[John Byrns]

In article ,
Ian Thompson-Bell wrote:

John Byrns wrote:

For those of us that are not "pro audio" designers, can you tell us what
the reference is for these gain figures, gain from input to what point?



Usually it is from mic input to a final output.

So all of the specified gain, ignoring makeup gain, is not necessarily
in the microphone amplifier proper?


Regards,

John Byrns

--
Surf my web pages at, http://fmamradios.com/

[Arny Krueger]

"John Byrns" wrote in message


Aren't THD and IMD levels pretty highly correlated unless
there are serious frequency response anomalies involved?

Yes.

The relationship between THD and IMD depend on:

(1) The amount and order of the nonlinearity.

(2) Frequency dependencies of the nonlinearity and system gain

(3) Which flavor of IM test that is involved, and signal level of the test
signal.

Hold all three constant, and the relationship between THD and IMD is
constant.

[Eeyore]

John Byrns wrote:

Patrick Turner wrote:

So you'd have to agree with me that utterly negligible IMD is caused by
my design idea.

Yes, but you are going to have a hard time getting Eeyore to buy in.

Show me the calculations !

In any event I think it's a crummy mixing method.

Graham

[Eeyore]

RickH wrote:

Oh you'd rather have 32 DtoA conversions up front

The NUMBER of conversions taking place in parallel is completely irrelevant.

It's a classic red herring that seeks to imply a defect through an implied
misunderstanding of the importance of same.

You should be ashamed of yourself for resorting to such an intellectually feeble
argument.

Graham

[Eeyore]

John Byrns wrote:

Patrick Turner wrote:

A large number of fabulous recordings of the golden days of vinyl
were made with gear stuffed full of EF86 in either pentode or triode,
and also full of transformers, and THD was often 0.1% or more,
and not all just 2H.

Yes, well that's VINYL for you !


Yes, and the premium versions, I forget the type numbers, of the 6SJ7
were used to good effect here in the U.S., The 5879 also ccquitted
itself well.

And the data I have unearthed shows the very excellent (for the period)
Siemens/Telefunken tube mic amps showing 0.5% THD specs. Of course that would be
swamped by pickup cartridge errors.

Graham

[Eeyore]

John Byrns wrote:

Patrick Turner wrote:

Their choice of loading the tubes and amount of FB and tube choice will
vary....

It's not pointless at all, Eeyore has injected "triode sound" into this
discussion as a supposed requirement.

Do you prefer 'pentode sound' ?

Come out of the closet and tell us.

Graham

[Eeyore]

Ian Thompson-Bell wrote:

Eeyore wrote:
Ian Thompson-Bell wrote:
Eeyore wrote:
Ian Thompson-Bell wrote:

Second, the buffer is after the fader and pan pot and it directly feeds
the mix bus. Therefore, all the buffers, used or not, contribute noise
to the mix bus. 8 buffers adds 9dB of noise.
I don't know where you get that from. The buffers I have in mind are
effectively noiseless.

Assume the buffers are identical and their output noise is -90dBu.

That's a big assumption. Is even a cathode follower that noisy ?

Doesn't matter what the actual noise level is, the principle still
applies. If the buffer noise was -110dBu then I'd agree you could ignore it.

At -110 then yes, I'd not consider it to be any big deal.


Select two channels and the total buffer noise on the bus will rise by
3dB. Select two more and it will rise another 3dB. Select four more and
it rises another 3dB - total 9dB.

I have been experimenting with CF buffers recently and it is hard to get
their output noise below -85dB. What were these noiseless buffers you
had in mind?

They're that bad ?

The ones I have at the moment are. I need to do more tests to confirm
it. The output Z is about 500 ohms so they are way noisier than the pure
resistive noise.

I'm puzzled by your results to be sure.


I was planning on cheating and using an enhanced complementary Class A emitter
follower solid-state buffer. It's easy to get sub 100 ohms output Z that way
and not influence the 'tube sound' adversely.

I think you may be on dodgy ground there. ;-)

We'll see. I know I'm not on technically dodgy ground (by a country mile) but then
the perception (or should I say religious view) may be different.

