Sign in with
Sign up | Sign in
Your question

Bypassing caps? How about with wire?

Tags:
  • Pro Audio
  • Audio
  • Product
Last response: in Home Audio
Share
Anonymous
July 24, 2004 9:28:40 AM

Archived from groups: rec.audio.pro (More info?)

This is a follow-on thread to the recent "Bypassing Caps" thread.

I too have a Soundcraft 200 series console waiting to be gone through. (A
reasonable beast to work on; lots of room on the boards and Harmon was
good about emailing the schematics.)

My plan has been to replace the TLO7x chips with better opamps that also
have zero (or darn near zero, <1mV) offset regardless of rail asymmetries,
and then hopefully getting rid of many coupling caps. (The OP275 is one
such device that usually holds zero offset under most power conditions.)

I've gone back and looked at several old threads related to this topic;
some have been quite interesting, but here are some additional questions:

Several threads had someone who was "about to" rechip a console, but I
didn't see any follow up as to which chip they finally liked best, or how
things turned out.

My current choices, based in part on personal experience (OP275) and also
several excellent posts by Monte McGuire (among others; sorry for not
mentioning all the great posters by name), are the OP275, 2604, and
OPA2132. There are other possibilities, but the prices start getting
rather high.

Any more recent thoughts on the best amps to use? My application is
primarily acoustic and classical recording/mixing; I'm not after a
particular "sound".

I will be adding power decoupling at each chip and will be watching
for oscillation, and am buying new linear power supplies. I will be upping
the wattage of the current limit resistors in each module to accommodate
the hungrier chips.

Any cautions about replacing coupling caps with a piece of wire? (PS
failure is one concern, but supposedly chips such as the 275s will
continue with zero offset even with a rail gone. I do plan to run
redundant supplies, coupled through diodes.)

Obviously I'll leave in the mic input couplers (pres are transformerless),
and perhaps one or two where the signal interfaces to the outside world,
but this console seems to have a large number of coupling caps in the
signal path.

Thanks for any new advice. I will make a report when this project is
concluded (the plan also includes adding EQ bypass switches in several
places and insert pre-post switching -- lots to do).

Frank
--
.

More about : bypassing caps wire

Anonymous
July 24, 2004 4:50:43 PM

Archived from groups: rec.audio.pro (More info?)

While I haven't rechipped & "de-capped" entire consoles, I do agree entirely with eleminating
coupling caps entirely where possible, replace with polypropylene and/or styrene film caps
where needed & as physically possible, and bypassing fresh Nichicon Muse lytics with styrene
films where there's not enough room for whole film.
I would strongly advocate using the OPA2134PA, with the OP275GP being my next choice.

--
Stephen Sank, Owner & Ribbon Mic Restorer
Talking Dog Transducer Company
http://stephensank.com
5517 Carmelita Drive N.E.
Albuquerque, New Mexico [87111]
505-332-0336
Auth. Nakamichi & McIntosh servicer
Payments preferred through Paypal.com
"Frank Stearns" <franks.pacifier.com@pacifier.net> wrote in message
news:10g3ss83vot7d34@corp.supernews.com...
> This is a follow-on thread to the recent "Bypassing Caps" thread.
>
> I too have a Soundcraft 200 series console waiting to be gone through. (A
> reasonable beast to work on; lots of room on the boards and Harmon was
> good about emailing the schematics.)
>
> My plan has been to replace the TLO7x chips with better opamps that also
> have zero (or darn near zero, <1mV) offset regardless of rail asymmetries,
> and then hopefully getting rid of many coupling caps. (The OP275 is one
> such device that usually holds zero offset under most power conditions.)
>
> I've gone back and looked at several old threads related to this topic;
> some have been quite interesting, but here are some additional questions:
>
> Several threads had someone who was "about to" rechip a console, but I
> didn't see any follow up as to which chip they finally liked best, or how
> things turned out.
>
> My current choices, based in part on personal experience (OP275) and also
> several excellent posts by Monte McGuire (among others; sorry for not
> mentioning all the great posters by name), are the OP275, 2604, and
> OPA2132. There are other possibilities, but the prices start getting
> rather high.
>
> Any more recent thoughts on the best amps to use? My application is
> primarily acoustic and classical recording/mixing; I'm not after a
> particular "sound".
>
> I will be adding power decoupling at each chip and will be watching
> for oscillation, and am buying new linear power supplies. I will be upping
> the wattage of the current limit resistors in each module to accommodate
> the hungrier chips.
>
> Any cautions about replacing coupling caps with a piece of wire? (PS
> failure is one concern, but supposedly chips such as the 275s will
> continue with zero offset even with a rail gone. I do plan to run
> redundant supplies, coupled through diodes.)
>
> Obviously I'll leave in the mic input couplers (pres are transformerless),
> and perhaps one or two where the signal interfaces to the outside world,
> but this console seems to have a large number of coupling caps in the
> signal path.
>
> Thanks for any new advice. I will make a report when this project is
> concluded (the plan also includes adding EQ bypass switches in several
> places and insert pre-post switching -- lots to do).
>
> Frank
> --
> .
Anonymous
July 24, 2004 8:34:45 PM

Archived from groups: rec.audio.pro (More info?)

"Frank Stearns"

> I too have a Soundcraft 200 series console waiting to be gone through. (A
> reasonable beast to work on; lots of room on the boards and Harmon was
> good about emailing the schematics.)
>
> My plan has been to replace the TLO7x chips with better opamps that also
> have zero (or darn near zero, <1mV) offset regardless of rail asymmetries,
> and then hopefully getting rid of many coupling caps. (The OP275 is one
> such device that usually holds zero offset under most power conditions.)
>

** Be aware of the input bias current for the (bi-polar) OP275 is up to
400 nA. If the input resistance is say 100kohms - this means 40 mV input
offset is possible. The input bias current of the humble (bi-fet) TL072 is
about 1000 times less.


>
> Obviously I'll leave in the mic input couplers (pres are transformerless),
> and perhaps one or two where the signal interfaces to the outside world,
> but this console seems to have a large number of coupling caps in the
> signal path.
>


** It actually has remarkably few ( per channel) - they are mostly only
fitted prior to switches and pots / faders to eliminate clicks and noise
when rotated / moved. Bi-polar electros do this job very well without the
THD issues at low frequencies of polarised electros.





............. Phil
Related resources
Anonymous
July 24, 2004 8:34:46 PM

Archived from groups: rec.audio.pro (More info?)

Phil Allison <philallison@tpg.com.au> wrote:
>"Frank Stearns"
>
>> I too have a Soundcraft 200 series console waiting to be gone through. (A
>> reasonable beast to work on; lots of room on the boards and Harmon was
>> good about emailing the schematics.)
>>
>> My plan has been to replace the TLO7x chips with better opamps that also
>> have zero (or darn near zero, <1mV) offset regardless of rail asymmetries,
>> and then hopefully getting rid of many coupling caps. (The OP275 is one
>> such device that usually holds zero offset under most power conditions.)
>>
>
> ** Be aware of the input bias current for the (bi-polar) OP275 is up to
>400 nA. If the input resistance is say 100kohms - this means 40 mV input
>offset is possible. The input bias current of the humble (bi-fet) TL072 is
>about 1000 times less.

Much as I dislike Phil's quoting style, he has a point here. I'll also
put in a good word for the sound quality of the TL072, which is really
not bad at all in a well-designed circuit.

I suggest you put sockets on one channel strip, then try a couple different
types of ICs in different spots and get a sense of how they sound and what
they do. BUT, if your intention is to later eliminate coupling capacitors,
you may want to be careful about offset of op-amps.

You may find that the TL072 is the best-sounding chip in at least some of
the locations. It's certainly among the best bifet chips that you can get
these days now that the Motorola stuff is all discontinued.

Do you care if the controls pop? If you can live with clicking and popping,
say when you enable and disable the EQ, you can live with substantially more
DC offset than the console was originally designed with. Do keep DC off of
the pots, of course. Of course, if you're using it as a PA console, the
popping is a major problem....
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Anonymous
July 24, 2004 8:34:47 PM

Archived from groups: rec.audio.pro (More info?)

"Scott Dorsey" <kludge@panix.com> wrote in message
news:cdto5r$mu2$1@panix3.panix.com...

