[Id ita]

O Rats,

Consider a standard passive EQ network, inserted after a transformer
that is coupled to a triode plate. The transformer's pri/sec ratio is
1:1. For the sake of accuracy, we have calculated R'p as a source
resistance and included it in the network. In practice, R'p and a few
other stray factors are variable, but my question is academic: How does
the inclusion of a transformer ultimately affect the source resistance?
If the ratio is 1:1, the secondary Z is about equal to plate resistance,
but how
might transformer losses figure in? Is there a rule of thumb for adding
them?

Regards,
Barry Bialos

[Eeyore]

Id ita wrote:

O Rats,

Consider a standard passive EQ network, inserted after a transformer
that is coupled to a triode plate. The transformer's pri/sec ratio is
1:1. For the sake of accuracy, we have calculated R'p as a source
resistance and included it in the network. In practice, R'p and a few
other stray factors are variable, but my question is academic: How does
the inclusion of a transformer ultimately affect the source resistance?

If it's a perfect transformer it doesn't.

Why on earth would you choose to use one ?

Graham

[Patrick Turner]

Id ita wrote:

O Rats,

Consider a standard passive EQ network, inserted after a transformer
that is coupled to a triode plate. The transformer's pri/sec ratio is
1:1. For the sake of accuracy, we have calculated R'p as a source
resistance and included it in the network. In practice, R'p and a few
other stray factors are variable, but my question is academic: How does
the inclusion of a transformer ultimately affect the source resistance?
If the ratio is 1:1, the secondary Z is about equal to plate resistance,
but how
might transformer losses figure in? Is there a rule of thumb for adding
them?

Regards,
Barry Bialos


The transformer has primary inductance which shunts both the load and Ra
of the tube in parallel.

When the reactance of the Lp = ( RL//Ra )ohms, the LF gain will be down
3dB,
and so Lp needs to be a high figure to avoid bass attenuations.

The shunt C and leakage L of the transformer will also affect the
outcome at HF,
and unless the transformer is termiated properly, a horribly peaked HF
response between 5k and 50k will result.

Get rid of the tranny in this location.
The use of a CCS anode dc load or high value choke feed to the triode
driver tube would be better
with a large value cap drive to the RIAA network.

Patrick Turner.

[Ian Iveson]

Barry Bialos wrote

Consider a standard passive EQ network, inserted after a transformer
that is coupled to a triode plate. The transformer's pri/sec ratio
is
1:1. For the sake of accuracy, we have calculated R'p as a source
resistance and included it in the network. In practice, R'p and a
few
other stray factors are variable, but my question is academic: How
does
the inclusion of a transformer ultimately affect the source
resistance?
If the ratio is 1:1, the secondary Z is about equal to plate
resistance,
but how
might transformer losses figure in? Is there a rule of thumb for
adding
them?

A google for "transformer equivalent circuit" gets lots of hits. Here
is one:

http://www.midcom-inc.com/Tech/tn17.asp

Note only the central portion of the circuit shown is relevant...see
bottom of diagram.

The resistances may not be very significant in your case. That leaves
the primary inductance, and winding capacitance in parallel, and
leakage inductance in series as shown.

cheers, Ian

[Arny Krueger]

"Id ita" wrote in message

O Rats,

Consider a standard passive EQ network, inserted after a
transformer that is coupled to a triode plate. The
transformer's pri/sec ratio is 1:1. For the sake of
accuracy, we have calculated R'p as a source resistance
and included it in the network. In practice, R'p and a
few other stray factors are variable, but my question is
academic: How does the inclusion of a transformer
ultimately affect the source resistance? If the ratio is
1:1, the secondary Z is about equal to plate resistance,
but how
might transformer losses figure in? Is there a rule of
thumb for adding them?

Your system design appears to be abysmal from a dynamic range viewpoint.

Part of that is simply because it is passive eq, which sucks from a dynamic
range perspective, and putting a questionably linear part like a transformer
before the eq just makes things worse. Looks like you are headed towards
coming up with the zero-feedback SET of RIAA phono front ends!

[Phil Allison]

"Id ita"


Consider a standard passive EQ network, inserted after a transformer
that is coupled to a triode plate. The transformer's pri/sec ratio is
1:1. For the sake of accuracy, we have calculated R'p as a source
resistance and included it in the network. In practice, R'p and a few
other stray factors are variable, but my question is academic: How does
the inclusion of a transformer ultimately affect the source resistance?
If the ratio is 1:1, the secondary Z is about equal to plate resistance,
but how
might transformer losses figure in? Is there a rule of thumb for adding
them?