Graham

[Eeyore]

John Byrns wrote:

Ian Thompson-Bell wrote:
Eeyore wrote:
Patrick Turner wrote:

so you need to spell out clearly what mics you plan to use
so that levels and attenuation can be all tailored to suit
a best outcome.

This is NOT an issue any more than it is for modern mic amps. Variable gain
between 20 and 80 dB plus a 20dB pad for high level signals will handle all
eventualities.

I would venture to suggest that for the vast majority of applications
and mics that 60dB max gain will be ample. At 80dB gain, even the
quietest preamps will give a S/N no better than about 50dB from a 200
ohm source.

For those of us that are not "pro audio" designers, can you tell us what
the reference is for these gain figures, gain from input to what point?

Look ****wit STOP TROLLING.

VOLTAGE GAIN = OUTPUT VOLTAGE / INPUT VOLTAGE !

How much ****ing clearer could it be ?

Graham

[Eeyore]

Ian Thompson-Bell wrote:

John Byrns wrote:

Aren't THD and IMD levels pretty highly correlated unless there are
serious frequency response anomalies involved?

Both depend on the non-linearities of transfer function of the device.
So for a particular device they will doubtless be correlated but between
different devices I don't know. As a rule IM tends to be higher than THD
and more unacceptable to the ear.

And IM truly STINKS !

In particular, high power valve power amps produce IM like it's going out of
fashion. A lot of that is likely due to output transformer issues.

I'm old enought to have mixed live bands using valve 'PAs'. The sound STANK
due to intermodulation. It was truly BAD ! Transistors were like a breath of
fresh air in comparison. A live performance with a transistorised PA could
be surprisingly close to the recording.

Graham

[Eeyore]

John Byrns wrote:

Eeyore wrote:
John Byrns wrote:
Eeyore wrote:

Some how I doubt you made any SPICE models ever.

Can you explain how you come to that conclusion?

You're not technically literate enough.

Your logic is flawed, which isn't a big surprise, how does a lack of
technical literacy prevent one from messing about with the creation of
spice models?

You're still an idiot with a chip on your shoulder.

[Eeyore]

Ian Thompson-Bell wrote:

John Byrns wrote:

For those of us that are not "pro audio" designers, can you tell us what
the reference is for these gain figures, gain from input to what point?

Usually it is from mic input to a final output.

John's not the smartest guy around.

Gain is invariably defined like that. It's no surprise to me that he doesn't
even undertand that much.

Graham

[Eeyore]

John Byrns wrote:

I was simply trying to point out that IMD and THD are highly correlated and
as a rule you don't have
one without the other, and they pretty much track one another

Pure nonsense. The mechanisms are quite different.

Graham

[Eeyore]

John Byrns wrote:

Ian Thompson-Bell wrote:
John Byrns wrote:

For those of us that are not "pro audio" designers, can you tell us what
the reference is for these gain figures, gain from input to what point?


Usually it is from mic input to a final output.

So all of the specified gain, ignoring makeup gain, is not necessarily
in the microphone amplifier proper?

Mic amps don't have 'makeup gain' you blithering MORON. You've learnt a few
buzzwords but quite clearly haven'y the tiniest idea what they mean. Spewing
tham at random as you're doing only makes you look stupider than you did before.
Were you dropped on your head as child perchance ?

You're an ignorant and highly offensive and disruptive prick Byrns ! Do you LIKE
lying and making things up ?

Graham

[Ian Thompson-Bell]

John Byrns wrote:
In article ,
Ian Thompson-Bell wrote:

John Byrns wrote:
For those of us that are not "pro audio" designers, can you tell us what
the reference is for these gain figures, gain from input to what point?


Usually it is from mic input to a final output.

So all of the specified gain, ignoring makeup gain, is not necessarily
in the microphone amplifier proper?



In practice it is pretty much all in the microphone amplifier proper. It
has to do with the balance between headroom and noise floor that I
mentioned before.

Starting with a 200 ohm source the noise seen at the mic amp input due
to that noise is about -130dBu. A perfect mic amp set to 80dB gain would
amplify this noise to -130 + 80dBu = -50dBu. Modern mic pres are pretty
near perfect and will only worsen this figure by a dB or so.