> You may find that the TL072 is the best-sounding chip in at least some of
> the locations. It's certainly among the best bifet chips that you can get
> these days now that the Motorola stuff is all discontinued.

Be careful about loading, though; the TL07x has an insipid output stage, and
it hates driving anything under 10k (don't forget to include the feedback
network when you calculate total load). It's really happiest with a 50k load
or higher.

Peace,
Paul
Anonymous
July 24, 2004 8:34:48 PM

Archived from groups: rec.audio.pro (More info?)

"Paul Stamler" <pstamlerhell@pobox.com> wrote in message
news:7gvMc.316443$Gx4.46615@bgtnsc04-news.ops.worldnet.att.net
> "Scott Dorsey" <kludge@panix.com> wrote in message
> news:cdto5r$mu2$1@panix3.panix.com...
>
>> You may find that the TL072 is the best-sounding chip in at least
>> some of the locations. It's certainly among the best bifet chips
>> that you can get these days now that the Motorola stuff is all
>> discontinued.
>
> Be careful about loading, though; the TL07x has an insipid output
> stage, and it hates driving anything under 10k (don't forget to
> include the feedback network when you calculate total load). It's
> really happiest with a 50k load or higher.

Not to mention the opportunities for latch-up.

http://focus.ti.com/docs/prod/folders/print/tl072.html seems to claim that
the current parts are "latchup-free". That's a tacit admission that there
was once a problem. I wonder when they fixed it?
July 24, 2004 9:57:36 PM

Archived from groups: rec.audio.pro (More info?)

"Stephen Sank" <bk11@thuntek.net> wrote in message news:<cdub0d$uco$1@reader2.nmix.net>...
> While I haven't rechipped & "de-capped" entire consoles, I do agree entirely with eleminating
> coupling caps entirely where possible, replace with polypropylene and/or styrene film caps
> where needed & as physically possible, and bypassing fresh Nichicon Muse lytics with styrene
> films where there's not enough room for whole film.
> I would strongly advocate using the OPA2134PA, with the OP275GP being my next choice.
>

How about somebody do this to say one channel only and then post a
before and after pic of a waveform or spectrum analyzer or other
measurable difference you think this makes...

Lets see a spectun alayzer pic of low frequency distortion caused by
electrolytic caps and see it go away when the cap is replaced. (And
I'm not talking about replacing a defective cap that is leaky causing
a shift of the bias point. If you want to tell me that electolytic
caps have a poor reliability record, OK, but I would ask that you show
me evidence of a non leaky electrolytic cap causing distortion.)

Lets get some science going here.

Mark
Anonymous
July 24, 2004 10:36:54 PM

Archived from groups: rec.audio.pro (More info?)

Paul Stamler <pstamlerhell@pobox.com> wrote:
>"Scott Dorsey" <kludge@panix.com> wrote in message
>news:cdto5r$mu2$1@panix3.panix.com...
>
>> You may find that the TL072 is the best-sounding chip in at least some of
>> the locations. It's certainly among the best bifet chips that you can get
>> these days now that the Motorola stuff is all discontinued.
>
>Be careful about loading, though; the TL07x has an insipid output stage, and
>it hates driving anything under 10k (don't forget to include the feedback
>network when you calculate total load). It's really happiest with a 50k load
>or higher.

Hmm... how is the output stage biased? Is the current-drive poor because
the output transistors are too small, or because they are biased way linear?

The Motorolas had no current drive ability to speak of, since the output
stage was running into class A.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Anonymous
July 24, 2004 11:49:09 PM

Archived from groups: rec.audio.pro (More info?)

Thank you, gentlemen, for many excellent ideas and observations regarding
rechipping and decapping my old Soundcraft 200.

This project is still 30-60 days out; will try to post some "here's what I
finally did" followups then.

Frank
--
.
Anonymous
July 25, 2004 12:15:43 AM

Archived from groups: rec.audio.pro (More info?)

One of the most enormous failings of the audio electronics industry through the years has been
a direct result of people like you, Mark. "If it looks good on paper(and on test equipment),
it has to sound good." One of the WORST sounding power amps I have ever heard was the Yamaha
M70. THD & TIM were rated 0.002% from 250mw to 250w at 8, 4 or 2 ohms. Slew rate 200V/us.
S/N 105db below rated output. Drew an absolutely perfect squarewave at any audio frequency &
any power level. But the amp, with music, sounded absolutely HORRIBLE. Regardless of speaker,
preamp, cables or any other peripherals, the upper treble was downright screechy, and despite
the apparently lively treble, detail in the high treble was obsured badly. When I switched
from that amp(which I regrettably owned for a while) to a much older McIntosh MC2105, which is
also a transistor amp, and which has far "worse" measured performance in every way, it was such
a VAST improvement in sound in every describable way, I couldn't believe I'd ever been able to
stand listening to the Yamaha.
In short, if you are actually trying to assert that any non-defective electrolytic sounds the
same as any other non-defective electrolytic, you are either completely lacking in actual
listening experience, have quite a lot of hearing loss, or are self deluded. Even
*manufacturer's* published data will plainly spell out that different grades of lytics can have
very different ESR curves, with most published ESR curves showing very significant upswing of
resistance well within the audio band, often even starting below 10kHz. If you think such
differences ALONE aren't enough to audibly effect the sound of lytics used as coupling caps,
then you defy even your own "If it looks good on paper, it must sound good" apparent mentality.
BTW, where did I mention "low frequency distortion" in my previous post? Not that there ain't
any with lytic coupling, but I did not state anything to that effect.

--
Stephen Sank, Owner & Ribbon Mic Restorer
Talking Dog Transducer Company
http://stephensank.com
5517 Carmelita Drive N.E.
Albuquerque, New Mexico [87111]
505-332-0336
Auth. Nakamichi & McIntosh servicer
Payments preferred through Paypal.com
"Mark" <makolber@yahoo.com> wrote in message
news:3367f36e.0407241657.6bd385fb@posting.google.com...
> "Stephen Sank" <bk11@thuntek.net> wrote in message news:<cdub0d$uco$1@reader2.nmix.net>...
> > While I haven't rechipped & "de-capped" entire consoles, I do agree entirely with
eleminating
> > coupling caps entirely where possible, replace with polypropylene and/or styrene film caps
> > where needed & as physically possible, and bypassing fresh Nichicon Muse lytics with
styrene
> > films where there's not enough room for whole film.
> > I would strongly advocate using the OPA2134PA, with the OP275GP being my next choice.
> >
>
> How about somebody do this to say one channel only and then post a
> before and after pic of a waveform or spectrum analyzer or other
> measurable difference you think this makes...
>
> Lets see a spectun alayzer pic of low frequency distortion caused by
> electrolytic caps and see it go away when the cap is replaced. (And
> I'm not talking about replacing a defective cap that is leaky causing
> a shift of the bias point. If you want to tell me that electolytic
> caps have a poor reliability record, OK, but I would ask that you show
> me evidence of a non leaky electrolytic cap causing distortion.)
>
> Lets get some science going here.
>
> Mark
Anonymous
July 25, 2004 9:41:25 AM

Archived from groups: rec.audio.pro (More info?)

"Scott Dorsey" <kludge@panix.com> wrote in message
news:cduoa6$s68$1@panix3.panix.com...
> Paul Stamler <pstamlerhell@pobox.com> wrote:
> >Be careful about loading, though; the TL07x has an insipid output stage,
and
> >it hates driving anything under 10k (don't forget to include the feedback
> >network when you calculate total load). It's really happiest with a 50k
load
> >or higher.
>
> Hmm... how is the output stage biased? Is the current-drive poor because
> the output transistors are too small, or because they are biased way
linear?

IIRC (it's been years since the last set of lab tests I did), the output
transistors are biased quite low. The harmonic distortion was ugly and
spiky-looking, much nastier than, say, an OPA604 or OPA134 (these are the
one-amp-to-a-package versions of the 2604 and 2134). As Monte noted, they're
also quite susceptible to high distortion when driven by high impedances at
high levels, in voltage-follower situations.

At lower impedances, driving higher impedance loads, they can be decent.
Unlike some other users, I've never had any problem with latchup. But you do
have to be careful about how you apply them.

Peace,
Paul
Anonymous
July 25, 2004 4:09:33 PM

Archived from groups: rec.audio.pro (More info?)