** Yep - the case of a 1:1 tranny is very easy.

Just sum the primary and secondary resistances and add that to the source
resistance at mid band frequencies.

For a given tranny - that is the best you can do.




......... Phil

[Id ita]

Date: Thu, May 10, 2007, 7:31am (PDT+7) From:
(Eeyore)
Id ita wrote:

[...] How does the inclusion of a transformer ultimately affect the
source resistance?

If it's a perfect transformer it doesn't.

Why on earth would you choose to use one ?


It's not a perfect transformer, of course.
Otherwise it would raise no concerns for you or I.

It's being considered as the sonic alternative to a coupling cap, with
measurable accuracy the lesser priority.

Thank you for your interest.

Regards,
Barry

[Id ita]

Date: Thu, May 10, 2007, 9:27am (PDT+7) From:
(Patrick*Turner)
Id ita wrote:

[...]

The transformer has primary inductance which shunts both the load
and Ra of the tube in parallel.
When the reactance of the Lp = ( RL//Ra )ohms, the LF gain will be
down
3dB,
and so Lp needs to be a high figure to avoid bass attenuations.
The shunt C and leakage L of the transformer will also affect the
outcome at HF,


Hi Patrick. Excuse the botched formatting. You've described how a real
world transformer performs as a bandpass filter. Every coupling
transformer down the chain is going to nibble off some bandwidth, and
subtract from the system's aggregate response. That's a given at any
location, including where an xformer is loaded with a straight
resistance. One of my concerns right here is that the reactive
properties of the xformer will interact with the complex network that is
loading it in a frequency-dependent way, and screw the playback curve.

...and unless the transformer is termiated properly, a horribly
peaked HF response between 5k and 50k will result.

Conversely, I'm concerned that the network will not or cannot correctly
load the secondary in order to avoid peaking. The network values must be
scaled to provide this, and at the same time present a practical load to
the tube. I'd like the EQ series resistor to be 7 x Rp, but no higher
than neccessary. It's a juggling act.

Thanks very much for your interest.

Regards,
Barry Bialos

[Eeyore]

Id ita wrote:

Date: Thu, May 10, 2007, 7:31am (PDT+7) From:
(Eeyore)
Id ita wrote:

[...] How does the inclusion of a transformer ultimately affect the
source resistance?

If it's a perfect transformer it doesn't.

Why on earth would you choose to use one ?

It's not a perfect transformer, of course.
Otherwise it would raise no concerns for you or I.

You can get transformers from the likes of Lundahl that are perfect enough to be
of no concern.


It's being considered as the sonic alternative to a coupling cap, with
measurable accuracy the lesser priority.

Well..... a transformer will certainly have some 'sonic colouration' and a cap
won't. Depends what you want really !

Graham

[Id ita]

Date: Thu, May 10, 2007, 10:03am (PDT+7) From:
(Ian*Iveson)
Barry Bialos wrote

[...]
A google for "transformer equivalent circuit" gets lots of hits.
Here is one:
http://www.midcom-inc.com/Tech/tn17.asp


Thank you for sending this along. I'm aware of these models, but they
seem useful only insofar that a known transformer is under test.


The resistances may not be very significant in your case.


I'd be thrilled to take that for granted. We'll see.


Regards,
Barry Bialos

[Id ita]

Date: Thu, May 10, 2007, 9:22pm (PDT+17) From:
(Phil*Allison)
"Id ita"

...how might transformer losses figure in? Is there a rule of thumb for
adding them?

** Yep - * the case of a 1:1 tranny is very easy.
Just sum the primary and secondary resistances and add that to the
source resistance at mid band frequencies.
For a given tranny - that is the best you can do.
........ * Phil


Ah-- that's the meat of it! My best guess was close-- to add in just
the secondary resistance, and center all considerations at 1 kHz.

Your reply is appreciated.

Regards,
Barry Bialos

[John Byrns]

In article ,
"Arny Krueger" wrote:

"Id ita" wrote in message

O Rats,

Consider a standard passive EQ network, inserted after a
transformer that is coupled to a triode plate. The
transformer's pri/sec ratio is 1:1. For the sake of
accuracy, we have calculated R'p as a source resistance
and included it in the network. In practice, R'p and a
few other stray factors are variable, but my question is
academic: How does the inclusion of a transformer
ultimately affect the source resistance? If the ratio is
1:1, the secondary Z is about equal to plate resistance,
but how
might transformer losses figure in? Is there a rule of
thumb for adding them?