Let's look at the noise output of our perfect mic pre at other gain setting:

Gain 60dB - output noise -70dBu
Gain 40dB - output noise -90dBu
Gain 20dB - output noise -110dBu
Gain 0dB - output noise -130dBu

In practice, at around 30 to 40dB of gain, the output noise of the mic
pre itself begins to dominate - a good mic pre will have an output noise
of around -100dB, so at around 30dB or less gain the mic pre noise
begins to dominate and at lower gains it remains at about -100dBu.

So the lowest possible level of noise from the output of our mic pre is
-100dBu and at gains over 30dB it gets worse due to input noise.

The next question is what level should we mix at? Since our noise level
is always -100dBu or worse we want to mix at as high a level as
possible; that is we want out signal to be as far above the noise as we
can get it. Since a CD can give a theoretical 96dB S/N we want to our
mix to achieve at least this S/N which implies a mixing level around
0dBu. Since we know we can design preamps that will deliver +20dBu,
choosing a 0dBu mix level gives us 20dB headroom and a potential 120dB
dynamic range.

So since our output level also wants be be in the region of 0dBu we want
want any additional gain after mixing (except to make up for mix losses)
so in practice all our gain ends up being in the mic pre.

Cheers

ian

[Ian Thompson-Bell]

Eeyore wrote:

Ian Thompson-Bell wrote:

Eeyore wrote:

I have been experimenting with CF buffers recently and it is hard to get
their output noise below -85dB. What were these noiseless buffers you
had in mind?

They're that bad ?

The ones I have at the moment are. I need to do more tests to confirm
it. The output Z is about 500 ohms so they are way noisier than the pure
resistive noise.

I'm puzzled by your results to be sure.



I have just done another test of a CF (using a 6AU6 in triode mode) with
its input shorted (ac) and this gave an output noise of just a shade
worse than -92dBu. I suspect to improve on this figure a beefier tube
with a higher current and lower output Z would be necessary.

Cheers

Ian

[Eeyore]

Ian Thompson-Bell wrote:

Eeyore wrote:
Ian Thompson-Bell wrote:
Eeyore wrote:

I have been experimenting with CF buffers recently and it is hard to get
their output noise below -85dB. What were these noiseless buffers you
had in mind?

They're that bad ?

The ones I have at the moment are. I need to do more tests to confirm
it. The output Z is about 500 ohms so they are way noisier than the pure
resistive noise.

I'm puzzled by your results to be sure.

I have just done another test of a CF (using a 6AU6 in triode mode) with
its input shorted (ac) and this gave an output noise of just a shade
worse than -92dBu. I suspect to improve on this figure a beefier tube
with a higher current and lower output Z would be necessary.

Well .... I recall noise figures of as low as 2uV (audio band) for modern
triodes. That's equivalent to ~ -110dBu !

Something's wrong here methinks.

Certainly my idea of a SS buffer will be vastly quieter still.

Graham

[Ian Thompson-Bell]

Eeyore wrote:

Ian Thompson-Bell wrote:

Eeyore wrote:
Ian Thompson-Bell wrote:
Eeyore wrote:
I have been experimenting with CF buffers recently and it is hard to get
their output noise below -85dB. What were these noiseless buffers you
had in mind?
They're that bad ?
The ones I have at the moment are. I need to do more tests to confirm
it. The output Z is about 500 ohms so they are way noisier than the pure
resistive noise.
I'm puzzled by your results to be sure.
I have just done another test of a CF (using a 6AU6 in triode mode) with
its input shorted (ac) and this gave an output noise of just a shade
worse than -92dBu. I suspect to improve on this figure a beefier tube
with a higher current and lower output Z would be necessary.

Well .... I recall noise figures of as low as 2uV (audio band) for modern
triodes. That's equivalent to ~ -110dBu !

Those figures are usually referred to the input (the EF86 for example
quotes a noise voltage of 2uV referred to g1 in a 10KHz bandwidth) in
common cathode mode. Output noise is not mentioned but of course in a CC
stage with any reasonable gain the input noise dominates. I don't know
of any figures for CF configuration where output noise presumably plays
ae much larger role.