"Arny Krueger" <
> "Paul Stamler"
> >
> > Be careful about loading, though; the TL07x has an insipid output
> > stage, and it hates driving anything under 10k (don't forget to
> > include the feedback network when you calculate total load). It's
> > really happiest with a 50k load or higher.


** The TL072 in fact drives loads down to around 2 kohms with 0.001 % THD.

See: http://www.dself.dsl.pipex.com/ampins/webbop/072.htm

Several other op-amps are detailed on similar pages at the same site -
well worth a read.


> Not to mention the opportunities for latch-up.
>
> http://focus.ti.com/docs/prod/folders/print/tl072.html seems to claim that
> the current parts are "latchup-free". That's a tacit admission that there
> was once a problem. I wonder when they fixed it?


** The TL0xx series op-amps were described by TI as "latch up free" from
the day they were released - I have a 1977 brochure which states that.

However, they will phase invert the output if the inputs are driven close
to a supply rail making for a *very nasty* overdrive sound.




............... Phil
Anonymous
July 25, 2004 4:09:34 PM

Archived from groups: rec.audio.pro (More info?)

In article <2mgj2qFmm4egU1@uni-berlin.de>,
"Phil Allison" <philallison@tpg.com.au> wrote:

> "Arny Krueger" <
> > "Paul Stamler"
> > >
> > > Be careful about loading, though; the TL07x has an insipid output
> > > stage, and it hates driving anything under 10k (don't forget to
> > > include the feedback network when you calculate total load). It's
> > > really happiest with a 50k load or higher.
>
>
> ** The TL072 in fact drives loads down to around 2 kohms with 0.001 % THD.

If you think that's good enough, then have at it. Chips like the
OPA2132 can do a lot better and the end result sounds that way too.
Also check the THD at something other than 1KHz. It's not so pretty as
you go further up.

One of the more annoying aspects of a JFET input amp is that it ends up
producing a lot of second harmonic when it is asked to run with lots of
input signal presented as a high impedance common mode signal, basically
when it is run as a noninverting stage with a high source Z, something
you'd think a JFET would be perfect for.

IIRC, Walt Jung built a circuit that bootstrapped the op amp's substrate
(V-) with the common mode input voltage to nullify the effects of the
nonlinear drain to substrate capacitance (that causes the common mode
problems) and a bunch of distortion went away. The other solution is to
use a low source Z, in which case you don't need a JFET input in the
first place.


Regards,

Monte McGuire
monte.mcguire@verizon.net
Anonymous
July 25, 2004 4:09:35 PM

Archived from groups: rec.audio.pro (More info?)

On Sun, 25 Jul 2004 05:09:02 GMT, Monte McGuire
<monte.mcguire@verizon.net> wrote:

>IIRC, Walt Jung built a circuit that bootstrapped the op amp's substrate
>(V-) with the common mode input voltage to nullify the effects of the
>nonlinear drain to substrate capacitance (that causes the common mode
>problems) and a bunch of distortion went away. The other solution is to
>use a low source Z, in which case you don't need a JFET input in the
>first place.

Yet another is to cascode the input diff pair(s). Not too easy
for a given monolythic op-amp, natch.

Chris Hornbeck
"Vote or Die" - P. Diddy
Anonymous
July 25, 2004 4:09:35 PM

Archived from groups: rec.audio.pro (More info?)

"Monte McGuire" <monte.mcguire@verizon.net> wrote in message
news:monte.mcguire-6A1C1D.01090525072004@news.verizon.net


> IIRC, Walt Jung built a circuit that bootstrapped the op amp's
> substrate (V-) with the common mode input voltage to nullify the
> effects of the nonlinear drain to substrate capacitance (that causes
> the common mode problems) and a bunch of distortion went away. The
> other solution is to use a low source Z, in which case you don't need
> a JFET input in the first place.

This seems to be described in

http://www.analog.com/UploadedFiles/Application_Notes/7...

The primary means of dealing with this problem that was recommended was to
drive both inputs of the opamp with similar source impedances.
Anonymous
July 25, 2004 6:33:13 PM

Archived from groups: rec.audio.pro (More info?)

makolber@yahoo.com (Mark) writes:

> "Stephen Sank" <bk11@thuntek.net> wrote in message news:<cdub0d$uco$1@reader2.nmix.net>...
>> While I haven't rechipped & "de-capped" entire consoles, I do agree entirely with eleminating
>> coupling caps entirely where possible, replace with polypropylene and/or styrene film caps
>> where needed & as physically possible, and bypassing fresh Nichicon Muse lytics with styrene
>> films where there's not enough room for whole film.
>> I would strongly advocate using the OPA2134PA, with the OP275GP being my next choice.
>>

> How about somebody do this to say one channel only and then post a
> before and after pic of a waveform or spectrum analyzer or other
> measurable difference you think this makes...

Get theee to a library! Wireless World did this, and more some years ago.

--
Paul Repacholi 1 Crescent Rd.,
+61 (08) 9257-1001 Kalamunda.
West Australia 6076
comp.os.vms,- The Older, Grumpier Slashdot
Raw, Cooked or Well-done, it's all half baked.
EPIC, The Architecture of the future, always has been, always will be.
Anonymous
July 25, 2004 7:40:07 PM

Archived from groups: rec.audio.pro (More info?)

"Monte McGuire"
> "Phil Allison" :
> > "Paul Stamler"
> > > >
> > > > Be careful about loading, though; the TL07x has an insipid output
> > > > stage, and it hates driving anything under 10k (don't forget to
> > > > include the feedback network when you calculate total load). It's
> > > > really happiest with a 50k load or higher.
> >
> >
> > ** The TL072 in fact drives loads down to around 2 kohms with 0.001 %
THD.
>
> If you think that's good enough, then have at it. Chips like the
> OPA2132 can do a lot better and the end result sounds that way too.


** Got DBT results to prove that ??

Thought not.


> Also check the THD at something other than 1KHz. It's not so pretty as
> you go further up.


** 0.003% at 10 kHz with a 3.3k ohms load sure aint ugly.


>
> IIRC, Walt Jung built a circuit that bootstrapped the op amp's substrate
> (V-) with the common mode input voltage to nullify the effects of the
> nonlinear drain to substrate capacitance (that causes the common mode
> problems) and a bunch of distortion went away. The other solution is to
> use a low source Z, in which case you don't need a JFET input in the
> first place.


** As I already pointed out, the very low input bias current is major plus
with the TL072 almost 10,000 times less than good bi-polar op-amps like the
5532. If you are after low *output* offset voltages then fet op-amps are
the way to go.

The DC source resistance seen by the inputs of an op-amp is a seperate
issue from the signal's AC source impedance.



............ Phil
July 25, 2004 10:58:33 PM

Archived from groups: rec.audio.pro (More info?)

"Stephen Sank" <bk11@thuntek.net> wrote in message news:<cdv52t$j08$1@reader2.nmix.net>...
> One of the most enormous failings of the audio electronics industry through the years has been
> a direct result of people like you, Mark. "If it looks good on paper(and on test equipment),
> it has to sound good." One of the WORST sounding power amps I have ever heard was the Yamaha
> M70. THD & TIM were rated 0.002% from 250mw to 250w at 8, 4 or 2 ohms. Slew rate 200V/us.
> S/N 105db below rated output. Drew an absolutely perfect squarewave at any audio frequency &
> any power level. But the amp, with music, sounded absolutely HORRIBLE. Regardless of speaker,
> preamp, cables or any other peripherals, the upper treble was downright screechy, and despite
> the apparently lively treble, detail in the high treble was obsured badly. When I switched
> from that amp(which I regrettably owned for a while) to a much older McIntosh MC2105, which is
> also a transistor amp, and which has far "worse" measured performance in every way, it was such
> a VAST improvement in sound in every describable way, I couldn't believe I'd ever been able to
> stand listening to the Yamaha.


Why did the Yamaha sound horrible Stephen. Were you able to make any
measurment that would account for what you (thought you) heard? Was
it crossover distortion, high frequency intermod, or what, you can't
just say, I hear it but I don't know what it is.

I also understand that some people think their speakers sound better
when they lift the speakers cable off the nylon carpet. Something
about lossy dielectrics.

Science please. If you think you hear something, fine, maybe you do,
but the job isn't done until you identify what it is that you hear.

Mark
Anonymous
July 26, 2004 6:31:33 AM

Archived from groups: rec.audio.pro (More info?)