Your system design appears to be abysmal from a dynamic range viewpoint.

Part of that is simply because it is passive eq, which sucks from a dynamic
range perspective, and putting a questionably linear part like a transformer
before the eq just makes things worse. Looks like you are headed towards
coming up with the zero-feedback SET of RIAA phono front ends!

Oh c'mon Arny it isn't as bad as all that. Back in the days before
stereo discs passive equalizers were frequently used in broadcasting
applications. A passive LCR equalizer was often connected between a low
impedance phonograph pickup and a microphone input of the broadcast
console to play transcriptions and phonograph records, it worked just
fine. Japanese audiophiles seem to love these passive LCR networks, and
there is, or at least recently was, at least one Japanese company still
making them. Today the Japanese audiophiles seem to use them following
a stage of amplification, probably because phonograph pickups of the
correct impedance aren't commonly available today, and audiophile
preamps typically don't have low impedance microphone inputs.


Regards,

John Byrns

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

[John Byrns]

In article ,
(Id ita) wrote:

Date: Thu, May 10, 2007, 9:27am (PDT+7) From:
(Patrick*Turner)
Id ita wrote:

[...]

The transformer has primary inductance which shunts both the load
and Ra of the tube in parallel.
When the reactance of the Lp = ( RL//Ra )ohms, the LF gain will be
down
3dB,
and so Lp needs to be a high figure to avoid bass attenuations.
The shunt C and leakage L of the transformer will also affect the
outcome at HF,


Hi Patrick. Excuse the botched formatting. You've described how a real
world transformer performs as a bandpass filter. Every coupling
transformer down the chain is going to nibble off some bandwidth, and
subtract from the system's aggregate response. That's a given at any
location, including where an xformer is loaded with a straight
resistance. One of my concerns right here is that the reactive
properties of the xformer will interact with the complex network that is
loading it in a frequency-dependent way, and screw the playback curve.

...and unless the transformer is termiated properly, a horribly
peaked HF response between 5k and 50k will result.

Conversely, I'm concerned that the network will not or cannot correctly
load the secondary in order to avoid peaking. The network values must be
scaled to provide this, and at the same time present a practical load to
the tube. I'd like the EQ series resistor to be 7 x Rp, but no higher
than neccessary. It's a juggling act.

How about including some sort of conjugate network so that the
transformer sees a constant constant resistive load, or at least a
nearly constant resistive load?


Regards,

John Byrns

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

[Eeyore]

John Byrns wrote:

Japanese audiophiles seem to love these passive LCR networks

For the response errors ?

Graham

[Arny Krueger]

"Eeyore" wrote in
message
John Byrns wrote:

Japanese audiophiles seem to love these passive LCR
networks

For the response errors ?

For the *euphonic* nonlinear distortion and noise due to the bad set of
choices, vis-a-vis dynamic range?

We've been here before - with SETs.

[Arny Krueger]

"John Byrns" wrote in message

In article ,
"Arny Krueger" wrote:

Part of that is simply because it is passive eq, which
sucks from a dynamic range perspective, and putting a
questionably linear part like a transformer before the
eq just makes things worse. Looks like you are headed
towards coming up with the zero-feedback SET of RIAA
phono front ends!

Oh c'mon Arny it isn't as bad as all that.

I think it would be amusing to do a proper DBT and see whether it sounds as
bad as it looks on paper.

Back in the
days before stereo discs passive equalizers were
frequently used in broadcasting applications.

Please correct me if I'm wrong, but isn't there a possibility that
expectations for dynamic range on AM radio were less then than now for a
21st centruy high quality audio system ?

A passive
LCR equalizer was often connected between a low impedance
phonograph pickup and a microphone input of the broadcast
console to play transcriptions and phonograph records, it
worked just fine.

Weren't a lot of the recordings being played 78s, and weren't a lot of the
radio stations broadcasting in AM?

I seem to recall that the modern stereo LP had a lot to do with the rise of
FM.