Something's wrong here methinks.


As an aside I left the CF test running for an hour and when I cam back
to it the noise had dropped a further 6dB to a slightly more respectable
-98dBu. Not quite sure why this should be so but I have turned it off to
cool down and I shall later power it up again and check the noise level
vs time. Maybe this is why the old tube console manuals said to switch
on and wait 30 minutes before using?

Cheers

Ian

[Eeyore]

Ian Thompson-Bell wrote:

Eeyore wrote:
Ian Thompson-Bell wrote:
Eeyore wrote:
Ian Thompson-Bell wrote:
Eeyore wrote:
I have been experimenting with CF buffers recently and it is hard to get
their output noise below -85dB. What were these noiseless buffers you
had in mind?
They're that bad ?
The ones I have at the moment are. I need to do more tests to confirm
it. The output Z is about 500 ohms so they are way noisier than the pure
resistive noise.
I'm puzzled by your results to be sure.
I have just done another test of a CF (using a 6AU6 in triode mode) with
its input shorted (ac) and this gave an output noise of just a shade
worse than -92dBu. I suspect to improve on this figure a beefier tube
with a higher current and lower output Z would be necessary.

Well .... I recall noise figures of as low as 2uV (audio band) for modern
triodes. That's equivalent to ~ -110dBu !

Those figures are usually referred to the input (the EF86 for example
quotes a noise voltage of 2uV referred to g1 in a 10KHz bandwidth) in
common cathode mode.

Yes, that's the kind of thing I had in mind.


Output noise is not mentioned but of course in a CC
stage with any reasonable gain the input noise dominates. I don't know
of any figures for CF configuration where output noise presumably plays
ae much larger role.

Well ... by inspection .. if the noise is 2uV on the grid with the cathode AC
grounded, then the differential noise voltage in any other application should be
the same.


Something's wrong here methinks.

As an aside I left the CF test running for an hour and when I cam back
to it the noise had dropped a further 6dB to a slightly more respectable
-98dBu. Not quite sure why this should be so but I have turned it off to
cool down and I shall later power it up again and check the noise level
vs time. Maybe this is why the old tube console manuals said to switch
on and wait 30 minutes before using?

Maybe I can convince you that my ultra-fidelista transistor anode buffer is the
best answer ?

Graham

[Ian Thompson-Bell]

Eeyore wrote:

Ian Thompson-Bell wrote:


Output noise is not mentioned but of course in a CC
stage with any reasonable gain the input noise dominates. I don't know
of any figures for CF configuration where output noise presumably plays
ae much larger role.

Well ... by inspection .. if the noise is 2uV on the grid with the cathode AC
grounded, then the differential noise voltage in any other application should be
the same.


Er, the noise *referred* to the grid is 2uV when connected in CC
configuration. Stage gain could be 30dB which means noise in the anode
circuit would be nearly 60uV. What the sources of this noise are ,
internal to the valve, is not stated. So in CC some of these will appear
in the cathode circuit. Typically a CF reduces noise induced in its
anode circuit. In other tests with this CF (when the PSU was not as hum
free as it should be) this factor was about 10, so you might expect
about 6uV in the cathode. I currently measure about 10, maybe less but
it is at the limit of my current instrumentation so it might be lower
and I am not using any bandwidth limitation so that could be a factor too.

By the way I discovered why the original reading was 6dB higher - it was
crap induced in the test leads by my temperature controlled soldering
iron - switch off the iron and the noise drops 6dB.

Cheers

ian

[Ian Thompson-Bell]

Ian Thompson-Bell wrote:

By the way I discovered why the original reading was 6dB higher - it was
crap induced in the test leads by my temperature controlled soldering
iron - switch off the iron and the noise drops 6dB.


I just added a few nF across the output to reduce the bandwidth a bit
and to my surprise the noise went up a bit. Thinking perhaps the cap was
picking up crap from somewhere (this test rig is completely unscreened)
I decided to turn of my PC with is LCD display which sits on the
adjacent bench. Lo and behold the noise drops yet again to 3 or 4uV, so
this CF now achieves nearly -106dBu - that's more like it.