On 25 Jul 2004 18:58:33 -0700, makolber@yahoo.com (Mark) wrote:

> you can't
>just say, I hear it but I don't know what it is.

Really? Why not?

Chris Hornbeck
"Vote or Die" - P. Diddy
Anonymous
July 26, 2004 5:40:38 PM

Archived from groups: rec.audio.pro (More info?)

On Mon, 26 Jul 2004 05:40:10 -0400, "Arny Krueger" <arnyk@hotpop.com>
wrote:

>In each of these situations, the case that the UUT measures good and sounds
>bad is in essence, based on ignorance.

I believe that's what Stephen said, too.

Chris Hornbeck
"Vote or Die" - P. Diddy
Anonymous
July 26, 2004 8:33:42 PM

Archived from groups: rec.audio.pro (More info?)

"Stephen Sank" <bk11@thuntek.net> wrote in message
news:cdv52t$j08$1@reader2.nmix.net...
> One of the most enormous failings of the audio electronics industry
through the years has been
> a direct result of people like you, Mark. "If it looks good on
paper(and on test equipment),
> it has to sound good." One of the WORST sounding power amps I have
ever heard was the Yamaha
> M70. THD & TIM were rated 0.002% from 250mw to 250w at 8, 4 or 2
ohms. Slew rate 200V/us.
> S/N 105db below rated output. Drew an absolutely perfect squarewave
at any audio frequency &
> any power level. But the amp, with music, sounded absolutely
HORRIBLE. Regardless of speaker,
> preamp, cables or any other peripherals, the upper treble was
downright screechy, and despite
> the apparently lively treble, detail in the high treble was obsured
badly. When I switched
> from that amp(which I regrettably owned for a while) to a much older
McIntosh MC2105, which is
> also a transistor amp, and which has far "worse" measured
performance in every way, it was such
> a VAST improvement in sound in every describable way, I couldn't
believe I'd ever been able to
> stand listening to the Yamaha.

Judging by the words in all caps, this amplifier must have been
unacceptable to anyone. Which brings up the question: Why did Yamaha
put this abomination on the market, and why did so many people buy it?

Another question might be: How would you go about designing an
amplifier that meets the specs you mentioned above, yet sounds
"HORRIBLE"? I've mulled the situation over, and I can't think of any
way to accomplish this feat.

Norm Strong
Anonymous
July 26, 2004 8:33:43 PM

Archived from groups: rec.audio.pro (More info?)

"normanstrong" <normanstrong@comcast.net> wrote in message
news:GnaNc.165810$IQ4.101918@attbi_s02

> Another question might be: How would you go about designing an
> amplifier that meets the specs you mentioned above, yet sounds
> "HORRIBLE"? I've mulled the situation over, and I can't think of any
> way to accomplish this feat.

The obvious way is to slip it past the tests. For example, make an amp that
works great with resistive loads, but clips at the first sign of a reactive
load. That's two transistors and a few resistors, less if asymetrical bad
behavior is desired, even less if the parts are already there and all you
need to do is misdesign them. You might call the results, a "DC-300". It
would be your err, *crowning* achievement as it were. ;-)

A built-in grounding problem might also be possible.
Anonymous
July 26, 2004 8:33:43 PM

Archived from groups: rec.audio.pro (More info?)

normanstrong <normanstrong@comcast.net> wrote:
>
>Judging by the words in all caps, this amplifier must have been
>unacceptable to anyone. Which brings up the question: Why did Yamaha
>put this abomination on the market, and why did so many people buy it?

I dunno, but it wouldn't be the first time Yamaha has made something that
sounded really horrible but sold well.

>Another question might be: How would you go about designing an
>amplifier that meets the specs you mentioned above, yet sounds
>"HORRIBLE"? I've mulled the situation over, and I can't think of any
>way to accomplish this feat.

By taking the attitude that circuit linearity isn't really very important
since all flaws can be cured with negative feedback. This seemed to be
a very common philosophy in the 1970s, and it resulted in a lot of very
disasterously bad-sounding gear. The terrible sound of some of the early
transistor gear is part of what fueled the whole tube audio renaissance
a decade and a half later.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Anonymous
July 26, 2004 8:33:43 PM

Archived from groups: rec.audio.pro (More info?)

Remembering the schematic as best as I can in my head, one glaring aspect was the rather
numerous stages that had small(pf values) caps in the immediate feedback look, which strongly
suggested that the Yamaha people had some difficulties with the HF stability of the amp. In my
experience, when you see things like that in an amp circuit, regardless of amount of global
feedback, the amp will not yield the best sound. In fact, in this particular line of Yamaha
amps(M50 & M70, carried fundamentally over to M40/M60/M80, and then M45/65/85), if you made the
mistake of replacing the predriver transistors with anything but the exact originals, the amp
would go ballistic with ultrasonic oscillation & blow up.
In regard to how they sold the darn things in large numbers, the plain fact is that a lot of
people get impressed by systems that have hyped treble(including a hell of a lot of recording
people, sadly, judging by the average cd release), and will buy something in a hifi store(an
environment usually acoustically absorbtive of treble) after a few minutes listening. How many
of these folks found their ears growing tired of the sound they ended up with is something that
might well be judged by surveying the audio listings on ebay.

--
Stephen Sank, Owner & Ribbon Mic Restorer
Talking Dog Transducer Company
http://stephensank.com
5517 Carmelita Drive N.E.
Albuquerque, New Mexico [87111]
505-332-0336
Auth. Nakamichi & McIntosh servicer
Payments preferred through Paypal.com
"normanstrong" <normanstrong@comcast.net> wrote in message
news:GnaNc.165810$IQ4.101918@attbi_s02...
>
> "Stephen Sank" <bk11@thuntek.net> wrote in message
> news:cdv52t$j08$1@reader2.nmix.net...
> > One of the most enormous failings of the audio electronics industry
> through the years has been
> > a direct result of people like you, Mark. "If it looks good on
> paper(and on test equipment),
> > it has to sound good." One of the WORST sounding power amps I have
> ever heard was the Yamaha
> > M70. THD & TIM were rated 0.002% from 250mw to 250w at 8, 4 or 2
> ohms. Slew rate 200V/us.
> > S/N 105db below rated output. Drew an absolutely perfect squarewave
> at any audio frequency &
> > any power level. But the amp, with music, sounded absolutely
> HORRIBLE. Regardless of speaker,
> > preamp, cables or any other peripherals, the upper treble was
> downright screechy, and despite
> > the apparently lively treble, detail in the high treble was obsured
> badly. When I switched
> > from that amp(which I regrettably owned for a while) to a much older
> McIntosh MC2105, which is
> > also a transistor amp, and which has far "worse" measured
> performance in every way, it was such
> > a VAST improvement in sound in every describable way, I couldn't
> believe I'd ever been able to
> > stand listening to the Yamaha.
>
> Judging by the words in all caps, this amplifier must have been
> unacceptable to anyone. Which brings up the question: Why did Yamaha
> put this abomination on the market, and why did so many people buy it?
>
> Another question might be: How would you go about designing an
> amplifier that meets the specs you mentioned above, yet sounds
> "HORRIBLE"? I've mulled the situation over, and I can't think of any
> way to accomplish this feat.
>
> Norm Strong
>
>
Anonymous
July 26, 2004 10:18:19 PM

Archived from groups: rec.audio.pro (More info?)

In article <GnaNc.165810$IQ4.101918@attbi_s02> normanstrong@comcast.net writes:

> Another question might be: How would you go about designing an
> amplifier that meets the specs you mentioned above, yet sounds
> "HORRIBLE"? I've mulled the situation over, and I can't think of any
> way to accomplish this feat.

One thing is to not specify the things that make it sound horrible, or
measure them in ways that make bad things look good. An easy on, if
there's a 40 or 50 kHz oscillation (and this really can happen with a
badly designed amplifier) is to measure of the bandwidth of 20 Hz to
20 kHz. While I wouldn't say "it sounds horrible" I've run across some
gear that has noise specificed as "A weighted" which pretty thoroughly
attenuates line frequency hum.

I couldn't tell from Stephen's message whether "it measured" or he
actually made measurements himself, trying determine if he could
analyze the problem he was hearing.

