventional right?


To elaborate, I'm primarily referring to the classic emitter follower
output stage with all the output devices biased into conduction at idle. I
already said that design tweaks can ameliorate cross-over THD, so yes, I'm
excluding examples of those.


Which is like saying there were problems with the model T Ford. So what?


Such designs are still in large scale production today. I would say they still
represent the 'norm', certainly in run-of-the-mill consumer hi-fi.

Graham



  #71   Report Post  
Posted to rec.audio.tech
Arny Krueger Arny Krueger is offline
external usenet poster
 
Posts: 17,262
Default Stereo receievers: THD of .04 vs. .08

"Mr.T" MrT@home wrote in message
u
"Eeyore" wrote
in message ...
To elaborate, the cross-over region introduces a highly
non-linear 'kink' into the transfer characteristic of an
amplifier at exactly the audibly most critical point. No
amount of obfuscation can deny that basic fact. The
smartest techniques to ameliorate the situation include
'moving' the kink away from the zero signal region by
various methods and 'stretching' the kink so its effect
at any given level is reduced.


And yet many modern amplifiers when properly measured at
1W (regardless of max power rating, and price) still
exhibit miniscule levels of THD and IMD. The theory is
correct, but not necessarily a real problem.


I suspect that Graham and the last AP test set he ever used have been
separated for some time. Me, I never had one, but figured out how to fake it
with really good sound cards. A fairly mediocre one in the right hands
could educate the man about modern reality.


  #72   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



Arny Krueger wrote:

"Mr.T" wrote
"Eeyore" wrote

To elaborate, the cross-over region introduces a highly
non-linear 'kink' into the transfer characteristic of an
amplifier at exactly the audibly most critical point. No
amount of obfuscation can deny that basic fact. The
smartest techniques to ameliorate the situation include
'moving' the kink away from the zero signal region by
various methods and 'stretching' the kink so its effect
at any given level is reduced.


And yet many modern amplifiers when properly measured at
1W (regardless of max power rating, and price) still
exhibit miniscule levels of THD and IMD. The theory is
correct, but not necessarily a real problem.


I suspect that Graham and the last AP test set he ever used have been
separated for some time.


Only about a year. I'm working on fixing that too !


Me, I never had one, but figured out how to fake it
with really good sound cards. A fairly mediocre one in the right hands
could educate the man about modern reality.


So what's the THD of your QSC USA amplifier @ 100mW / ~ 900mV with a speaker
load ? That's YOUR reality.

Graham


  #73   Report Post  
Posted to rec.audio.tech
Arny Krueger Arny Krueger is offline
external usenet poster
 
Posts: 17,262
Default Stereo receievers: THD of .04 vs. .08

"Eeyore" wrote in
message
Arny Krueger wrote:

"Mr.T" wrote
"Eeyore" wrote

To elaborate, the cross-over region introduces a highly
non-linear 'kink' into the transfer characteristic of
an amplifier at exactly the audibly most critical
point. No amount of obfuscation can deny that basic
fact. The smartest techniques to ameliorate the
situation include 'moving' the kink away from the zero
signal region by various methods and 'stretching' the
kink so its effect at any given level is reduced.

And yet many modern amplifiers when properly measured at
1W (regardless of max power rating, and price) still
exhibit miniscule levels of THD and IMD. The theory is
correct, but not necessarily a real problem.


I suspect that Graham and the last AP test set he ever
used have been separated for some time.


Only about a year. I'm working on fixing that too !


Me, I never had one, but figured out how to fake it
with really good sound cards. A fairly mediocre one in
the right hands could educate the man about modern
reality.


So what's the THD of your QSC USA amplifier @ 100mW / ~
900mV with a speaker load ? That's YOUR reality.


I rummaged around and found some test data I took some years back at a power
level of 1 watt into an 8 ohm resistive load.

Test 1: sine wave test with 1 KHz tone.

Stimulus 1 was 1 watt @ 1 KHz into 8 ohm resistive load:

All harmonics through approx 40 KHz were 96 dB or more below 1 watt.

Test 2: a standard dynamic range test signal ( 1 KHz test tone 60 dB below
FS = 1 watt)

Stimulus 2 was 1 KHz tone at 60 dB below FS = 1 watt. If FS was 1 watt,
then 60 dB down would be 1 microwatt, right?

All harmonics were 114 or more dB below FS. IOW 54 or more dB below the
test tone.

Actually, there was nothing I would properly call a harmonic. The baseline
was all noise or hash.






  #74   Report Post  
Posted to rec.audio.tech
Mr.T Mr.T is offline
external usenet poster
 
Posts: 2,108
Default Stereo receievers: THD of .04 vs. .08


"Eeyore" wrote in message
...
Miniscule ? Care to give a number ? Can you even give an example. I've

never
seen such a number quoted on a spec.


No need to take a manufacturers word for it, sound cards with 0.001% THD &
IMD are readily available for you to do your own FFT testing.
If you have never tested an amplifier yourself, why pretend you are an
expert? Especially when it is now so easy at the low power levels you are
worried about.

The theory is correct, but not necessarily a real problem.


I'll bet that the 100mW THD is still MORE than the 'just prior to clip'

THD.

Usually, but still insignificant, so what is the problem?

MrT.