Just shows you how careful you need to be when measuring very small
noise voltages.

Cheers

Ian

[RickH]

On Jan 13, 1:45*pm, Eeyore
wrote:
RickH wrote:
Oh you'd rather have 32 DtoA conversions up front

The NUMBER of conversions taking place in parallel is completely irrelevant.

[Arny Krueger]

"Eeyore" wrote in
message
RickH wrote:

Oh you'd rather have 32 DtoA conversions up front

The NUMBER of conversions taking place in parallel is
completely irrelevant.

Practically so.

In fact parallel ADCs can be used to improve dynamic range, albeit
expensively.

But that isn't really what he's talking about. If it was, it would be a sign
that he has a clue.

It's a classic red herring that seeks to imply a defect
through an implied misunderstanding of the importance of
same.

The guy is obviously technically clueless.

You should be ashamed of yourself for resorting to such
an intellectually feeble argument.

It seems to come with the territory. ;-)

[Arny Krueger]

"RickH" wrote in message


Maybe the OP shuld just go digital then.

Only if he wants the best possible SQ.

[Arny Krueger]

"Eeyore" wrote in
message
John Byrns wrote:

I was simply trying to point out that IMD and THD are
highly correlated and as a rule you don't have
one without the other, and they pretty much track one
another

Pure nonsense. The mechanisms are quite different.

??????????

[RickH]

On Jan 13, 1:45*pm, Eeyore
wrote:
RickH wrote:
Oh you'd rather have 32 DtoA conversions up front

The NUMBER of conversions taking place in parallel is completely irrelevant.

[John Byrns]

In article ,
Eeyore wrote:

John Byrns wrote:

Yes, and the premium versions, I forget the type numbers, of the 6SJ7
were used to good effect here in the U.S., The 5879 also ccquitted
itself well.

And the data I have unearthed shows the very excellent (for the period)
Siemens/Telefunken tube mic amps showing 0.5% THD specs. Of course that would
be
swamped by pickup cartridge errors.

I don't know if it is the case for the Telefunken tube mic amplifiers
you found, but the specifications for much vintage "pro audio" equipment
seem to be based on absolute worst case conditions, I have noticed that
the actual performance typically achieved is far better than specified,
for various reasons the old guys seem to have been very conservative in
determining their ratings.


Regards,

John Byrns

--
Surf my web pages at, http://fmamradios.com/

[John Byrns]

In article ,
Eeyore wrote:

John Byrns wrote:

Patrick Turner wrote:

Their choice of loading the tubes and amount of FB and tube choice will
vary....

It's not pointless at all, Eeyore has injected "triode sound" into this
discussion as a supposed requirement.

Do you prefer 'pentode sound' ?

As a matter of fact I do prefer "pentode sound", as I said earlier it is
hard to beat the sound of a 5879.


Regards,

John Byrns

--
Surf my web pages at, http://fmamradios.com/

[John Byrns]

In article ,
Eeyore wrote:

John Byrns wrote:

Ian Thompson-Bell wrote:
Eeyore wrote:
Patrick Turner wrote:

so you need to spell out clearly what mics you plan to use
so that levels and attenuation can be all tailored to suit
a best outcome.

This is NOT an issue any more than it is for modern mic amps. Variable
gain
between 20 and 80 dB plus a 20dB pad for high level signals will handle
all
eventualities.

I would venture to suggest that for the vast majority of applications
and mics that 60dB max gain will be ample. At 80dB gain, even the
quietest preamps will give a S/N no better than about 50dB from a 200
ohm source.

For those of us that are not "pro audio" designers, can you tell us what
the reference is for these gain figures, gain from input to what point?

Look ****wit STOP TROLLING.

VOLTAGE GAIN = OUTPUT VOLTAGE / INPUT VOLTAGE !

How much ****ing clearer could it be ?

It was not at all clear, it wasn't clear if he was simply talking about
the gain of a microphone amplifier in isolation, or the gain of the
entire box including the microphone amplifier, the mixer, and the line
amplifier.