--
I'm really Mike Rivers (mrivers@d-and-d.com)
However, until the spam goes away or Hell freezes over,
lots of IP addresses are blocked from this system. If
you e-mail me and it bounces, use your secret decoder ring
and reach me here: double-m-eleven-double-zero at yahoo
Anonymous
July 27, 2004 12:07:59 AM

Archived from groups: rec.audio.pro (More info?)

On Mon, 26 Jul 2004 16:33:42 GMT, "normanstrong"
<normanstrong@comcast.net> wrote:

>Another question might be: How would you go about designing an
>amplifier that meets the specs you mentioned above, yet sounds
>"HORRIBLE"? I've mulled the situation over, and I can't think of any
>way to accomplish this feat.

In addition to what Arny and Scott have mentioned, a problem
not so long ago was to specify performance only at full output.

The assumption that all flaws will decrease with signal level
has become *less* true with most modern improvements.

Full signal performance spec's are almost completely
irrelevant for current designs.

Chris Hornbeck
"Vote or Die" - P. Diddy
July 27, 2004 12:11:57 AM

Archived from groups: rec.audio.pro (More info?)

My point is that if you claim that an amp sounds bad, or changing caps
sounds good, I think you at least owe the folks to whom you are making
the claim, a reason for your claim, a reason backed up by
measurements. Somthing like the bad bypass cap caused an ultrasonic
oscillation that created terrible high frequency IM distortion (which
is easily measured).

I guess our fundamental disagreement is that I beleive that any
difference that can be heard can also be measured. With FFT
analyzers that can measure 100 dB down and have sub Hz resolution BW,
it is my firm belief that anything you can hear can be measured. It
may be difficult to identify the particular parameter to measure to
account for what you hear, but that is part of the challange.

Mark
Anonymous
July 27, 2004 11:11:00 AM

Archived from groups: rec.audio.pro (More info?)

"Mark" <makolber@yahoo.com> wrote in message
news:3367f36e.0407261911.1ae2b56@posting.google.com...
> My point is that if you claim that an amp sounds bad, or changing caps
> sounds good, I think you at least owe the folks to whom you are making
> the claim, a reason for your claim, a reason backed up by
> measurements. Somthing like the bad bypass cap caused an ultrasonic
> oscillation that created terrible high frequency IM distortion (which
> is easily measured).
>
> I guess our fundamental disagreement is that I beleive that any
> difference that can be heard can also be measured. With FFT
> analyzers that can measure 100 dB down and have sub Hz resolution BW,
> it is my firm belief that anything you can hear can be measured. It
> may be difficult to identify the particular parameter to measure to
> account for what you hear, but that is part of the challange.

In principle, yes, unless you believe in ghosts and gremlins, the causes of
bad sound can be found and measured. I think Stephen's point is that the
traditional measurements of the time (frequency response, THD, SMPTE-IMD and
damping factor at one frequency, all into resistive loads) didn't identify a
problem, but his ears did. We now have some more sophisticated tests, and
sometimes they'll explain what the problem is -- marginal instability into
real-world speaker loads, for example. Or current-limiting baddies, as Scott
described. There are still, however, stubborn cases of amps that measure
well on all the tests we have, yet sound crummy. That doesn't mean
measurements can't find the problem, just that we haven't devised the right
measurements yet.

Peace,
Paul
Anonymous
July 27, 2004 11:24:19 AM

Archived from groups: rec.audio.pro (More info?)

"Paul Stamler" <pstamlerhell@pobox.com> wrote in message
news:8enNc.140313$OB3.123267@bgtnsc05-news.ops.worldnet.att.net

> In principle, yes, unless you believe in ghosts and gremlins, the
> causes of bad sound can be found and measured.

It doesn't take belief in gremlins, just a firm belief in the infallibility
of one's own ears. That can be a horrible mistake - sort of like saying "I'm
a professional, I don't need to take the precautions that the most skilled
and productive modern professionals take as a matter of course."

I am reminded of this every time I see the History Channel show about the
DC10 crash at O'Hare. The guys who did the prerequisite slipshod maintenance
on these jets were obviously convinced of their professionalism. So were
their bosses. So was the FAA. The same basic sloppy procedure was used by at
least three airlines. 273 people died getting this straightened out. Ain't
it great that audio is rarely about life and death?

>I think Stephen's
> point is that the traditional measurements of the time (frequency
> response, THD, SMPTE-IMD and damping factor at one frequency, all
> into resistive loads) didn't identify a problem, but his ears did.

I really didn't see a retrospective tone in Stephen's comments. I read about
outdated technology, though.

> We now have some more sophisticated tests, and sometimes they'll explain
> what the problem is -- marginal instability into real-world speaker
> loads, for example. Or current-limiting baddies, as Scott described.

So far so good.

> There are still, however, stubborn cases of amps that measure well on
> all the tests we have, yet sound crummy.

I'd like to see the specifics of even just one such case. IME, when an amp
sounds bad in a reliable listening test, it is one sick puppy.

Case in point, the Dyna 120. One of the most-criticized amps in the history
of audio. Its worst problem was that it was fragile. It's potential
fragility was obvious to anybody who was familiar with the technology of the
day. Here is a 60 WPC amp built with an output pair that virtually all of
its competitors thought made a pretty good 30-50 WPC amp. Unlike tubes, SS
devices aren't very elastic, and that was well-known at the time.

Lots of people must have spent too much time listening to a Dyna 120 that
would show up as being terribly broken in the simplest of bench tests. I
suspect I could slide one in good operational condition past *everybody* who
reads RAP, in a bias-controlled listening test. But not if they could see
what they were listening to. All the old-timers would *know*.

> That doesn't mean
> measurements can't find the problem, just that we haven't devised the
> right measurements yet.

I question the choice of words. There's no doubt that if an amp does it, and
we can hear it, we can measure it with margins of like 100:1. Measuring
waveforms with -100 dB accuracy is a relatively simple operation. I've
accurately and reliably measured the dynamic range of power amps with 110 dB
dynamic range using less than $1,000 worth of equipment. In a few years that
may drop to $100.

The challenge is setting up the tests and doing the analysis and
interpretation of what we measure. To do that you need to be productive
with signals and systems analysis, and experimental design, and
psychoacoustics. How many technical people in audio, more specifically
audio production, are conversant and productive in all three areas?

"Ear versus Gear" was vastly improved from its chaotic state in say 1973, by
means of improving both ear and gear. The ear side of the equation nets out
to be a matter of personal discipline. The gear side of the equation takes
far less self-discipline = you get to play with fancy toys. Guess which one
more people are interested in?

;-)
Anonymous
July 27, 2004 1:55:25 PM

Archived from groups: rec.audio.pro (More info?)

Mark <makolber@yahoo.com> wrote:
>My point is that if you claim that an amp sounds bad, or changing caps
>sounds good, I think you at least owe the folks to whom you are making
>the claim, a reason for your claim, a reason backed up by
>measurements. Somthing like the bad bypass cap caused an ultrasonic
>oscillation that created terrible high frequency IM distortion (which
>is easily measured).

I wish I had measurements for everything that I think sounds good or bad.
It would sure be nice. And don't think I have spent a lot of time looking.

>I guess our fundamental disagreement is that I beleive that any
>difference that can be heard can also be measured. With FFT
>analyzers that can measure 100 dB down and have sub Hz resolution BW,
>it is my firm belief that anything you can hear can be measured. It
>may be difficult to identify the particular parameter to measure to
>account for what you hear, but that is part of the challange.

That is a substantial and significant challenge. We're finally getting to
the point with the G-L metric that we might have a useful distortion
measurement, too.

The problem with FFT boxes is that they _can_ measure 100 dB down and have
sub Hz bandwidth resolution... and so you are seeing vast amounts of
information, most of which isn't useful. Finding the information that
_is_ useful is the problem.

I have a microphone on the bench today that just has some awful sibilance
issues... but the swept sine response looks just fine, and the impulse
response looks okay. I'm guessing it's some sort of midrange distortion
issue but I don't have any good way of measuring that.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Anonymous
July 27, 2004 3:45:08 PM

Archived from groups: rec.audio.pro (More info?)

"Scott Dorsey" <kludge@panix.com> wrote in message
news:ce5msd$a55$1@panix2.panix.com
>
> I have a microphone on the bench today that just has some awful
> sibilance issues... but the swept sine response looks just fine, and
> the impulse response looks okay. I'm guessing it's some sort of
> midrange distortion issue but I don't have any good way of measuring
> that.