  #75   Report Post  
Posted to rec.audio.tech
Mr.T Mr.T is offline
external usenet poster
 
Posts: 2,108
Default Stereo receievers: THD of .04 vs. .08


"Eeyore" wrote in message
...
Such designs are still in large scale production today. I would say they

still
represent the 'norm', certainly in run-of-the-mill consumer hi-fi.


Sure, and we all agree that most cheap consumer stereo or surround systems
are crap.
If you weren't talking about real HiFi then just say so.

MrT.




  #76   Report Post  
Posted to rec.audio.tech
Mr.T Mr.T is offline
external usenet poster
 
Posts: 2,108
Default Stereo receievers: THD of .04 vs. .08


"Arny Krueger" wrote in message
. ..
I suspect that Graham and the last AP test set he ever used have been
separated for some time.


If indeed he has ever used one, pity he didn't do some testing at the levels
he claims are a problem.
OTOH I doubt it since he keeps mentioning manufacturers published specs
rather than his own testing.

Me, I never had one, but figured out how to fake it
with really good sound cards.


The Lynx and Spectra-Lab is hardly "faking it" IMO :-)
More than adequate to prove the fallacy of his argument. But since he keeps
redefining it more narrowly each day, he will probably be right sooner or
later :-)

A fairly mediocre one in the right hands
could educate the man about modern reality.


I doubt it, first you have to WANT to learn.

MrT.


  #77   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



Arny Krueger wrote:

"Eeyore" wrote in

So what's the THD of your QSC USA amplifier @ 100mW / ~
900mV with a speaker load ? That's YOUR reality.


I rummaged around and found some test data I took some years back at a power
level of 1 watt into an 8 ohm resistive load.

Test 1: sine wave test with 1 KHz tone.

Stimulus 1 was 1 watt @ 1 KHz into 8 ohm resistive load:

All harmonics through approx 40 KHz were 96 dB or more below 1 watt.


In that case your test was defective.

Even by QSC's specs I'd expect no better than -60dB.

Graham

  #78   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



Arny Krueger wrote:

"Eeyore" wrote

So what's the THD of your QSC USA amplifier @ 100mW / ~
900mV with a speaker load ? That's YOUR reality.


I rummaged around and found some test data I took some years back at a power
level of 1 watt into an 8 ohm resistive load.

Test 1: sine wave test with 1 KHz tone.

Stimulus 1 was 1 watt @ 1 KHz into 8 ohm resistive load:

All harmonics through approx 40 KHz were 96 dB or more below 1 watt.

Test 2: a standard dynamic range test signal ( 1 KHz test tone 60 dB below
FS = 1 watt)

Stimulus 2 was 1 KHz tone at 60 dB below FS = 1 watt. If FS was 1 watt,
then 60 dB down would be 1 microwatt, right?

All harmonics were 114 or more dB below FS. IOW 54 or more dB below the
test tone.

Actually, there was nothing I would properly call a harmonic. The baseline
was all noise or hash.


You don't know how to run a proper test.

Therefore ALL the data you have EVER presented is suspect as are the conclusions
you have drawn from them.

Graham




  #79   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



"Mr.T" wrote:

"Eeyore" wrote:

Miniscule ? Care to give a number ? Can you even give an example. I've

never seen such a number quoted on a spec.

No need to take a manufacturers word for it, sound cards with 0.001% THD &
IMD are readily available for you to do your own FFT testing.


Right, so you're actually MAKING UP what you say. You've NEVER seen such a spec
NOR have you ever made a measurement to support your idiotic assertion.


If you have never tested an amplifier yourself


Just how ****ing STUPID are you.

I DESIGN amplifiers for ****'s sake ! Do you seriously think I don't run tests
on them ? I use Audio Precision test equipment.
http://ap.com/

Little IDIOT pricks like you and your stupid ideas are a blot on humanity.

Graham

  #80   Report Post  
Posted to rec.audio.tech
Eeyore Eeyore is offline
external usenet poster
 
Posts: 8,474
Default Stereo receievers: THD of .04 vs. .08



"Mr.T" wrote:

"Eeyore" wrote

Such designs are still in large scale production today. I would say they
still represent the 'norm', certainly in run-of-the-mill consumer hi-fi.


Sure, and we all agree that most cheap consumer stereo or surround systems
are crap.


You're an ignorant stupid prick with no experience of anything so SHUT UP.

The world is already infested with know-nothing idiots. We don't need another
one posting here.

Graham

Reply
Thread Tools
Display Modes

Posting Rules

Smilies are On
[IMG] code is On
HTML code is Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
FA: 360 Systems Model 2800 Programmable stereo Parametric EQ for stereo bus or mastering kellykevm Pro Audio 0 February 16th 07 03:54 AM
Escort '97 - Can I add Stereo RCA input plugs to my factory stereo? David Car Audio 0 November 29th 04 09:46 PM
AM receievers and tuners,1. Patrick Turner Vacuum Tubes 14 June 26th 04 07:26 PM
"Lost" left channel into stereo headphones through 3.0 / 3.5 mm stereo jack socket / plug Clive Long,UK General 0 June 9th 04 05:57 PM
Mazda Tribute - Stereo upgrades/mods, 7 speaker cd and cassette stereo - upgrd prairieboy Car Audio 0 March 9th 04 03:51 PM


All times are GMT +1. The time now is 09:19 PM.

Powered by: vBulletin
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Copyright ©2004-2024 AudioBanter.com.
The comments are property of their posters.
 

About Us

"It's about Audio and hi-fi"