Japanese audiophiles seem to love these passive LCR networks,

The whims and fancies of Japanese audiophiles are legendary, and not
necessarily for sound quality.

and there is, or at least
recently was, at least one Japanese company still making
them. Today the Japanese audiophiles seem to use them
following a stage of amplification, probably because
phonograph pickups of the correct impedance aren't
commonly available today, and audiophile preamps
typically don't have low impedance microphone inputs.

Just because some people want audio in the worst way doesn't mean we should
give it to them in the worst way.

[Patrick Turner]

Id ita wrote:



Date: Thu, May 10, 2007, 9:27am (PDT+7) From:
(Patrick Turner)
Id ita wrote:

[...]

The transformer has primary inductance which shunts both the load
and Ra of the tube in parallel.
When the reactance of the Lp = ( RL//Ra )ohms, the LF gain will be
down
3dB,
and so Lp needs to be a high figure to avoid bass attenuations.
The shunt C and leakage L of the transformer will also affect the
outcome at HF,

Hi Patrick. Excuse the botched formatting. You've described how a real
world transformer performs as a bandpass filter. Every coupling
transformer down the chain is going to nibble off some bandwidth, and
subtract from the system's aggregate response. That's a given at any
location, including where an xformer is loaded with a straight
resistance. One of my concerns right here is that the reactive
properties of the xformer will interact with the complex network that is
loading it in a frequency-dependent way, and screw the playback curve.

...and unless the transformer is termiated properly, a horribly
peaked HF response between 5k and 50k will result.

Conversely, I'm concerned that the network will not or cannot correctly
load the secondary in order to avoid peaking. The network values must be
scaled to provide this, and at the same time present a practical load to
the tube. I'd like the EQ series resistor to be 7 x Rp, but no higher
than neccessary. It's a juggling act.

Thanks very much for your interest.

Regards,
Barry Bialos

Lets assume the transformer primary inductance is high enough at the low
signal level
then the sec R termination for HF damping becomes R1 +R2 of a RIAA
passive filter.
The sum of these values just need to be kept low enough for damping at
above 5kHz.

Some trial and error using a reverse RIAA filter input signal should get
you a flat response between
20Hz and 20kHz, -3dB points.

However the tranny will have non linear inductance at low sig levels
unless its designed for the
application.
Billington or someone would have something, but the tranny might not
like the DC.

What about a balanced input, with a 1 : 1 OPT transformer?

The OPT has to have CT windings, to allow CT B+ connection and balanced
DC.
The sec can be grounded one end for a single ended output to power the
RIAA filter.

A 6DJ8 with 7mA per side and Ea = 80V would be fine, and give a gain of
about 30.

The Ra-a = 6k approx, and Lp should be more than 95H if you want -3dB at
10Hz.

If the Lp was only 9.5H, response would be -3 at 100Hz, attrocious.

You have to work all this basic stuff right out BEFORE you waste all
your time trying
to achieve what nodody has ever bothered to do.

The only way to lower transformer distortions is to drive it with
the low output resistance of a cathode follower or have a loop of NFB.

Soon the parts count rises, and you'll wonder why you tried....

I prefer using a CCS active load for the triode, or a µ-follower.
Then the standard cap coupled passive works fine.

What you are doing would be only good for MM, since MC signal is so low
that
the recovered signal at the RIAA output will be low and compete with
noise....

Transformers are used to get a phase inversion in a power amp,
or get a balanced line input to convert to se, or se output to line out,
or to get an impedance transformation downwards, as in the case of power
amp OPT,
or to get transformation upwards, as in the case of MC to tube input,
which raises voltage 10 times and transforms typical MC Rout of 10 ohms
to 1,000 ohms
without increasing noise much.

I cannot at all see any advantage that you gain sonically with a 1:1
tranny after a triode
phono input stage.

Patrick Turner.

[John Byrns]

In article ,
Eeyore wrote:

John Byrns wrote:

Japanese audiophiles seem to love these passive LCR networks

For the response errors ?

Could be, you would have to ask them. On the other hand there don't
have to be response errors anymore than with a modern RIAA equalizer.


Regards,

John Byrns

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

[John Byrns]

In article ,
"Arny Krueger" wrote:

The whims and fancies of Japanese audiophiles are legendary, and not
necessarily for sound quality.

Sound quality is a secondary issue in music listening, the important
thing is whatever it is that floats your boat, tastes vary, viva la
difference.