Regards,

John Byrns

--
Surf my web pages at, http://fmamradios.com/

[John Byrns]

In article ,
Eeyore wrote:

Ian Thompson-Bell wrote:

John Byrns wrote:

Aren't THD and IMD levels pretty highly correlated unless there are
serious frequency response anomalies involved?

Both depend on the non-linearities of transfer function of the device.
So for a particular device they will doubtless be correlated but between
different devices I don't know. As a rule IM tends to be higher than THD
and more unacceptable to the ear.

And IM truly STINKS !

In particular, high power valve power amps produce IM like it's going out of
fashion. A lot of that is likely due to output transformer issues.

I'm old enought to have mixed live bands using valve 'PAs'. The sound STANK
due to intermodulation. It was truly BAD ! Transistors were like a breath of
fresh air in comparison. A live performance with a transistorised PA could
be surprisingly close to the recording.

I'm not sure what "PAs" have to do with mixing, but I am old enough to
have mixed live orchestras using valve mixers, the resulting sound was
truly excellent!


Regards,

John Byrns

--
Surf my web pages at, http://fmamradios.com/

[John Byrns]

In article ,
Eeyore wrote:

John Byrns wrote:

Eeyore wrote:
John Byrns wrote:
Eeyore wrote:

Some how I doubt you made any SPICE models ever.

Can you explain how you come to that conclusion?

You're not technically literate enough.

Your logic is flawed, which isn't a big surprise, how does a lack of
technical literacy prevent one from messing about with the creation of
spice models?

You're still an idiot with a chip on your shoulder.

That may well be, but with that statement you have implicitly admitted
that your logic was flawed.


Regards,

John Byrns

--
Surf my web pages at, http://fmamradios.com/

[John Byrns]

In article ,
Eeyore wrote:

John Byrns wrote:

I was simply trying to point out that IMD and THD are highly correlated and
as a rule you don't have
one without the other, and they pretty much track one another

Pure nonsense. The mechanisms are quite different.

This is a perfect example of your failure to understand the underlying
theory. How are the mechanisms different? Can you propose a mechanism
that produces IMD without producing HD, and which is likely to be
encountered in an audio amplifier circuit? I didn't think so, both
types of distortion are produced by the same basic mechanism.


Regards,

John Byrns

--
Surf my web pages at, http://fmamradios.com/

[John Byrns]

In article ,
Ian Thompson-Bell wrote:

John Byrns wrote:
In article ,
Ian Thompson-Bell wrote:

John Byrns wrote:
For those of us that are not "pro audio" designers, can you tell us what
the reference is for these gain figures, gain from input to what point?


Usually it is from mic input to a final output.

So all of the specified gain, ignoring makeup gain, is not necessarily
in the microphone amplifier proper?



In practice it is pretty much all in the microphone amplifier proper. It
has to do with the balance between headroom and noise floor that I
mentioned before.

Starting with a 200 ohm source the noise seen at the mic amp input due
to that noise is about -130dBu. A perfect mic amp set to 80dB gain would
amplify this noise to -130 + 80dBu = -50dBu. Modern mic pres are pretty
near perfect and will only worsen this figure by a dB or so.

Let's look at the noise output of our perfect mic pre at other gain setting:

Gain 60dB - output noise -70dBu
Gain 40dB - output noise -90dBu
Gain 20dB - output noise -110dBu
Gain 0dB - output noise -130dBu

In practice, at around 30 to 40dB of gain, the output noise of the mic
pre itself begins to dominate - a good mic pre will have an output noise
of around -100dB, so at around 30dB or less gain the mic pre noise
begins to dominate and at lower gains it remains at about -100dBu.

So the lowest possible level of noise from the output of our mic pre is
-100dBu and at gains over 30dB it gets worse due to input noise.

OK, I think I follow you this far, for the case of the mic pre example
you are discussing, although I don't think Henry Pasternack would care
much for your use of the term "output noise".

The next question is what level should we mix at? Since our noise level
is always -100dBu or worse we want to mix at as high a level as
possible; that is we want out signal to be as far above the noise as we
can get it. Since a CD can give a theoretical 96dB S/N we want to our
mix to achieve at least this S/N which implies a mixing level around
0dBu. Since we know we can design preamps that will deliver +20dBu,
choosing a 0dBu mix level gives us 20dB headroom and a potential 120dB
dynamic range.