As soon as we step away from simple stuff like electronics, and start
measuring acoustical transducers, everything gets crazy.

While "sounds good and measures bad" is controversial with electronics, with
acoustical transducers, its the way of life.
July 27, 2004 6:19:20 PM

Archived from groups: rec.audio.pro (More info?)

kludge@panix.com (Scott Dorsey) wrote in message news:<ce5msd$a55$1@panix2.panix.com>...
> Mark <makolber@yahoo.com> wrote:
> >My point is that if you claim that an amp sounds bad, or changing caps
> >sounds good, I think you at least owe the folks to whom you are making
> >the claim, a reason for your claim, a reason backed up by
> >measurements. Somthing like the bad bypass cap caused an ultrasonic
> >oscillation that created terrible high frequency IM distortion (which
> >is easily measured).
>
> I wish I had measurements for everything that I think sounds good or bad.
> It would sure be nice. And don't think I have spent a lot of time looking.
>
> >I guess our fundamental disagreement is that I beleive that any
> >difference that can be heard can also be measured. With FFT
> >analyzers that can measure 100 dB down and have sub Hz resolution BW,
> >it is my firm belief that anything you can hear can be measured. It
> >may be difficult to identify the particular parameter to measure to
> >account for what you hear, but that is part of the challange.
>
> That is a substantial and significant challenge. We're finally getting to
> the point with the G-L metric that we might have a useful distortion
> measurement, too.
>
> The problem with FFT boxes is that they _can_ measure 100 dB down and have
> sub Hz bandwidth resolution... and so you are seeing vast amounts of
> information, most of which isn't useful. Finding the information that
> _is_ useful is the problem.
>
> I have a microphone on the bench today that just has some awful sibilance
> issues... but the swept sine response looks just fine, and the impulse
> response looks okay. I'm guessing it's some sort of midrange distortion
> issue but I don't have any good way of measuring that.
> --scott


Scott,
I'm certainly willing to leave mics and speakers out of this
discussion, they ARE hard to measure. The original discussion was
about putting a 0.01uF cap in parallel with a 10uF electrolytic
coupling cap makes an audible difference in an amplifier which has an
electrical input and an electrical output. The effects of a cap mod
on an amplifier (if any) should be very easiliy measureable.

Mark
Anonymous
July 27, 2004 7:40:48 PM

Archived from groups: rec.audio.pro (More info?)

"Chris Hornbeck" <chrishornbeckremovethis@att.net> wrote in message
news:hpoag0h5m96grfv0vmqtiq0se8msog11vg@4ax.com...
> On Mon, 26 Jul 2004 16:33:42 GMT, "normanstrong"
> <normanstrong@comcast.net> wrote:
>
> >Another question might be: How would you go about designing an
> >amplifier that meets the specs you mentioned above, yet sounds
> >"HORRIBLE"? I've mulled the situation over, and I can't think of
any
> >way to accomplish this feat.
>
> In addition to what Arny and Scott have mentioned, a problem
> not so long ago was to specify performance only at full output.
>
> The assumption that all flaws will decrease with signal level
> has become *less* true with most modern improvements.
>
> Full signal performance spec's are almost completely
> irrelevant for current designs.

Did you read the specs? They are applicable from full output down to
250 milliwatts.

Norm Strong
Anonymous
July 27, 2004 8:10:07 PM

Archived from groups: rec.audio.pro (More info?)

On Tue, 27 Jul 2004 15:40:48 GMT, "normanstrong"
<normanstrong@comcast.net> wrote:

>> Full signal performance spec's are almost completely
>> irrelevant for current designs.
>
>Did you read the specs? They are applicable from full output down to
>250 milliwatts.

250 mW into speakers of 89dB SPL/1W/1M sensitivity is 83dB SPL
at one meter from one speaker, 86dB SPL for uncorrelated signals
from two speakers. This is usually greater than 0VU.

The next 60dB or so down from 0VU is what matters. And our
unconscious assumption of monotonicity can lead us astray.

Chris Hornbeck
"Vote or Die" - P. Diddy
Anonymous
July 27, 2004 8:10:08 PM

Archived from groups: rec.audio.pro (More info?)

Yes, that first 250mW is usually very important in a home or studio environment. The spec
cutoff on THD & TIM at that point is a very well known tactic of manufacturers to avoid bad
looking(relative to given specs) figures that are affected by the amp's noise floor & crossover
distortion. However, I have done low level THD & TIM measurements on the Yamaha amps I
mentioned, and they are still excellent measuring amps down to even 50mW, relative to other
amps I have thusly tested. Hence my opinion that the really bad sound is an aspect of
performance deficiency not measurable with the conventional equipment. I doubt even FFT would
reveal very much to explain the horrible sound of these & most other great measuring/bad
sounding amps.
It has actually become a very widespread opinion among people who listen to reproduced music(as
opposed to those who seem to listen to equipment, in such a way as to convince their own ears
to conform to what the measurements reveal) that one should be highly suspicious of an amp,
preamp, etc., that has really good measurements. This is because it is so frequently the case,
and I think the majority of cases, that the better the specs are, the worse the gear
sounds(within rational limits). Obviously, if one sees really bad specs, like 1% THD or TIM,
or 22dB S/N, one would not expect great sound. But the best amps I have ever heard are often
in the 0.1-0.3% THD/TIM range, such as the Nakamichi PA-7AII, Threshold S/500II and McIntosh
MC2105. Can't name one amp I loved the sound of that was lower than 0.02% on the distortion
specs.

--
Stephen Sank, Owner & Ribbon Mic Restorer
Talking Dog Transducer Company
http://stephensank.com
5517 Carmelita Drive N.E.
Albuquerque, New Mexico [87111]
505-332-0336
Auth. Nakamichi & McIntosh servicer
Payments preferred through Paypal.com
"Chris Hornbeck" <chrishornbeckremovethis@att.net> wrote in message
news:k0vcg0lsf1thku6bedn3f44j9rnnoot6mn@4ax.com...
> On Tue, 27 Jul 2004 15:40:48 GMT, "normanstrong"
> <normanstrong@comcast.net> wrote:
>
> >> Full signal performance spec's are almost completely
> >> irrelevant for current designs.
> >
> >Did you read the specs? They are applicable from full output down to
> >250 milliwatts.
>
> 250 mW into speakers of 89dB SPL/1W/1M sensitivity is 83dB SPL
> at one meter from one speaker, 86dB SPL for uncorrelated signals
> from two speakers. This is usually greater than 0VU.
>
> The next 60dB or so down from 0VU is what matters. And our
> unconscious assumption of monotonicity can lead us astray.
>
> Chris Hornbeck
> "Vote or Die" - P. Diddy
Anonymous
July 27, 2004 9:27:32 PM

Archived from groups: rec.audio.pro (More info?)

"Paul Stamler"

> In principle, yes, unless you believe in ghosts and gremlins, the causes
of
> bad sound can be found and measured. I think Stephen's point is that the
> traditional measurements of the time (frequency response, THD, SMPTE-IMD
and
> damping factor at one frequency, all into resistive loads) didn't identify
a
> problem, but his ears did.


** Stephen did not produce for us a single scrap of proof of their being a
problem with the amp. He even failed to do ANY sort of scientific
comparison test - so his wild claims should no be taken as fact by anyone.


> There are still, however, stubborn cases of amps that measure
> well on all the tests we have, yet sound crummy.


** It is just too easy to make throw away remarks like that - Paul.

NOTE: When you can prove even *one example* that complies with the above
assertion exists, Paul Stamler will become the most famous man in audio.


>That doesn't mean measurements can't find the problem, just that we haven't
devised the right
> measurements yet.


** More audiophool mantras - how boring.




............ Phil
Anonymous
July 27, 2004 9:27:33 PM

Archived from groups: rec.audio.pro (More info?)

Phil Allison <philallison@tpg.com.au> wrote:
>"Paul Stamler"
>
>> There are still, however, stubborn cases of amps that measure
>> well on all the tests we have, yet sound crummy.
>
>** It is just too easy to make throw away remarks like that - Paul.
>
> NOTE: When you can prove even *one example* that complies with the above
>assertion exists, Paul Stamler will become the most famous man in audio.

There are examples like that all the time, and that is why we are getting
newer and better measurements all the time. As people find more things
that are audible, they design new measurements to measure them.