Regards,

John Byrns

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

[Eeyore]

John Byrns wrote:

Eeyore wrote:
John Byrns wrote:

Japanese audiophiles seem to love these passive LCR networks

For the response errors ?

Could be, you would have to ask them. On the other hand there don't
have to be response errors anymore than with a modern RIAA equalizer.

Show me an inductor with a tolerance of 1% !

Graham

[John Byrns]

In article ,
Eeyore wrote:

John Byrns wrote:

Eeyore wrote:
John Byrns wrote:

Japanese audiophiles seem to love these passive LCR networks

For the response errors ?

Could be, you would have to ask them. On the other hand there don't
have to be response errors anymore than with a modern RIAA equalizer.

Show me an inductor with a tolerance of 1% !

I'm not sure such isn't available, but why is it necessary?

Inductors were commonly used in equalizers in the days before opamps and
such became the fashion.


Regards,

John Byrns

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

[Chris Hornbeck]

On Thu, 10 May 2007 07:19:21 -0400, "Arny Krueger"
wrote:

Your system design appears to be abysmal from a dynamic range viewpoint.

Part of that is simply because it is passive eq, which sucks from a dynamic
range perspective, and putting a questionably linear part like a transformer
before the eq just makes things worse. Looks like you are headed towards
coming up with the zero-feedback SET of RIAA phono front ends!

My only personal comment is that you're being way, way, way
too tolerant of poor design.

But that's just me.

Much thanks, as always,

Chris Hornbeck

[west[_3_]]

"Phil Allison" wrote in message
...

"Id ita"


Consider a standard passive EQ network, inserted after a transformer
that is coupled to a triode plate. The transformer's pri/sec ratio is
1:1. For the sake of accuracy, we have calculated R'p as a source
resistance and included it in the network. In practice, R'p and a few
other stray factors are variable, but my question is academic: How does
the inclusion of a transformer ultimately affect the source resistance?
If the ratio is 1:1, the secondary Z is about equal to plate resistance,
but how
might transformer losses figure in? Is there a rule of thumb for adding
them?



** Yep - the case of a 1:1 tranny is very easy.

Just sum the primary and secondary resistances and add that to the source
resistance at mid band frequencies.

For a given tranny - that is the best you can do.

It's ironic Phil. You get so many angry at you, yet you make a 1 sentence
remark and end the discussion. Way to go. Shssh, don't tell anyone, but I am
somewhat of a fan, even though I'm an autistic, criminal, f--k wit...

west


........ Phil

[robert casey]

Show me an inductor with a tolerance of 1% !


If you really need to, you could use adjustable inductors.

[Id ita]

Date: Thu, May 10, 2007, 1:28pm (PDT+7) From:
(John*Byrns)
In article ,
(Id ita) wrote:
Date: Thu, May 10, 2007, 9:27am (PDT+7) From:
(Patrick*Turner)
Id ita wrote:

[...]
... I'm concerned that the network will not or cannot correctly load
the secondary in order to avoid peaking. The network values must be
scaled to provide this, and at the same time present a practical load
to the tube. I'd like the EQ series resistor to be 7 x Rp, but no
higher than neccessary. It's a juggling act.

How about including some sort of conjugate network so that the
transformer sees a constant constant resistive load, or at least a
nearly constant resistive load?
Regards,
John Byrns
--


A resistive divider across the secondary would simplify several
factors- it would swamp out changes in Rp, for one. But the loss factor
would be fatal to gain and noise. I'd like to drive the filter from a
600 ohm winding, for that matter, but it's unaffordable without adding
an extra stage.

Regards,
Barry

[Id ita]

Date: Fri, May 11, 2007, 3:37am (PDT+7) From:
(Chris*Hornbeck)

On Thu, 10 May 2007 07:19:21 -0400, "Arny Krueger"
wrote:

Your system design appears to be abysmal from a dynamic range
viewpoint.
Part of that is simply because it is passive eq, which sucks from a
dynamic range perspective, and putting a questionably linear part
like a transformer before the eq just makes things worse. Looks like
you are headed towards coming up with the zero-feedback SET of RIAA
phono front ends!

My only personal comment is that you're being way, way, way too
tolerant of poor design.
But that's just me.
Much thanks, as always,
Chris Hornbeck


The only purpose of Arny's troll, it seemed to me, was to superimpose
his belief system on
his presupposition of my belief system. An irrelevant response deserved
no reply.