You seem to have left some assumptions out of this part of the
discussion and your logic here seems to be dependent on the particular
mic pre you defined in the previous paragraph. Not that it isn't
necessarily true, but you haven't presented evidence that it is
necessary to mix at 0 dBu to achieve at least a 96 dB S/N. For example
if we had a mic pre with 10 dB less gain, and 10 dB less "output noise",
it isn't obvious to me that we couldn't mix at a level of -10 dBu and
still achieve the desired 96 dB S/N. I guess I will have to try and
flesh out the example myself, but this gets back to my earlier question
about the noise contribution of the mixer network.

So since our output level also wants be be in the region of 0dBu we want
want any additional gain after mixing (except to make up for mix losses)
so in practice all our gain ends up being in the mic pre.

So mixing at -10 dBu as in my example would require an additional 10 dB
gain following the mixing network, but since the output noise of the mic
pre is also 10 dB lower, it all washes out as far as the noise
contribution of the mic pre goes. However any noise contribution of the
mixing network/circuit would be increased by 10 dB, this is the part I
don't have a grip on yet, how the actual mixing network contributes to
the noise at the output of the "mixer unit"?


Regards,

John Byrns

--
Surf my web pages at, http://fmamradios.com/

[Arny Krueger]

"RickH" wrote in message

On Jan 13, 1:45 pm, Eeyore

wrote:
RickH wrote:
Oh you'd rather have 32 DtoA conversions up front

The NUMBER of conversions taking place in parallel is
completely irrelevant.

It's a classic red herring that seeks to imply a defect
through an implied misunderstanding of the importance of
same.

You should be ashamed of yourself for resorting to such
an intellectually feeble argument.

If the mix is a sum how is that irrelevant? 32 AtoD
conversions each introduce noise, then it's all summed mathematically in
the CPU.

Almost all of the noise comes from the mic preamps and microphone, which you
have regardless.

You've obviously never done any serious audio work with an eye to finding
out where your noise and distortion come from.

Once converted to digital there is no need for there to be any added noise,
distotion, or frequency response problems.

Vs. an analog sum that is then converted AtoD once for the write.

All that does is economize on ADCs, at the cost of the slings and arrows of
analog signal processing and summing.

SOTA ADCs run around $10, while ADCs with far less noise than mic preamps
and microphones run more like $1.

[Arny Krueger]

"John Byrns" wrote in message

In article ,
Eeyore wrote:

John Byrns wrote:

I was simply trying to point out that IMD and THD are
highly correlated and as a rule you don't have
one without the other, and they pretty much track one
another

Pure nonsense. The mechanisms are quite different.

This is a perfect example of your failure to understand
the underlying theory. How are the mechanisms different?
Can you propose a mechanism that produces IMD without
producing HD, and which is likely to be encountered in an
audio amplifier circuit? I didn't think so, both types
of distortion are produced by the same basic mechanism.

The only way to have HD without IM or IM without HD is to use cleverly
designed bandpass circuits.

[Eeyore]

Ian Thompson-Bell wrote:

Ian Thompson-Bell wrote:

By the way I discovered why the original reading was 6dB higher - it was
crap induced in the test leads by my temperature controlled soldering
iron - switch off the iron and the noise drops 6dB.

I just added a few nF across the output to reduce the bandwidth a bit

Don't you know what the measurement bandwidth is ? That's CRITICAL !

Graham

[Eeyore]

RickH wrote:

Eeyore wrote:
RickH wrote:
Oh you'd rather have 32 DtoA conversions up front

The NUMBER of conversions taking place in parallel is completely irrelevant.

It's a classic red herring that seeks to imply a defect through an implied
misunderstanding of the importance of same.

You should be ashamed of yourself for resorting to such an intellectually feeble
argument.

Graham

If the mix is a sum how is that irrelevant? 32 AtoD conversions each
introduce noise, then it's all summed mathematically in the CPU.

See. You're an idiot. You've already made a stupid assertion that has no scientific
basis.

Join the club of ****wits that infest this group.