If every possible aspect of sound can already be measured, why are there
all of these metrology papers at AES and AAS meetings every year?
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Anonymous
July 27, 2004 9:27:34 PM

Archived from groups: rec.audio.pro (More info?)

"Scott Dorsey" <kludge@panix.com> wrote in message
news:ce5ne5$lju$1@panix2.panix.com

> Phil Allison <philallison@tpg.com.au> wrote:

>> "Paul Stamler"

>>> There are still, however, stubborn cases of amps that measure
>>> well on all the tests we have, yet sound crummy.

>> ** It is just too easy to make throw away remarks like that -

>> NOTE: When you can prove even *one example* that complies with the
>> above assertion exists, Paul Stamler will become the most famous man
>> in audio.

Thinking about it, there might be an argument about what constitutes
measuring *well* as compared to measuing *unwell*.

For example JJ (ex AT&T Labs) once suggested that anything that measures
with all distortions and noises 100 dB down, is sonically transparent. I
think that criteria is going to stand for an awfully long time.

You see, the actual threshold for hearing noise and distortion is generally
below 70 dB on an instantaneous basis. JJ's other 30 dB relate to dynamics,
exceptional cases, and basic conservatism.

> There are examples like that all the time, and that is why we are
> getting newer and better measurements all the time.

Really?

I was thinking of "new measurements", and I came up with the Geddes/Geddes
criteria for nonlinear distortion. But, its not a new measurement, its a
modified way to analyze data that has been measured for decades. It's not so
much about thresholds but rather about comparing different amounts of
audible distortion, and saying which one "sounds worse" given that both are
audible.

> As people find more things that are audible, they design new measurements
to measure them.

It has been known that nonlinear distortion has been audible for at least 70
years. The Geddes/Geddes paper is about psychoacoustic weighting factors for
distortion products that have been measured in a highly detailed fashion for
decades. It doesn't so much as lower the known thresholds for detection, as
raise them selectively.

> If every possible aspect of sound can already be measured, why are
> there all of these metrology papers at AES and AAS meetings every
> year?

(1) Are they really about metrology, or are they really about something
else, like psychoacoustics?
(2) Three words: Publish or Perish.
(3) Eight words: Professional papers as publicity for a new product.
Anonymous
July 27, 2004 9:27:35 PM

Archived from groups: rec.audio.pro (More info?)

Arny Krueger <arnyk@hotpop.com> wrote:
>"Scott Dorsey" <kludge@panix.com> wrote in message
>
>> If every possible aspect of sound can already be measured, why are
>> there all of these metrology papers at AES and AAS meetings every
>> year?
>
>(1) Are they really about metrology, or are they really about something
>else, like psychoacoustics?

They are more and more about psychoacoustics, as more of the systems that
we need to measure are systems that use perceptual-encoding methods.

But they are all in some sense about psychoacoustics, because the whole
name of the game, starting in the 1930s with the THD measurement at Western
Electric, is to figure out a way to measure what people are hearing.

>(3) Eight words: Professional papers as publicity for a new product.

There is a huge amount of this, but there is actually less of it in the
metrology world than elsewhere, probably because there isn't much money
in new products there. The JAES is very good about not allowing
this sort of thing but there is a huge amount of it at conventions, though
it's usually pretty obvious. Many of the section folks put the obvious
advertisements together into a single session so everyone can avoid them,
or drop them into the posters area.

Personally, I have nothing against those papers, as long as they do actually
have real technical content. (And some of the ones that don't are still
worthwhile for amusement.... some of the Pioneer-sponsored papers on 96k
audibility just had people laughing in the aisles.)
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Anonymous
July 27, 2004 9:52:34 PM

Archived from groups: rec.audio.pro (More info?)

"Stephen Sank" <bk11@thuntek.net> wrote in message
news:ce6hn7$ock$1@reader2.nmix.net

> Yes, that first 250mW is usually very important in a home or studio
> environment. The spec cutoff on THD & TIM at that point is a very
> well known tactic of manufacturers to avoid bad looking(relative to
> given specs) figures that are affected by the amp's noise floor &
> crossover distortion. However, I have done low level THD & TIM
> measurements on the Yamaha amps I mentioned, and they are still
> excellent measuring amps down to even 50mW, relative to other amps I
> have thusly tested. Hence my opinion that the really bad sound is an
> aspect of performance deficiency not measurable with the conventional
> equipment.

Specifically, which Yamaha amp(s) are you talking about?
Anonymous
July 28, 2004 6:50:01 AM

Archived from groups: rec.audio.pro (More info?)

On Tue, 27 Jul 2004 17:52:34 -0400, "Arny Krueger" <arnyk@hotpop.com>
wrote:

>Specifically, which Yamaha amp(s) are you talking about?

The M series Yamaha's of that era have several known defects,
including hot glue which has by now baked to a corrosive and
partially conductive state, and solder joints old enough to
vote and frequently marginal now.

And, of course, the inevitable electrolytics well past their
prime. But even when you restore all the things that you know
about, they're still not great.

Design choices that emphasize .00x THD figures sacrifice some
other virtues which actually matter more. Stability under
dynamic and especially ultrasonic stress, small signal
linearity, spectral content of residual distortion products,
insensitivity to ground and power supply contamination,
recovery from transient overloads, etc.

These and similar issues don't fit easily into a "spec sheet"
but are critical in real electronics.

As you know. I've seen your name in old construction articles
from 20 plus years ago; you've gotten your hands dirty too.

Chris Hornbeck
"Vote or Die" - P. Diddy
Anonymous
July 28, 2004 8:38:05 AM

Archived from groups: rec.audio.pro (More info?)

On Tue, 27 Jul 2004 21:59:02 -0600, "Stephen Sank" <bk11@thuntek.net>
wrote:

>I imagine you are also familiar with the succeding MX-xxx series that replaced the M85/65/45.
>Sounded even worse, and loved to blow up as ballistically as Phase Linear 400's.

Well, I've been off the front lines for a while now. And I wouldn't
really want to pan Yamaha specifically; my gripes are kinda generic.

Besides, I *love* their little $100 mixer. Maybe not a Mackie, but
not bad at all. It even feels good.

And the 1980's were a difficult time anywho. We were all just trying
to find our way. Or something.

Anyway, thanks for your comments,

Chris Hornbeck
"Vote or Die" - P. Diddy
Anonymous
July 28, 2004 1:22:57 PM

Archived from groups: rec.audio.pro (More info?)

Phil Allison <philallison@tpg.com.au> wrote:
>
>
> BTW He still will * not * name the TV repair business in Sydney that he
>alleged I run.

I will again. http://www.allisonaudio.com.au.

Perhaps your shouting about lack of evidence has to do with the fact that
people present you with urls and citations that you ignore?
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Anonymous
July 29, 2004 4:19:06 AM

Archived from groups: rec.audio.pro (More info?)

"Scott Dorsey"

>
> Either one of these amplifiers will horribly fail a TIMD test. As will
> the Phase Linears and Dynaco ST120s.
>

** TIMD ( and SID) have been well and truly been proved to be a myths as
far as the reproduction of music program is concerned.

Matti Otala was wrong as well - his published maths contained a basic
error.



> The reason we _have_ TIMD testing is because of all the amps in the
seventies
> that sounded bad but measured well.


** That is another wild and wrong assertion.


> So somebody had to come up with a test
> that explained why they sounded bad.


** No formal listening tests ever showed that 70s amps "sounded bad".

Quite the opposite was found in fact when controlled tests were done.



>
> Now we're living in a world of perceptual encoding systems which measure
> well and sound bad,


** Wrong again.




............. Phil
Anonymous
July 29, 2004 4:19:07 AM

Archived from groups: rec.audio.pro (More info?)

Phil Allison <philallison@tpg.com.au> wrote:
>"Scott Dorsey"
>>
>> Either one of these amplifiers will horribly fail a TIMD test. As will
>> the Phase Linears and Dynaco ST120s.
>>
>
>** TIMD ( and SID) have been well and truly been proved to be a myths as
>far as the reproduction of music program is concerned.
>
> Matti Otala was wrong as well - his published maths contained a basic
>error.

Have you got any citations for these statements?

And, can you give me any better explanation for the ST120 sounding so bad,
for instance?