I submitted a design problem that was not intended to endorse any one
particular
approach. I have no manifesto, and would not have posed my questions had
it been assumed the approach in mind was workable. It surprises the hell
out of me, Chris, that you would ignore that distinction to recieve your
portion of the prey.

Regards,
Barry Bialos

[Eeyore]

John Byrns wrote:

Eeyore wrote:
John Byrns wrote:
Eeyore wrote:
John Byrns wrote:

Japanese audiophiles seem to love these passive LCR networks

For the response errors ?

Could be, you would have to ask them. On the other hand there don't
have to be response errors anymore than with a modern RIAA equalizer.

Show me an inductor with a tolerance of 1% !

I'm not sure such isn't available, but why is it necessary?

For an accurate RIAA curve. In comparison 1% resistors are commonplace and 1% or
2% caps aren't that tricky to find either.

Graham

[Eeyore]

robert casey wrote:

Show me an inductor with a tolerance of 1% !

If you really need to, you could use adjustable inductors.

Ok, how about tempco then ?

Graham

[Eeyore]

John Byrns wrote:

Inductors were commonly used in equalizers in the days before opamps and
such became the fashion.

They still are used by some ppl. I've just been working on a 'boutique' eq
section that uses LCR filters and op-amps. It's a replica of a Neve design from
the late 70s.

Graham

[John Byrns]

In article ,
(Id ita) wrote:

I submitted a design problem that was not intended to endorse any one
particular
approach. I have no manifesto, and would not have posed my questions had
it been assumed the approach in mind was workable.

So if you considered the approach to be unworkable, why did you pose
your questions in the first place?

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

[Arny Krueger]

"Id ita" wrote in message

Date: Fri, May 11, 2007, 3:37am (PDT+7) From:
(Chris Hornbeck)

On Thu, 10 May 2007 07:19:21 -0400, "Arny Krueger"
wrote:

Your system design appears to be abysmal from a
dynamic range viewpoint. Part of that is simply because
it is passive eq, which sucks from a
dynamic range perspective, and putting a questionably
linear part like a transformer before the eq just
makes things worse. Looks like you are headed towards
coming up with the zero-feedback SET of RIAA phono
front ends!

My only personal comment is that you're being way,
way, way too tolerant of poor design. But that's just
me.
Much thanks, as always,
Chris Hornbeck

The only purpose of Arny's troll, it seemed to me, was
to superimpose his belief system on
his presupposition of my belief system. An irrelevant
response deserved no reply.

Yeah sure, I'm just commenting here in order to make trouble among the holy
and sacred. I'm sure that's what you think.


I submitted a design problem that was not intended to
endorse any one particular approach.

Good, then when you hear some negative comments, you would be happy to
consider them carefully before you dismiss them. Trouble is, you didn't do
that. In fact you immediately dismissed the negative comments.

I have no manifesto, and would not have posed
my questions had it been assumed the approach in mind was
workable. It surprises the hell out of me, Chris, that
you would ignore that distinction to recieve your portion
of the prey.

Id; your defensive, hostile, dismissive response makes everything you've
just said, say shall we say, highly questionable.

[Id ita]

Date: Fri, May 11, 2007, 7:22am (PDT+2) From:
(John*Byrns)
In article ,
(Id ita) wrote:

**I submitted a design problem that was not intended to endorse any
one particular
approach. I have no manifesto, and would not have posed my questions had
it been assumed the approach in mind was workable.

So if you considered the approach to be unworkable, why did you pose
your questions in the first place?
--

Correction: I would have not posed my questions assuming the outcome of
the experiment for better or worse.

It was expected that some readers would dismiss this idea right off the
bat for ideological reasons, or preferably some technical reason that
would benefit me to know. That comes with the territory. I build phono
amps continually, and pick a different poison every time. Hitting a
design wall-- a useful truth-- has come to be painless.

Regards,
Barry

[John Byrns]

In article ,
flipper wrote:

On Fri, 11 May 2007 07:22:45 -0500, John Byrns
wrote:

In article ,
(Id ita) wrote:

I submitted a design problem that was not intended to endorse any one
particular
approach. I have no manifesto, and would not have posed my questions had
it been assumed the approach in mind was workable.

So if you considered the approach to be unworkable,

That isn't what he said.

That was the way I parsed it, but then I was never very good in English
class. He said "would not have posed my questions had it been assumed
the approach in mind was workable." To me that sounds like another way
of saying that he considered the idea unworkable.