I'm DISGUSTED how technically clueless ****s like you think their posts have any
meaning. Did you learn electronics from the Beano by any chance ?

Graham

[Patrick Turner]

Ian Thompson-Bell wrote:

Patrick Turner wrote:

The THD and IMD are often quoted as being the root cause of percieved
sound differences between two circuits doing the same job.

In comparisons of 4 different brands of 6CG7 a few years ago,
some were percieved as being better/worse/different to each other.
4 listeners were present on the day to make the comparisons.
For the record, NOS Siemans 6CG7 were percieved as more detailed, far
less harsh
than new made Russian EH6CG7. We concluded that nobody in their right
mind would use an EH6CG7.

Maybe it was a bad batch, but why? ask the GOT, maybe He knows.

I later tested the preamp in which the tests were made and found the THD
could NOT have exceeded 0.01% at any time during the tests.


Yes, but what was the IM distortion?

No idea.

But if you have a single triode with 0.01% THD, mainly 2H,
then using say 1V of 75Hz and 0.25V of 5kHz, you could work it out,
and you'd find the IMD was insignificant.

RDH4 speaks about tests done on audiences to discover what
levels of THD could be detected in wide bandwidth music
and settled on 0.5%.

If we gave the benefit of any doubt to a modern audience and said the
threshold of detectability
of any distortions was where THD 0.1%, then how detectable is
distortion with only 0.01% ?

These levels of THD always bring IMD as a result.

Its easy to set up a single 1/2 6CG7 with a CCS dc feed and RL 50k so
that
at 0.25Vrms output for average levels the THD/IMD is simply not audible
and less than noise in the source signal and total circuit being used.

Bootstrapped µ-follower stages can easily be rigged for about 1/3 of the
THD
you get with the typical DC load R of 47k.

I have often seen 0.2% at 10.0 Vrms, with THD reducing about in
proportion to Vo.
All without any loop NFB.

Typical noise at the anode output of such a triode stage with gain = 18
is 36uV.

So whatever does make such tubes sound different isn't worth
quantifying.

Patrick Turner.







Ian

[Eeyore]

Patrick Turner wrote:

But if you have a single triode with 0.01% THD, mainly 2H

Under what conditions does a typical triode, say an ECC83, produce ONLY 0.01%
THD ?

I've been digging up data on valve mic preamps and was fairly shocked to see THD
specced in the 0.4% region.

Graham

[Tynan AgviÐr]

Not enough noise and distortion for you, I take it?



I will take noise and distortion any , i mean any day of the week over
sterility(ie TELARC and about 90% of the recordings of the last 5 years). I
went down the road of millennias, DPAs, GML 8302s, DCS clarity, Gordon
model 5, Sonodores, sonomas, and didnt like it at all. My ears are the
judge and jury, and if you put a truckload of neumann Ds and a fully
strapped Pyramix/DAD DAW beside one pair of RCA 44DX , a single tfunken
201 and a Stellavox and told me to choose one of em, itd take me a
nanosecond to choose the latter, though specs wise, id be going with the
losers .

Ears matter, nothing else(unless you are an engineer and have to kiss
corporate ass to put food on the table, which I do not)

How do specs sound?

[Tynan AgviÐr]

"Arny Krueger" wrote in
:

"RickH" wrote in message


Maybe the OP shuld just go digital then.

Only if he wants the best possible SQ.



I do want the best possible SQ..hence my query here.

I dont give half a **** about specs, chips, number of conversions, DSPs,
ADCs, DACs, LSBs, FFTs, BCDs, ECCs, or SMPTEs.


I am intriqued by just why during the course of any discussion about
*analog* recording/audio ;;digital acolytes have to chime in with *"man,
why dont you just go digital"* it never ever fails. there should be an
equivalent to Godwin's law for the phenomenon.

*I dont want to go digital/solid state etc because I dont ****ing like the
way it sounds for ****s sake!!!!!!Damn!In my 15 years as a professional
musician, around orchestras and bands and countless live performances, I am
damn well sure of what I like and what I dont, thank you very much. If my
precious paper specs suffer , then hell with it. My ears will be pleased at
least.*


(not directed at anyone in particular , just venting)