>** No formal listening tests ever showed that 70s amps "sounded bad".
>
> Quite the opposite was found in fact when controlled tests were done.

How can you expect to have any credibility when you make statements like
this? These amps were so noxious that a deaf elephant could tell how
harsh they sounded.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Anonymous
July 29, 2004 4:19:07 AM

Archived from groups: rec.audio.pro (More info?)

"Phil Allison" <philallison@tpg.com.au> wrote in message
news:2mpqujFp339cU1@uni-berlin.de
> "Scott Dorsey"

>> Either one of these amplifiers will horribly fail a TIMD test. As
>> will the Phase Linears and Dynaco ST120s.

AFAK, there's no AES or other comparable organization standard TIM test. I'm
willing to go this this suggestion:

http://www.benchmarkmedia.com/appnotes-a/caig/caig08-10...

"If slewing induced intermodulation and/or transient intermodulation
distortion exist within the system, it will only occur at the higher
frequencies. The best way in a bandwidth limited system to detect their
presence is with the CCIF twin tone IM distortion measurement. By using 14
and 15 kHz tones mixed 1:1, a 1 kHz IM product is easily detected if SID/TIM
exists in an FM broadcast system. A 19 and 20 kHz pair may be used to
evaluate A-to-D converters or other equipment in the audio chain. We believe
that every broadcast facility should perform twin tone IM distortion
measurements to see the truth about heir high frequency audio performance. "

I've recently done twin-tone tests on a number of ca. 70s power amps
including the Dyna ST120, and it doesn't seem to measure out as having this
problem. Admittedly, I might have one of the few ST120's that appears to be
operating with its original output devices, not that spares aren't on hand!


> ** TIMD ( and SID) have been well and truly been proved to be a
> myths as far as the reproduction of music program is concerned.

> Matti Otala was wrong as well - his published maths contained a basic
error.

He published corrections, later on.

Otala's 1977 article seems to agree the Benchmark Media paper mentioned
above:

"The CCIF-IM method gives a reliable indication of dynamic intermodulation
distortion."

>> The reason we _have_ TIMD testing is because of all the amps in the
>> seventies that sounded bad but measured well.

> ** That is another wild and wrong assertion.

It's also self-contradictory.

Marshall Leach first published the Audio Magazine construction articles for
his "Low TIM" amplifier in February 1976. Therefore, it would seem to be
safe to say that there were a goodly number of low-TIMD solid state
amplifiers in the 1970s since the technology for building them was
well-known just after the middle of the decade.

http://users.ece.gatech.edu/~mleach/papers/lowtim/feb76...

>> So somebody had to come up with a test
>> that explained why they sounded bad.

> ** No formal listening tests ever showed that 70s amps "sounded bad".

As a general rule, no these amps didn't all sound bad.

> Quite the opposite was found in fact when controlled tests were
> done.

Agreed.

>> Now we're living in a world of perceptual encoding systems which
>> measure well and sound bad,

> ** Wrong again.

It is true that at least some of the best-known bad sounding perceptual
coders also measured bad. Please see:

http://www.pcavtech.com/techtalk/compare_perceptual/ind...
Anonymous
July 29, 2004 4:19:08 AM

Archived from groups: rec.audio.pro (More info?)

Phil Allison <philallison@tpg.com.au> wrote:
>Kludge wrote:
>
> The slew rates contained in program material can be easily shown to never
>approach the rates needed to induce TIMD or SID in commercial SS hi-fi amps.

I'm not sure I buy that, but I'm looking up the issues to see.

>> And, can you give me any better explanation for the ST120 sounding so bad,
>> for instance?
>
>
>** That is called "begging the question" - a favourite cheat used by
>liars.

No. You might have plenty of other good explanations, like high order
distortion products. Or the fact that the amplifier is on the bare edge
of instability and breaks out into ultrasonic oscillation on some source
material peaks.

>> >** No formal listening tests ever showed that 70s amps "sounded bad".
>> >
>> > Quite the opposite was found in fact when controlled tests were done.
>>
>>
>> How can you expect to have any credibility when you make statements like
>> this?
>
>
>** Then kindly supply evidence from formally conducted listening tests that
>backs you up.

Suppression of Slew Rate and Transient IM Distortions in Audio Power Amplifiers
Marshall W. Leach, AES preprint 1137

(This is pretty much the grandaddy of all of these things and predates the
Audio magazine articles that Leach did. And yes, there IS a listening test and
it does correlate.)

Slewing Induced Distortion and Its Effect On Audio Amplifier Performance--
With Correlated Measurement/Listening Results
Jung, Walter G.; Stephens, Mark L.; Todd, Craig C., AES preprint 1252

Envelope Distortion in Audio Amplifiers
Takahashi, Susumu; Funada, Saburo, AES preprint 1344

Interface Intermodulation Distortion (IIM) in Power Amplifiers
Corveleyn, L.; Bossuyt, F.; Sansen, W., AES preprint 1869

(This paper is rather different than the others since it is specifically
relating to intermodulation effects induced by the load nonlinearity).

Transient Distortion in Feedback Audio Power Amplifiers
Glowacki, Miecyslaw; Stanclik, Josef; Pierzchala, Marian, AES preprint 3605

(The listening tests in this one are fairly minimal, but the model being
used is a fairly important one that gets cited a lot in other testing.)

--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Anonymous
July 29, 2004 4:19:08 AM

Archived from groups: rec.audio.pro (More info?)

"Scott Dorsey" <kludge@panix.com> wrote in message
news:ce8e48$k89$1@panix2.panix.com...
> And, can you give me any better explanation for the ST120 sounding
so bad,
> for instance?
>
> >** No formal listening tests ever showed that 70s amps "sounded
bad".
> >
> > Quite the opposite was found in fact when controlled tests
were done.
>
> How can you expect to have any credibility when you make statements
like
> this? These amps were so noxious that a deaf elephant could tell
how
> harsh they sounded.

I don't know about deaf elephants, but apparently the designer of the
ST120 didn't think his amp sounded bad--or he wouldn't have put it on
the market. Additionally, since the ST120 sold briskly, I would have
to assume that its faults were not immediately evident to customers.

BTW, was the ST120 a Hafler design?

Norm Strong
Anonymous
July 29, 2004 4:19:09 AM

Archived from groups: rec.audio.pro (More info?)

normanstrong <normanstrong@comcast.net> wrote:
>
>I don't know about deaf elephants, but apparently the designer of the
>ST120 didn't think his amp sounded bad--or he wouldn't have put it on
>the market. Additionally, since the ST120 sold briskly, I would have
>to assume that its faults were not immediately evident to customers.
>
>BTW, was the ST120 a Hafler design?

I don't know if it was a Hafler design or not. It has some interesting
little tricks to it, like the huge inductor wound around a capacitor can
that is used on the output in an attempt to keep the thing stable and some
biasing goofiness to avoid using a bipolar supply with a push-pull output
stage.

I do know that David Hafler does use it as an example of one of the bad
early power amp designs.

The thing is that the ST120, when compared with the tube amps of the
era, sounded very bright and forward (probably because of all the high
order harmonics). This seemed like a good thing to a lot of people back
then, but it's the sort of thing that eventually gave solid state
electronics a bad name. It took some listening, though, for people to
realize what was going on, because they were being presented with a set
of distortions that they weren't used to listening to.

This is always the case. When the Edison phonograph came out, many people
said that it was just like listening to the performer right there in the
room, it was so accurate. It took people a few years to get used to the
sound and understand the deficiencies.

The same thing happens every time there is a revolutionary change in
design. It happened with solid state electronics and it happened with
digital audio and it's happening right now with perceptual encoding systems.
It takes people a while to recognize what they are listening to.
--scott


--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Anonymous
July 29, 2004 4:19:10 AM

Archived from groups: rec.audio.pro (More info?)

On 28 Jul 2004 13:47:49 -0400, kludge@panix.com (Scott Dorsey) wrote:

>The same thing happens every time there is a revolutionary change in
>design. It happened with solid state electronics and it happened with
>digital audio and it's happening right now with perceptual encoding systems.
>It takes people a while to recognize what they are listening to.

When people in pre-technological societies are first shown photographs
they frequently can't distinguish between the portrait and the person.

The difficult 99% of perception is in the wetware. But our
unchallenged faith in "objectivity" is touching.

Chris Hornbeck
"Vote or Die" - P. Diddy
!