Haven't you ever had an 'idea' you weren't sure of, either way?

Sure but in that case I would have said I didn't know if the idea was
workable or not, and I was looking for some help in figuring out if it
was workable.

why did you pose
your questions in the first place?

He's probably asking himself the same thing by now.

I thought it was an interesting question, and worthy of discussion until
he said he didn't consider it workable. Unfortunately, workable or not,
there are too many in this group, like Arny, who consider ideas like
this to be unworthy on religious grounds, and put them down on that
basis.


Regards,

John Byrns

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

[Chris Hornbeck]

On Thu, 10 May 2007 22:20:50 -0700, (Id ita) wrote:

The only purpose of Arny's troll, it seemed to me, was to superimpose
his belief system on
his presupposition of my belief system.

Although greatly tempted to dismiss this as, well, whatever,
I'll just add that you really, really need to do some
simple calculations on the effect of the (unknown!)
source impedance on frequency response.

Then do some noise calculations including the minimum
20dB loss through the input network.

And, finally, get a grip. Jeez, it's just a hobby.

All good fortune,

Chris Hornbeck

[Patrick Turner]

Id ita wrote:


Date: Fri, May 11, 2007, 7:22am (PDT+2) From:
(John Byrns)
In article ,
(Id ita) wrote:

I submitted a design problem that was not intended to endorse any
one particular
approach. I have no manifesto, and would not have posed my questions had
it been assumed the approach in mind was workable.

So if you considered the approach to be unworkable, why did you pose
your questions in the first place?
--

Correction: I would have not posed my questions assuming the outcome of
the experiment for better or worse.

It was expected that some readers would dismiss this idea right off the
bat for ideological reasons, or preferably some technical reason that
would benefit me to know. That comes with the territory. I build phono
amps continually, and pick a different poison every time. Hitting a
design wall-- a useful truth-- has come to be painless.

Regards,
Barry

I understand Barry. I never build an amp which is exactly like the last
one either,
and I don't like boring myself to death.

But using a 1:1 tranny to couple an RIAA filter to a tube is a challenge
alright because if the issue of the primary L mainly,
and the loss of LF gain.

Possibly you could exploit the leakage L of the tranny as the HF
attenuation
element. And possibly you could get best possible bass performance if
you used
a 12AX7 for gain then with a cathode follower output, say 12AU7,
and then Rout will be less than 1,000 ohms to allow Lp to be 16H at
10Hz.

I can't see what you'd gain in sound quality the tranny and would put
the effort into
balanced and/or using j-fets to reduce noise.

Patrick Turner.

[robert casey]

Ok, how about tempco then ?


It'll drift out more than 1%? Okay...

[Phil Allison]

"robert casey"


Ok, how about tempco then ?


It'll drift out more than 1%? Okay...



** No if made right.

Ferrite inductors were widely used for 31 band graphic equalisers, master
tuning oscillators in early electronic organs ( 12 of them, one for each
semitone ) and precision notch filters with only a few Hz bandwidth in 10
kHz in AM tuners.

In all cases, drift was well under 1% for the room temp range.



........ Phil

[Patrick Turner]

Phil Allison wrote:

"robert casey"


Ok, how about tempco then ?


It'll drift out more than 1%? Okay...

** No if made right.

Ferrite inductors were widely used for 31 band graphic equalisers, master
tuning oscillators in early electronic organs ( 12 of them, one for each
semitone ) and precision notch filters with only a few Hz bandwidth in 10
kHz in AM tuners.

In all cases, drift was well under 1% for the room temp range.

Is this ferrite the similar sort used in ferrite cores shaped like
cotton reels
for speaker crossover inductors?

What is the distortion like in such cores?

In my own experiments making bridged T LCR notch filters to filter out
fundmental F from a sample signal, the ferrite core i first tried was
the ferrite rod
antenna stuff from DSE/Jaycar, and it sure made the Q higher but
introduced some 3H.
Even air cored windings reacted with a steel case to produce 3H, unless
carefully placed.

When I began to try to measure THD down to 0.002%, using L anywhere was
a PITA,
and opamps with R&C filter elements only were more reliable for low THD
and noise.
So anyone using L in an RIAA filter should worry about the distortion
caused by the iron.
Hum pick up is another worry in coils used for low level signals.

The other phenomena is that when testing THD using an LCR notch filter,
say
with 40Vrms of amp signal. My notch filter uses air core L only.
I can adjust the phase of the oscillator signal and the phase of the
1kHz notch filtered waves to within a few degrees, and thus get a very
deep null of the fundemental
right down to below the noise floor, maybe -100dB, 0.4mV, but then when
the
input signal being tested is reduced to 2Vrms, the oscillator and notch
filter have
to be re-adjusted finely with the extra fine set pots to get the deeper
null.

Seems to me the L or the C values change slightly with applied voltage
to cause a change to the Fo,
maybe by only a fraction of a Hertz, but its enough to prevent accurate
THD readings
with signal changes unless I constantly adjust for deep null, while
monitoring with a CRO to make sure I have the null I want.

Patrick Turner.





....... Phil

[Id ita]

Group: rec.audio.tubes Date: Sat, May 12, 2007, 12:45am (PDT+7) From:
(Patrick*Turner)

Possibly you could exploit the leakage L of the tranny as the HF
attenuation
element.

That would be some leak. It would be easier to exploit for the 3µS
rolloff you mentioned in another thread. Wouldn't the turnover tend to
wander with changes in the plate resistance? You'd lose the advantage of
C2 attenuating some of the noise generated across R1+R2, plus everything
else upstream.

And possibly you could get best possible bass performance if you
used
a 12AX7 for gain then with a cathode
follower output, say 12AU7, and then Rout
will be less than 1,000 ohms to allow Lp to be 16H at 10Hz.

So far, would rather use a single stage with R'p ~3K-10K, which still
allows the transformer to be a tenable beast. I did listen and learn
from your suggestion to make the stage balanced in the primary to avoid
inadequate L at low levels. Considering that also for the external noise
rejection benefits of a balanced cartridge input, and maybe reduced
sonic artifacts from the power supply/decoupling components.

Thank you again.
Barry Bialos

[Patrick Turner]

Id ita wrote:




Group: rec.audio.tubes Date: Sat, May 12, 2007, 12:45am (PDT+7) From:
(Patrick Turner)

Possibly you could exploit the leakage L of the tranny as the HF
attenuation
element.

That would be some leak. It would be easier to exploit for the 3µS
rolloff you mentioned in another thread. Wouldn't the turnover tend to
wander with changes in the plate resistance? You'd lose the advantage of
C2 attenuating some of the noise generated across R1+R2, plus everything
else upstream.

Quad exploit leakage inductance in their ESL57 step up tranny to give
attenuation of HF above quite a low F.

A 1:1 tranny wound on a C core with windings well apart on each leg of
the core
would givs a lot of leakage L, and the more turns put on for a high
amount of Lp,
the greater the LL, which is proportional to Lp.

Frankly, I think it'd be difficult to get the LL to give an exact
75uS time constant, and R&C are so much damn easier and precise.



And possibly you could get best possible bass performance if you
used
a 12AX7 for gain then with a cathode
follower output, say 12AU7, and then Rout
will be less than 1,000 ohms to allow Lp to be 16H at 10Hz.

So far, would rather use a single stage with R'p ~3K-10K, which still
allows the transformer to be a tenable beast. I did listen and learn
from your suggestion to make the stage balanced in the primary to avoid
inadequate L at low levels. Considering that also for the external noise
rejection benefits of a balanced cartridge input, and maybe reduced
sonic artifacts from the power supply/decoupling components.

The power supply noise simply should never be a problem because of
the excellent filtering and/or regulation given to
working rails in phono stages. See
http://www.turneraudio.com.au/preamp...hono-2005.html

Balanced design doesn't reduce noise unless it is common mode,
and one has to minimise it at the input first before anywhere else.

Hum pick up by stray magnetic induction pick up is very easy wherever
you have a transformer
in an input stage, so the tranny and amp should be within a thick steel
case
and shielded.

I have never built a pukka balanced phono stage which performed to
justify the extra components
and circuit complexity, and it seems pointless whe at the output someone
connects an
SE amplifier to make the speaker sound.
The SE phono circuits i routinely make have less than 0.05% N&D maximum
at any peak listening levels,
and the thd in the power amp/speakers will always be a lot more.
The input fet abolishes the noise problems one gets with tube phono
inputs without a step up tranny and MC.

Allen Wright has more to consider at http://www.vacuumstate.com

Patrick Turner.






Thank you again.
Barry Bialos